Upload 209 files

1-s2.txt

@@ -0,0 +1,51 @@
+Amidst several widely publicized suicides among adolescents with a minority sexual orientation in the past year and a half [1], there has been a national conversation about what can be done to reduce and prevent suicides among lesbian, gay, and bisexual (LGB) youths. Within this context, several individuals have initiated court cases against school districts whose policies may
+have harmed LGB students by their failure to adopt policies that protect LGB youths [2], including inclusive anti-bullying policies [3]. Although social science data are frequently used in court cases involving issues related to sexual orientation [4,5], there is currently a paucity of research examining the associations between anti-bullying policies and mental health outcomes for LGB students upon which to inform policy recommendations. The goal of the present study was to address this gap in the literature.
+Evaluating the associations between anti-bullying policies and LGB youths’ mental health has important implications for etiologic and prevention research. Population-based studies of adolescents in the United States have consistently shown that LGB youths’ rates of suicide attempts are between two and seven times higher than those of their heterosexual peers [6]. Although
+these disparities are well-documented, there is comparably less research on the processes that create risk for, or protection against, suicide attempts among LGB youths. Consequently, establishing associations between anti-bullying policies and reduced risk of suicide attempts among LGB youths would provide critical information on social and contextual protective factors within this population and aid in public health intervention efforts.
+Recent research has shown that social policies negatively targeting gays and lesbians, including constitutional amendments banning same-sex marriage [7] and the absence of employment nondiscrimination acts [8], are robust predictors of psychiatric morbidity among LGB adults. Whereas negative social policies appear to increase risk for psychopathology in LGB populations, supportive policies and programs may protect LGB individuals against the development of mental health problems [9,10]. For instance, LGB youths who attend schools with Gay-Straight Alliances report less suicidality than youths who attend schools without these programs [11]. These empirical findings are consistent with ecosocial [12] and ecological systems [13] theories, both of which highlight the importance of broad social and contextual influences, including family, school, and neighborhood factors, on health and development. Thus, several lines of evidence suggest that inclusive anti-bullying policies may be associated with reduced prevalence of suicide attempts among LGB youths [7—11]. The current study tested this hypothesis by evaluating whether LGB students living in counties with a greater proportion of school districts with inclusive antibullying policies have a lower risk of suicide attempts.
+Methods
+Sample and setting
+We obtained data from the Oregon Healthy Teens (OHT) study. Annual OHT surveys are administered to more than one third of Oregon’s eighth- and 11th-grade students attending public schools. Each year, a random sample of districts within counties and schools within districts is selected. Participating students came from 34 counties (no respondents were sampled in the remaining two counties in Oregon). The questionnaire was available in both English and Spanish. All participants were assured that the survey is anonymous and voluntary, and parents provided passive consent for their children to participate. For the current study, we pooled data from the years 2006 (when sexual orientation was first assessed) to 2008 (the most recently available data), to increase the sample size of LGB participants. Sampling for the 2007—2008 years was conducted so that each school would be asked to participate as part of the state sample once in the 2-year period, minimizing the likelihood that the same schools were sampled in multiple years. In 2008, 75.4% of the eighth- and 11th-grade students in participating schools completed the OHT survey.
+Measures
+Demographic variables including sex and race/ethnicity were obtained via self-report. Sexual orientation, which is only assessed in the survey of 11th graders, was measured with a single item asking respondents to indicate “which of the following best describes you.” Four response options were given: (1) heterosexual (straight); (2) gay or lesbian;
+(3) bisexual, and (4) not sure. Of the 33,714 original OHT respondents, 30,439 (90.3%) self-identified as heterosexual, 301 (.9%) self-identified as gay or lesbian, and 1,112 (3.3%) selfidentified as bisexual. We excluded from analyses participants who indicated that they were “not sure” about their sexual orientation (n = 653; 1.9%), which is consistent with previous studies [14]. An additional 1,209 respondents did not complete the sexual orientation item, and were also excluded. Consequently, the final sample size was 31,852. The sociodemographic characteristics of the LGB sample in the OHT study are provided in a previous report [9].
+Independent variable
+We obtained data on school anti-bullying policies at the district level from the Oregon Department of Education. We analyzed school district websites and high school student handbooks for 197 school districts. If we were not able to obtain policy information from this search (31 school districts), we contacted the individual school district to request this information. Of the 197 districts in Oregon, we were not able to obtain information for 18 districts, which we coded as missing. The missing data were largely clustered within four counties: Of the 36 counties in Oregon, 60% (21 counties) had no missing district data, 31% (11 counties) had only one or two districts with missing data, and 11% (four counties) had more than half of districts with missing data. We conducted sensitivity analyses by removing respondents from the four counties with the most missing data. The magnitude of the results remained unchanged when we removed these counties from the analyses, so the current report included all counties in the analyses.
+We first coded school district websites and student handbooks for whether the districts had any anti-bullying policies (these policies had to specifically mention bullying; harassment and antidiscrimination policies were not included in this category). Next, we coded the policies to indicate whether they contained an enumerated list of groups specifically covered by the policy, and finally, whether the enumerated list included sexual orientation. Policies had to include the phrase “sexual orientation” (e.g., in a list of protected class statuses) to be considered to protect LGB youth. Thus, these data made it possible to differentiate among (1) the absence of anti-bullying policies; (2) the presence of anti-bullying policies including specific categories (e.g., gender, race, religion), but not sexual orientation (which are hereafter referred to as “restrictive antibullying policies” [This category includes districts with antibullying policies but no enumeration of specific protected groups, as well as districts with anti-bullying policies with enumeration of groups, but no mention of sexual orientation]); and (3) anti-bullying policies that were inclusive of sexual orientation (which are hereafter referred to as “inclusive antibullying policies”).
+Because information on location of residence was available only at the county level, we aggregated the measures of antibullying policies from the district to the county level by dividing the number of school districts with anti-bullying policies by the total number of school districts in the county. We created variables of the proportion of school districts that had restrictive and inclusive anti-bullying policies within each of the Oregon counties. Of the school districts with available data, 7% had no anti-bullying policies; among districts with anti-bullying policies, 37% did not include sexual orientation as a protected
+class status. Of the counties with available data, 15% had no districts with inclusive anti-bullying policies; 18% had fewer than half of their school districts with inclusive policies; and only 15% of the counties had 100% of their school districts with inclusive policies.
+Outcome variable
+Participants were asked the number of times they attempted suicide during the past 12 months. Given the non-normal distribution, we examined suicide attempts as a dichotomous outcome. The suicide question used in the OHT was based on a measure from the Youth Risk Behavior Surveillance Survey, which showed excellent test—retest reliability (k = 76.4) [15,16].
+Covariates
+We were interested in examining whether anti-bullying policies were associated with reduced risk of suicide attempts after controlling for exposure to peer victimization, a risk factor for suicide attempts among sexual minority adolescents [17,18]. Exposure to peer victimization was assessed by asking participants, “During the last 30 days, have you been harassed at school (or on the way to or from school)?” This item had a “yes” or “no” response option.
+Statistical analysis
+The analytic strategy consisted of four steps corresponding to the four study aims. First, we calculated differences in suicide attempts and risk factors between LGB and heterosexual youth using basic descriptive cross-tabulations. Second, we tested whether the effect of inclusive anti-bullying policies on suicide attempts varies by sexual orientation. For this aim, we divided the inclusive anti-bullying policy into tertiles based on the distribution in the data. Third, we examined whether inclusive anti-bullying policies were significantly associated with suicide attempts among LGB youth after adjusting for individual-level risk factors (sociodemographic characteristics and peer victimization). For this aim, we entered inclusive anti-bullying policies as a continuous variable, with larger values indicating a higher proportion of districts with inclusive anti-bullying policies within the county. For the second and third study aims, we used Generalized Estimating Equations, a method developed for handling clustered data, in which the observations within each cluster are correlated with each other [19]. Given that OHT respondents were nested within their county of residence, we used Generalized Estimating Equations to account for the correlations among observations from each individual within the same county. Fourth, we repeated the second and third study
+aims to determine whether the presence of any anti-bullying policies (i.e., restrictive policies) buffered LGB youth against risk of suicide attempts, or whether these protective effects were only observed for policies that specifically include sexual orientation (i.e., inclusive policies). These analyses therefore tested the specificity of the protective effects of inclusive antibullying policies on rates of suicide attempts among LGB youth.
+Recent research that has disaggregated bisexuals from gay and lesbian youths has shown that bisexual adolescents are more likely to attempt suicide than gay and lesbian youths [20]; consequently, we separated these groups in all analyses. Given the relatively small number of lesbian and gay participants, we did not stratify analyses by sex. Statistical significance was set at a = .05.
+Results
+Lesbian, gay, and bisexual respondents were significantly more likely to have attempted suicide in the past 12 months than heterosexuals (%2 = 109.1; degrees of freedom = 2; p < .001). Approximately 21% of lesbian and gay youths and 23% of bisexual youths reported attempting suicide at least once in the previous 12 months, compared with 4.3% of their heterosexual peers. Lesbian, gay, and bisexual adolescents were also more likely to report past-30-day peer victimization (lesbian and gay: 60.2%; bisexual: 56.7%; heterosexual: 28.8%), compared with heterosexual youths. These group differences in peer victimization were statistically significant: %2 = 175.4; degrees of freedom = 2; p < .001.
+Associations between inclusive anti-bullying policies and suicide attempts
+We divided the inclusive anti-bullying policies into tertiles ranging from least inclusive (i.e., counties with the smallest proportion of school districts with inclusive policies) to most inclusive (i.e., counties with the largest proportion of school districts with inclusive policies). We examined the prevalence of suicide attempts within each tertile for the three different sexual orientation groups (lesbian/gay, bisexual, and heterosexual).
+Among lesbian and gay youths, the risk of suicide attempts was lowest in counties that had the greatest proportion of school districts with inclusive policies (Table 1). The proportion of lesbian and gay respondents attempting past-year suicide within the tertiles was as follows: most inclusive (16.67%); medium (19.05%); and least inclusive (31.08%). Lesbian and gay youths living in the least inclusive counties were 2.25 times (95% confidence interval [CI], 1.13—4.49) more likely to have attempted suicide in the past year compared with those in the most inclusive counties.
+In contrast, we did not observe this pattern for the bisexual or heterosexual youths. Among bisexual youths living in the most inclusive counties, 20.76% attempted suicide in the past year, compared with 25.65% in the medium and 22.11% in the least inclusive counties. Bisexual youths living in the least inclusive counties were not more likely to attempt suicide than those living in the most inclusive counties (odds ratio [OR] = 1.08; 95% CI, .75—1.56). Similarly, the proportion of heterosexual respondents attempting suicide was nearly identical across tertiles: least inclusive (4.72%); medium (3.77%); and most inclusive (4.45%). Heterosexual youths were no more likely to attempt suicide in the least inclusive compared with the most inclusive counties (OR = 1.06; 95% CI, .93—1.22).
+Having documented a protective effect of inclusive antibullying policies only among lesbian and gay youths, we next tested whether there was an association between inclusive antibullying policies and suicide attempts over and above peer victimization experiences (Table 2). In the full sample, peer victimization was significantly more likely to occur in the least inclusive (31.59%) compared with the most inclusive (29.69%) counties (Wald F = 4.44; p = .01). Even after adjusting for peer victimization and sociodemographic characteristics (sex and race/ethnicity), a higher proportion of districts with inclusive anti-bullying policies was associated with reduced risk for suicide attempts among lesbian and gay youths (OR = 0.18; 95% CI, .03—.92).
+Tests of specificity
+We conducted follow-up analyses to determine whether these effects were specific to inclusive anti-bullying policies. Results indicated that having any anti-bullying policy (i.e., restrictive policies that did not include sexual orientation as a protected class status) did not protect lesbian and gay youths from attempting suicide. The proportion of gay and lesbian respondents attempting suicide did not differ between the low-and high-inclusion categories: 21.56% and 20.00%, respectively. Moreover, after controlling for other established risk factors for suicide attempts (Table 3), restrictive anti-bullying policies did not buffer lesbian and gay youths against attempting suicide (OR = .38; 95% CI, .02—7.33).
+Discussion
+Suicide is the third leading cause of death among youths aged 15—24 years [21], and studies have consistently documented
+sexual orientation—related disparities in suicide attempts among adolescents [6,22]. However, the prevalence of suicide attempts among LGB youths does not appear to be invariant across social context. For instance, a recent study found that the risk of suicide attempts was 20% higher among LGB youths living in communities characterized by lower support for gays and lesbians (e.g., counties with a lower density of same-sex couples and fewer schools with protective policies), compared with LGB youths living in more supportive communities [9]. In addition, data from the pooled 2001—2009 Youth Risk Behavior Surveillance Survey studies showed that, across 13 states and cities that included a measure of sexual identity, rates of past-year suicide attempts among gay and lesbian youths ranged from a low of 15.1% to a high of 34.3%, over a twofold difference [23]. This geographic variation in the prevalence of suicide attempts among lesbian and gay adolescents suggests that social and contextual factors likely contribute to sexual orientation disparities in suicide attempts. The current study examined school policies, and in particular inclusive anti-bullying policies, as one social/contex-tual factor that may lower the risk of suicide attempts among LGB adolescents. We highlight four key findings below.
+First, as the proportion of school districts that adopted inclusive anti-bullying policies increased, rates of past-year suicide attempts among lesbian and gay youths decreased. Whereas 31% of lesbian and gay adolescents attempted suicide in counties where school districts were the least likely to adopt inclusive anti-bullying policies, only 17% attempted suicide in counties with the greatest proportion of school districts with inclusive policies. In models adjusted for established risk factors at the individual level (sex, race/ethnicity, and peer victimization), inclusive anti-bullying policies remained significantly associated with lower rates of suicide attempts among lesbian and gay youths.
+Second, peer victimization of all youth was also less likely to occur in counties with inclusive anti-bullying policies. These results not only suggest one potential mechanism linking inclusive anti-bullying policies to reduced risk of suicide attempts in lesbian and gay youth, but also demonstrate that policies protecting sexual minority adolescents may confer benefits for heterosexual youths as well [9].
+Third, the results documented specificity of the protective effects of inclusive anti-bullying policies to lesbian and gay youths. Inclusive anti-bullying policies did not reduce the risk of suicide attempts among bisexual youths. Recent studies that have disaggregated gay and lesbian from bisexual youths suggest one possible explanation for these results. This research has
+Data represent the Generalized Estimating Equations model predicting suicide attempts in the past 12 months. We entered restrictive anti-bullying policy policies as a continuous variable, ranging from 0 to 1.0. Higher values indicate a greater proportion of districts with inclusive anti-bullying policies. Sex: male = 0; female = 1. Race/ethnicity: non-white = 0; white = 1. Peer harassment (0 = no peer victimization in past 30 days).
+Abbreviations as in Table 1.
+illustrated that risk factors for mental health problems among bisexual youths are somewhat distinct from those for individuals with same-sex sexual orientations [24], which suggests that factors benefiting gay and lesbian youths do not always generalize to bisexual youths. Given the high rates of suicide attempts among bisexual youths observed in this study and others [20], the identification of social and contextual factors that protect bisexual youths from engaging in suicidal behaviors represents an important avenue for future inquiry. In addition, inclusive anti-bullying policies were not associated with a decreased risk for suicide attempts in the heterosexual sample. It is likely that these policies are more relevant to subgroups of heterosexual youths that are targets of bullying, such as the overweight or obese [25]. However, we did not code for other groups that were protected in these inclusive policies, which was beyond the scope of this study. This remains an important topic that can be examined in subsequent research with this sample.
+Fourth, the results documented specificity of the effects to inclusive anti-bullying policies. That is, policies had to include sexual orientation in the list of protected class statuses to be associated with significantly lower rates of suicide attempts among lesbian and gay youths. There was not sufficient evidence to indicate that restrictive anti-bullying policies (which did not enumerate sexual orientation) exerted a mental health benefit for lesbian and gay students. These results therefore suggest the importance of specifically including sexual orientation in antibullying policies that enumerate protected groups, to signal supportive and inclusive school environments for lesbian and gay youths. However, over three quarters of the school districts had restrictive anti-bullying policies; thus, most students, both LGB and heterosexual, were in districts with at least some antibullying policies. The limited range for this variable may have reduced our ability to detect significant results for the restrictive anti-bullying policies.
+This study had several limitations. The OHT survey assesses youths attending public schools. Results are therefore not generalizable to students attending private or alternative schools, or to adolescents who do not attend school. In addition, a quarter of school districts that were randomly selected declined to participate in the study. The OHT does not provide information on these school districts. Consequently, we cannot determine to what extent differential nonresponse by school district might affect the study’s results.
+In addition to issues of sampling, there are measurement limitations. In particular, the number of questions that can be included in large-scale surveys such as the OHT is necessarily limited, especially given the time constraints involved in
+administering questionnaires in classroom settings. Thus, in many cases, the OHT survey relied on single-item questions, including those for suicide attempts and peer victimization. Although the reliability of these measures has been well validated [15,16], future studies examining similar research questions would benefit from more detailed assessments of suicide attempts and associated risk factors.
+Our measure of school policies is also subject to a number of limitations. First, because the OHT study does not release information on the individual schools participating in the survey, it was not possible to obtain data on whether these policies were enforced in the schools. An important direction for future studies is to conduct detailed assessments of the extent to which school policies are consistent with daily practices. Second, school policies on bullying are determined at the district level; however, data had to be aggregated to the county, because participants’ residence was available only at this level of analysis. This approach could introduce potential error in the county variable; however, this would likely bias us toward the null, because we would not expect that misclassification is related to the proportion of students attempting suicide within the county. Consequently, these results are likely a conservative estimate of the association between anti-bullying policies and suicide attempts among lesbian and youth.
+A final study limitation is that the data are cross-sectional. Consequently, we are unable to determine whether antibullying policies are causally related to decreases in suicide attempts among lesbian and gay youth, or whether such policies are merely a marker of more supportive environments known to protect LGB youth [9]. Future studies with stronger research designs are needed to strengthen causal inferences regarding the effect of anti-bullying policies on LGB health. For instance, quasi-experimental designs can be used to compare rates of suicide attempts among LGB youth before and after inclusive antibullying policies are implemented.
+Despite these limitations, the current study has several methodological advantages for testing relationships between anti-bullying policies and suicide attempts. The large, population-based sample increases generalizability of the results and minimizes biases that may occur with convenience samples of LGB youths [26]. Moreover, unlike many previous studies [27], the LGB and heterosexual participants in the OHT study were recruited using identical sampling methods (i.e., through schools), which further diminished sampling biases [28]. An additional strength was the ability to document associations between social policies and mental health at geographic scales below the state level. Most studies that have examined the health consequences of policies targeting gays and lesbians have been conducted at the state level [7,8]. Because the OHT study released data at the county level, we were able to use measures of ecological environments that are more proximal to LGB youth.
+This study provides a significant contribution to the literature on social determinants of suicide attempts among sexual minority youths. In particular, the results indicate that the social environments in which lesbian and gay adolescents are embedded can shape their mental health, independent of individual-level characteristics. Schools are key social contexts in which important health and developmental processes unfold for adolescents [29]. In documenting associations between inclusive anti-bullying policies in schools and reduced risk of suicide attempts among lesbian and gay youth, this study lends further empirical support to the argument that social policies exert
+downstream health effects [30,31]. Consequently, altering negative social environments surrounding LGB youths through policylevel changes may ultimately lead to reductions in sexual orientation—related disparities in suicide attempts, an important public health priority [32].

A US national randomized study to guide how best to reduce stigma when describing drug-related impairment in practice and policy.txt

@@ -0,0 +1,66 @@
+INTRODUCTION
+Substance use disorders (SUD)—and opioid use disorders (OUD), in particular—are among the most stigmatized conditions in psychiatry and, indeed, throughout societies more generally [1-4]. Such stigma leads to fears of discrimination and negative repercussions that prevent or delay sufferers from seeking treatment leading to greater morbidity and mortality risk [5]. To help mitigate the
+negative personal and public health impact of stigma in relation to drug-related impairment, different medical terminology (e.g. ‘chronically relapsing brain disease’, ‘disorder’) has been adopted and deployed explicitly by US federal public health agencies [6] [e.g. National Institute on Drug Abuse (NIDA), National Institute on Alcohol Abuse and Alcoholism (NIAAA), Substance Abuse and Mental Health Services Administration (SAMHSA)], the American Psychiatric Association (APA)
+1758 John F Kelly et al.
+[7] (e.g. substance use ‘disorder’ category in DSM-5) and prominent addiction-specific medical organizations [e.g. American Society of Addiction Medicine (ASMSA)] [8]. The belief is that greater emphasis on the brain-based and medical nature of drug-related impairment will reduce stigma. At the same time, others have vehemently objected to the over-medicalization of substance-related impairment [9-11], arguing that doing so may undermine personal agency and self-efficacy among sufferers, thereby reducing prognostic optimism and the likelihood that someone would initiate or persevere in salutary change efforts. While these issues remain hotly debated, there are no existing rigorous empirical data to inform the field about which terms may be optimal and under what circumstances.
+The choice of language and terminology used is particularly important with regard to drug-related impairment, because whether or not we are aware of it, the use of certain terms can perpetuate stigmatizing attitudes that influence the selection and effectiveness of our social and public health policies for addressing them [12-14]. In fact, rigorous scientific investigations have now shown that certain common terms in the field used to describe individuals suffering from chronic drug-related impairment (e.g. ‘substance abuser’) may actually induce explicit and implicit cognitive biases that result in a perceived need for punishment rather than treatment [15-17]. Such research has made it difficult to trivialize and dismiss the terminology debate as merely ‘semantics’ or a linguistic preference for ‘political correctness’.
+Whereas several terms are used somewhat interchangeably across federal and state public health agencies, we are not aware of any rigorous research that may inform and guide the choice regarding which terms may be optimal in describing the phenomena of drug-related impairment itself. There have been recent efforts to re-assert the notion of addiction as a ‘disease’ characterized by brain-related structural deficits and functional impairment [18] (e.g. ‘brain disease’) or as a chronically relapsing variant thereof (e.g. ‘chronically relapsing brain disease’). It has also been described more generally as simply a medical ‘disease’ or ‘illness’ without specific emphasis on the brain per se, and also as a ‘disorder’ as in the current fifth edition of the Diagnostic and Statistical Manual (DSM-5) of the American Psychiatric Association (AMA) [7]. It is also often referred to more generically as a ‘problem’ (e.g. a ‘drug problem’).
+With the exception of the latter, most, if not all, these terms are used explicitly with the intention of placing the responsibility for addressing these problems squarely within the broad realm of medicine, psychiatry and public health, and to reduce blame and associated self- and public stigma in order that more people will seek and stay engaged with treatment [16,19]. Little is known,
+however, about any differential impact of the use of each of these commonly used terms when applied to drug-related impairment on well-studied dimensions of stigma, such as perceptions of blame, danger, social distance, treatment need and prognostic optimism. Although it is challenging to assess the impact of stigma directly, randomized experimental designs have been conducted to assess differences in attitudes that result from differential exposure to certain terminology typically presented within a vignette (e.g. Kelly & Westerhoff, 2010). Using this type of study design, compelling evidence has emerged from the mental health field that emphasizing more biomedically oriented genetic explanations of causes of mental illness may reduce blame attributions, but increase prognostic pessimism and perceptions of dangerousness [20] and that emphasizing brain-based neurobiological explanations for mental illness actually may increase perceptions of dangerousness, desire for social distance and pessimism about people’s likelihood of recovering.
+Although untested, it is thus conceivable that describing drug-related impairment as a ‘chronically relapsing brain disease’, while intended to diminish self-blame and stigma, may similarly increase perceptions that someone is chronically volatile and dangerous, thereby increasing attitudes of social exclusion and reducing beliefs that the affected person can recover. Knowledge of such attitudes are important, particularly in the general population, as public opinion can exert pressure for greater investment in therapeutic versus punitive criminal justice approaches to addressing drug-related impairment at local and national levels.
+To this end, using a nationally representative sample of the US general population, this randomized study is intended to shed light upon this hotly debated issue of whether differential use of commonly used terms produce attitudinal differences across dimensions of stigma and perceptions governing the likelihood of recovery. Given the recent dramatic rise in opioid use disorder and overdose deaths in the United States and several other nations, in this study we test an example relating to opioid-related impairment, in particular. Specifically, the study tested whether (1) commonly used terminology differentially affected perceptions of stigma; (2) perceptions of stigma differed depending upon whether the opioid-impaired person being portrayed was depicted as a man or a woman; and (3) any observed stigma differences across terminology depended upon whether the portrayed opioid-impaired person was a man or a woman. It is hoped that this investigation will provide some empirical basis for the choice of terminology in our clinical practices with patients, families and colleagues, as well as in our broader public health and social policies and communication efforts.
+METHOD
+Participants
+A nationally representative sample of non-institutionalized adults in the United States was recruited in partnership with Ipsos, an internationally recognized survey company, to participate in this experimental study. Participants enrolled in Ipsos’ ‘KnowledgePanel’—the largest probability-based on-line panel assembled via addressbased sampling and representative of the United States pop-ulation—were screened for eligibility in this study. The KnowledgePanel uses address-based sampling (ABS) to randomly select individuals from 97% of all US households based on the US Postal Service’s Delivery Sequence File. If necessary, Ipsos provides individuals with a web-enabled computer and free internet service. Using this Ipsos is able to include households that (a) have unlisted telephone numbers, (b) do not have landline telephones, (c) are cellphone only, (d) do not have current internet access and (e) do not have devices to access the internet. This type of broad-scale sampling helps to redress socio-economic differences in landline telephone use and internet access. Ipsos’ population-based probability sampling approach has been vetted and validated in dozens of published studies in the medical and behavioral health fields (e.g. Journal of the American Medical Association, JAMA Internal Medicine, Journal of Consulting and Clinical Psychology). Eligible people were adults aged 18 years and older and English-speakers. In order to produce unbiased estimates of population parameters from these respondents, survey data are weighted to account for selection probabilities, non-response and under-coverage. The sample is weighted to geo-demographic benchmarks obtained from the US Census Bureau’s Current Population Survey (CPS), including gender, age, race/ ethnicity, education, census region, household income, home ownership and metropolitan area. The resulting weights are used as a measure of size, which is then applied using probability-proportional-to-size, to select studyspecific samples. After data are collected, design weights are adjusted to account for differential non-response using iterative proportional fitting (i.e. raking). Finally, outlier weights are trimmed and resulting weights are scaled to the total sample size of eligible participants.
+The survey was pre-tested over 30 days in December 2019-January 2020 to estimate time to completion and identify potential pitfalls in the survey to be addressed. The official survey was administered over 16 days in February 2020. Of the 5998 participants who were sampled, 3635 completed the survey (61% completion rate). This response rate is comparable to most other current nationally representative surveys [e.g. National Epidemiologic Survey on Alcohol and Related Condi-tions-III (NESARC-III), 60.1% [21]; the 2018 National Survey on Drug Use and Health (NSDUH), 58.3%] [5].
+Non-responders to the screening question were sent e-mail reminders on days 3, 7 and 11 of the survey period. The median time it took for participants to complete the survey was 8 minutes. Participants were able to refuse to respond to an item or skip it entirely. If they skipped the item, they were provided with a warning notification.
+Procedures
+Participants were randomized to receive one of 12 vignettes (six terms x gender) describing a person who had become increasingly involved with opioids and was currently receiving treatment wherein they were learning about the exact nature of their condition described in one of six different ways as: ‘a chronically relapsing brain disease’, ‘a brain disease’, ‘a disease’, ‘an illness’, ‘a disorder’ or ‘a problem’. Vignettes also depicted the person as either male or female, but used the same gender-neutral name (Alex’) making for a total of 12 randomized cells (approximately n = 300 participants per cell). The vignette used was as follows:
+Alex was having serious trouble at home and work because of (his/her) increasing opioid use. (He/She) is now in a treatment program where (he/she) is learning from staff that (his/her) drug use is best understood as a (chronically relapsing brain disease/brain disease/ disease/illness/disorder/problem) that often impacts multiple areas of one’ss life. Alex is committed to doing all that (he/she) can to ensure success following treatment. In the meantime, (he/she) has been asked by (his/her) counselor to think about what (he/she) has learned with regard to understanding (his/her) opioid use as a (chronically relapsing brain disease/brain disease/disease/illness/disorder/problem).
+Participants were asked to read their specific assigned vignette and then answered 2 7 stigma-related questions reliably clustered within five subscales (stigma-blame; social distance/exclusion; prognostic optimism, need for continu-ingcare, perceived danger; a = 0.70-0.83). All study procedures were approved by the Massachusetts General Hospital Partners HealthCare Institutional Review Board. The study was not pre-registered on a publicly available platform, and thus results should be considered exploratory.
+Measures
+Demographic characteristics
+Demographic data were derived from the Ipsos’ existing KnowledgePanel sample of respondents (collected prior to the survey), as well as from our survey data for variables not assessed by Ipsos. Regarding existing demographic data, participants reported the following: (a) age, (b) sex
+Addiction, 116, 1757-1767
+at birth, (c) level of education, (d) race/ethnicity, (e) marital status, (f) employment, (g) household income and (h) US census region.
+Stigma and attributions
+Twenty-seven questions covering multiple dimensions of stigma and attitudes towards opioid-related impairment were administered as part of the stigma and attribution assessment. The measure comprised five distinct scales, (a ranged from 0.70 to 0.83) including: (1) blame attribution (five items), (2) prognostic optimism (five items), (3) need for continuing care (three items), social distance (five items) and attribution (nine items; AQ-9 [22,23], a = 0.72). For each question, response options ranged from ‘strongly disagree’ to ‘strongly agree’ on a corresponding 1-6 scale. Apart from the AQ-9, most items were constructed as distinct scales for the purpose of the study by the first author and/or adapted from prior work [24]
+Statistical analysis
+We first described the distribution of demographic characteristics in our study population. In this sample we evaluated the internal consistency and construct validity of the 27-item stigma measure using exploratory factor analyses and Cronbach’s alpha coefficient. Exploratory factor analysis of all 2 7 items was used to identify stigma subscales. First, a principal components analysis was used to determine the number of factors to extract. We plotted the eigenvalues on a scree plot to identify an inflection point that corresponded to the number of factors that explained a sufficient proportion of the variance (i.e. approximately 5% or greater) in stigma. Factor loadings were estimated using orthogonal or oblique rotation depending upon whether factors displayed a low or moderate/high intercorrelation, respectively. We eliminated items with low item-total correlations as well as low factor loadings (X < 0.35) and high uniqueness. Total scores for each subscale were calculated as the sum of all retained items. Using unadjusted linear regression models, we estimated the mean difference (i.e. beta coefficient) in each of the final stigma subscales as a function of: (1) opioid-related impairment terminology, (2) gender of the vignette character and (3) gender of the vignette character stratified by opioid-related impairment terminology. In the opioid-related impairment terminology regression models, ‘chronically relapsing brain disease’ was included as the reference group. In the models examining stigma as a function of the gender of the vignette character in the full sample, and when stratified by opioid-related impairment terminology, female was the reference group. Given the potential for more subtle nuances to be obscured if a participant had a language other than English as a first language, we conducted a sensitivity analysis excluding
+all people who spoke a language other than English in their home to determine whether language fluency influenced the observed findings. All analyses were conducted in Stata version 14 and incorporated sampling weights.
+RESULTS
+Characteristics of study population
+The sample included 3635 adults living in the northeast (17.5%), Midwest (20.8%), South, (3 7.9%), or Western (23.8%) region of the United States. On average, participants were 47.8 years of age and most were female at birth (52.4%), non-Hispanic white (63.1%), married (54.9%), had at least some college education (61.0%), were working (59.6%) and reported an income > US$50 000 (68.3%; Table 1).
+Psychometrics of the stigma and attribution scales
+We extracted five factors based on the eigenvalues, scree plot and variance explained in the overall stigma construct (see Supporting information, Appendix S1). Of the 27 items, three were initially dropped due to low factor loadings and high uniqueness (items 1, 13, 15). Oblique rotation was used to estimate the factor loadings due to the moderate-high correlation of extracted factors. Low internal consistency of the ‘need for continuing care’ subscale resulted in the removal of two additional items displaying low item-total correlations (items 17 and 19; see Table 2 for list of all items). Final stigma subscale reliability coefficients are shown also in Table 2).
+Differences in stigma and attitudes toward opioid-related impairment in the US adult population as a function of terminology and gender
+Main effect of terminology
+Relative to participants who were randomized to receive the vignette describing the opioid-impaired person as having a ‘chronically relapsing brain disease’, participants whose vignette included any of the other five terms (brain disease, disease, illness, disorder, problem) reported significantly higher levels of blame attribution toward the individual with the opioid-related impairment (Table 3). We identified a relationship suggesting that blame attributions were highest for individuals with a ‘problem’ or ‘disease’, moderate blame attribution for individuals with an ‘illness’ or ‘disorder’ and lower levels of blame attribution for individuals with a ‘brain disease’ followed by the lowest levels of blame attribution for individuals with a ‘chronic relapsing brain disease’ (Fig. 1).
+Of note, participants who were randomized to receive the vignette describing the depicted person as having an opioid ‘problem’ were more likely to attribute more
+Addiction, 116, 1757-1767
+Pct = percent; CI = confidence interval.
+personal blame for the opioid impairment; however, at the same time they were more likely to view that same person more positively in terms of viewing them as being less dangerous, viewing that person as being more able to recover from their opioid impairment and less likely to need continuing care than those participants who were randomized to the term ‘chronically relapsing brain disease’. In addition to characters portrayed as having a ‘problem’, characters described as having a ‘disorder’ or ‘brain disease’ were also perceived as less likely to require continuing care relative to characters portrayed as having a ‘chronically relapsing brain disease’. We did not find any significant differences in perceptions of need for social distance as a function of the different vignette terms.
+Main effect of gender
+When comparing the effect of gender of the character portrayed in the vignette (collapsing across terminology), we found that, when the person with the opioid-related
+impairment was described as a male, participants attributed significantly less blame, but a desire for greater social distance and expressed higher levels of perceived dangerousness relative to participants who were randomized to a vignette with a female character exhibiting opioid-related impairment. We did not identify significant differences in prognostic optimism or need for continuing care between participants who were randomized to a vignette with a female versus male character with opioid-related impairment.
+Effect of gender by terminology
+We further examined whether gender modified the effects of exposure to the opioid-related impairment terms (or null effects) observed in the main effects analysis (Table 3; Fig. 2). We found that decreased blame attribution toward male characters relative to female characters with opioid-related impairment was observed only when the term ‘chronically relapsing brain disease’ was included in
+the vignette. Specifically, participants were less likely to attribute blame to males with a chronically relapsing brain disease relative to females with a chronically relapsing brain disease. We did not identify any significant differences in blame attribution by gender when other terms were included in the vignettes.
+We found that males with a brain disease were significantly less likely to be perceived to need continuing care relative to females with a brain disease. Across all terms, participants expressed a desire for increased levels of social distance from males relative to females exhibiting opioid-related impairment; however, this was only statistically significant when the opioid-related impairment was referred to as a ‘brain disease’, a ‘disease’ or an ‘illness’. Males with a ‘disease’ or ‘illness’ were also perceived as more dangerous relative to females whose opioid impairment was described as a ‘disease’ or an ‘illness’. Results of the sensitivity analysis restricting our sample to participants who reported English as the primary language spoken at home (84.7% of the sample) did not reveal notable differences in the pattern of effects of gender or term on the stigma and attribution subscales.
+DISCUSSION
+Using a large nationally representative sample of the US general population and a randomized design, our study examined the impact of exposure to different common terms used to describe someone suffering from opioid-related impairment (i.e. ‘chronically relapsing brain disease’, ‘brain disease’, ‘disease’, ‘illness’, ‘disorder’, ‘problem’) on perceptions of several dimensions of stigma (e.g. blame, dangerousness), treatment need and prognostic optimism. Findings were nuanced with differential effects observed across terminology, gender and dimensions of stigma.
+In terms of the main effect of terminology, perhaps the most notable finding was that whereas there were beneficial stigma-reducing effects observed for certain terms on certain stigma dimensions, there was not one clear single term that produced beneficial effects across all dimensions of stigma, treatment need and prognostic optimism. Specifically, exposure to the ‘chronically relapsing brain disease’ term was associated with the lowest levels of stigmatizing blame attributions; in fact, exposure to any other term was associated with a significant increase in stigmatizing blame although, intriguingly, the blame effect was related in a linear ordinal fashion with ‘problem’, resulting in the greatest stigmatizing blame attribution. In contrast, study participants who were exposed to the person described as having an opioid ‘problem’ compared to ‘chronically relapsing brain disease’ exhibited the strongest beliefs that the person could recover (Fig. 1), were less dangerous and less likely to require continuing care. These findings support the use of the ‘chronically relapsing brain disease’
+term to reduce stigmatizing blame, but simultaneously suggest that this may not be the best term to use to convey the more positive notion that someone with opioid-related impairment is approachable and can recover; in that case, the less medical and more generic, ‘problem’ term may be optimal.
+In terms of gender of the subject, compared to a man, a woman exhibiting opioid-related impairment was judged significantly more harshly—as more to blame. Conversely, when study participants were exposed to a male versus a female character, they seemed more afraid and rated both social distance and danger higher for a man than a woman with the same level of opioid-related impairment. It is perhaps expected that a man would be viewed as more dangerous and for people to want to stay further away from a man than a woman due to greater perceived aggression,
+but it is noteworthy that a woman was judged more harshly and more personally to blame for exhibiting opioid-related impairment than a man.
+Aspects of this pattern became clearer and more pronounced when examining the results of the stratified models. When described using the ‘chronically relapsing brain disease’ terminology women may be viewed as more personally responsible, suggesting a potentially harsher and less forgiving social stance against women—even when exhibiting the same level of opioid-related impairment. This may be a case of socially stereotyped exonerating expectations that ‘boys will be boys’ (i.e. ‘bad’ behavior is to be expected and is excusable) and that ‘girls should behave’, thereby implicitly assigning greater levels of expected pre-programmed externalizing behavior and impulsivity regarding male behavior. However, the pattern
+Addiction, 116, 1757-1767
+is complex, as although men may be viewed less harshly for exhibiting opioid-related impairment when described in that manner, they are more likely to be viewed as more dangerous and thus socially ostracized and excluded overall compared to women.
+Limitations
+Observed differences were small in absolute magnitude, and the extent to which such differences may translate into actual real-world differences in terms of behavior of the general population is not known; for example, whether this would mean voting for a particular policy measure or not (e.g. increased appropriation for treatment). The set of six terms used as the levels of the independent variable is highly applicable within a US English-speaking cultural context; applicability in other cultures could vary. It would be very helpful to know also what terms drug-impaired person themselves would regard as either helping or discouraging them. Future research should examine this. Also, we used opioid-related impairment in this study as a specific example of ‘drug-related’ impairment—we do not know the extent to which observed differences in the stigma dimensions would generalize to other substances. Also, although the focus here was to examine general (main) effects in response to certain commonly used terminology, this pattern of findings could be moderated by specific respondent characteristics (e.g. personal history of a substance problem), which is worthy of further investigation. Finally, we explored the covariance and internal consistency of items to identify meaningful subscales; however,
+this 2 7-item measure has not been previously validated. We summed the item scores within each subscale for this analysis to be consistent with the way these items have been previously scored, which may have also introduced measurement error. Further research exploring the measurement and criterion validity of these stigma subscales is needed to confirm their ability to assess substance use stigma and related attitudes.
+Implications for practice and policy
+In summary, findings suggest that there may not be one single recommended term that can be applied across the board to meet all desired clinical and public health goals when attempting to reduce stigma. Choice of terminology may depend on the purpose of communication: to reduce stigmatizing blame, the more biomedical ‘chronically relapsing brain disease’ terminology may be optimal; to increase prognostic optimism and decrease perceived danger and social exclusion of affected people’s use of non-medical terminology (e.g. ‘opioid problem’) may be optimal. Findings also suggest that women may be judged more harshly than men, possibly due to broad cultural sex-based stereotypes governing differential acceptability of opioid-related impairment; and men, overall, may have more difficulty being trusted and reintegrating into society due to greater fears that they present more danger.
+Declaration of interests
+We have no conflicts of interest to report. Dr. John Kelly has received funding from the United States national institutes
+of health (NIH) as well as the US substance abuse and mental health services administration (SAMHSA), state governments, and private foundations to conduct research on addiction and its treatment.

A study on household headship, living arrange.txt

@@ -0,0 +1,91 @@
+1.	Introduction
+In most parts of the world, suicide rates in older adults are generally higher than their younger counterparts (Conwell, 2014). In order to comprehend the severity of late-life suicides, not only the absolute suicide rates but also their relative suicide age ratios (i.e., suicide rate ratios between older adults versus the younger age groups) within the respective population should be considered as well. For example, according to the Global Burden of Disease (“GBD”) Study in 2015, suicide
+rates in the population aged 60 years and above in China, Austria and Ukraine were similar at around 28.0 per 100,000. However, due to greater variations in suicide rates among the younger age groups, the old to non-old suicide rate ratios of those three countries were 4.64, 2.29, and 1.30, respectively. Suicide age ratios vary significantly among countries and regions of the world, thus providing a new perspective to understand the meaning of suicide rates in old age across the globe. Greater understanding of the age patterns of suicides could result in potential preventive solutions (Snowdon et al., 2017). To the best of the
+authors’ knowledge, there are no existing studies that have explored the global variations of suicide age ratios and their associated factors.
+Families are valuable resources in not only providing caregiving but also imparting a sense of worth, lasting emotional ties, and human dignity to elders in their later years (Walsh, 2016). From a life course perspective, previous researchers have found that higher risks in late-life suicides are associated with the unique experiences of the elderly adapting to age-related challenges and family dysfunction (Chan et al., 2014; Chang et al., 2017; Duberstein et al., 2004; Park and Moon, 2016; Purcell et al., 2012; Rubenowitz et al., 2001; Van Orden et al., 2015). However, whether the socioeconomic status of older people within families is associated with suicide risks has never been properly examined. In this study, loss of socioeconomic status in the older adults within their families was measured by three critical constructs: i) loss of household headship, ii) loss of residential independence, and iii) loss of pension support.
+With regard to family status in a culture-based index, being the “head of household” indicates the importance of a family member which is related to the power to control and allocate the family's economic and social resources (Phua et al., 2001). Loss of family headship therefore represents an important life-stage transition associated with the fundamental questions of independence and authority that lend sociological meaning to the concept of old age (Gordon et al., 1981).
+Living with one's children reflects loss of independence, which is a valued condition. Owing to the stigma of dependency in the dominant culture, most older adults in good health prefer to maintain a separate household from their children, yet sustaining frequent contacts, reciprocal emotional ties and mutual support in a pattern aptly termed “intimacy at a distance” (Blenkner, 1965; Walsh, 2016). With the advancement of telecommunication and transportation technologies, high levels of geographic mobility in modern societies have significantly supported the aforementioned living arrangement (Phua et al., 2001). Nonetheless, the transition from independent living to co-residence with the younger generations in later life is common today, reflecting reduced autonomy of the older adults in family life.
+Financial independence is important for the older adults in keeping their authority in the family. For instance, those who are more financially independent will be consulted more frequently than those who are supported economically by their children (Williams et al., 1999). In other words, economically independent elders could play a considerable role in family decisions. The pension scheme, a well-known policy to maintain financial security, led to significant reductions in poverty rates among older adults (Lloyd-Sherlock et al., 2012). Moreover, pension receipt directly affects well-being of retired older adults with a low economic status (Ju et al., 2017). Earlier studies have demonstrated that low financial status in older adults may act as a stressor that exacerbates any ongoing deterioration in psychological well-being and contributes to suicide risk (Almeida et al., 2012; Duberstein et al., 2004).
+In this study, it is hypothesized that lower late-life socioeconomic status within families would significantly increase suicide rate among the older adults. In order to take into account different base rates among different countries, suicide age ratios are used instead. To be more specific, loss of household headship, dependence in residence, and receiving no pension might elevate the suicide age ratios. The aim of this study is firstly to examine the variability of suicide age ratios in the world, and secondly to illustrate the associations of suicide age ratios with potential socioeconomic factors including household headship, living arrangement, and whether in receipt of pension in later life.
+2.	Methods
+2.1.	Data and measures
+Suicide age ratio was measured by the old (>=60 years) to non-old (<60 years) suicide rate ratios, which was our dependent variable of
+primary interest. The suicide data in the year of 2015 were obtained from the Global Burden of Disease (“GBD”) Study (Global Burden of Disease Collaborative Network, 2016) and suicide mortality was identified by the International Classification of Diseases, 10th Revision (“ICD-10") codes X60-X84 (self-harm). Suicide age ratios for the 173 regions were computed in the world.
+In this study, lower socioeconomic status within family was conceptualized by three aspects: losing household headship, living dependently, and having no pension, which reflected the honorary, residential and economic status of older adults, respectively. Furthermore, based on the data available for consistent comparisons across the nations worldwide, we used three variables to measure older people's socioeconomic status: i) the percentages of households with heads aged 60 years and above, ii) the percentages of households with both older adults aged 60 and above and children under 15, and iii) the proportions of the population above retirement age receiving a pension.
+The percentages of households with the heads aged 60 years and above were obtained from the United Nations Report on Household Size and Composition Around the World 2017. The head of household was nominated by family members in the census or survey. Elderly headship rate is calculated by dividing the number of heads aged 60 years or over identified on the household roster of the census or survey by the total number of household heads (United Nations, 2017). The data were the latest available estimates (i.e. the data for the most recent years) between 1990 and 2015 for 141 regions and ranged from 12% in North Korea to 44% in Italy.
+The percentages of households with both older adults aged 60 and above and children under 15 were also obtained from the United Nations Report on Household Size and Composition Around the World 2017. It is calculated by dividing the number of households with at least one member under age 15 years and at least one member aged 60 years or over by the total number of households (United Nations, 2017). The data represented estimates from 1990 to 2015 for 125 regions and ranged from near 0% in Germany and the Netherlands to 34% in Gambia.
+The proportions of the population above retirement age receiving a pension were extracted from the United Nations Statistics Division. It is calculated by dividing the number of population above retirement age receiving a pension by the total number of population above retirement age (United Nations Statistics Division, Department of Economic and Social Affairs, 2017). Among the latest available data from 132 regions of the world during the period 2010-2016, the values on pension ranged from 0.93% in Myanmar to 100% in many European Countries.
+2.2.	Analytic strategies
+Suicidal rates for each country were age-standardized by the standard structure of the world population in 2015. In this study, the threshold of the older adults was 60 years and above. In order to take into account the respective suicide rates in each of the countries, suicide age ratios were calculated and a world map was constructed according to the different levels of suicide age ratios including <1.0, [1.0, 2.0), [2.0, 3.0), [3.0, 4.0), and >=4. Scatterplots of three exposure variables (i.e., household headship, living with descendants, and recipient of pension) and log-transformed suicide age ratios were presented in the figures in the Appendix (Appendix Figs. A1-A3).
+Forest plots were performed to assess whether elderly household headship, living with descendants, and recipient of pension moderated the worldwide patterns of suicide age ratios. Three exposure factors were not modeled as continuous variables as the relationships between them and suicide age ratios were not linear. In the moderation analyses, percentages of the elderly heads were modeled as a dichotomized variable by the first quarter point: the higher (>19%) versus the lower (<=19%). Similarly, regions were classified by median value into higher (>11%) versus lower (<=11%) percentages of late-life co-re-sidence of the elderly with their descendants. Percentages of the elderly
+receiving a pension above retirement age were also grouped as a dichotomized variable by median value: the higher (>73%) versus the lower (<=73%). It could be seen clearly that from the United Nations Report on Household Size and Composition around the World 2017, the vast majority of countries in Africa and Asia had the very low percentages of elderly headship less than the first quarter point at 19%. Unlike the population in these Africa and Asia countries shared with the common practice of multi-generational living arrangements, older adults in other countries prefer to maintain a separate household from their children, which leads to higher percentages of elderly headship than 19%. Therefore, on the cultural and empirical bases, the first quarter point is preferable to the median split as the cut-off for the household headship variable. Pooled suicide age ratios with 95% confidence interval of associations between the suicide age ratios and the potential affecting factors were calculated using the Comprehensive Meta-Analysis software program. The software program takes population size of each analyzed region into account. Total between-group variance (“Total Qb-) was then calculated to examine the differential moderation effect among the different subgroups. Gender-specific forest plots were also constructed to examine the variance of the moderation effect between men and women.
+Crude and adjusted regression analyses were then utilized to estimate the extent to which elderly household headship, living with descendants, and recipient of pension affect suicidal age ratios in the world. Stratified analyses were also performed based on gender-specific data. The outcome variable was log-transformed as according to the Kline's rule, i.e., skew index absolute value <3; kurtosis index absolute values <10 (Kline, 2005), the distribution of suicide age ratios was not normal. The countries with missing data on the independent variables were handled by the Listwise Deletion method and were not used in the analyses.
+3.	Results
+From Fig. 1, there were significant variations of suicide age ratios across different regions in the world. On the whole, the suicide age ratios were higher than 1.00 in most parts of world, indicating that worldwide, suicide rates in older adults were generally higher than the younger population. The highest old to non-old suicide rate ratios were found in the Western Pacific and African regions.
+Appendix Table 1 shows that there were strong correlations among the suicide age ratios and the potential factors. To be specific, higher suicide age ratio was significantly correlated with lower percentage of elderly household head (r =-0.36, P <0.01), higher percentage of coresidence of the elderly with their descendants (r = 0.37, P <0.01), and lower percentage of the population receiving a pension above retirement age (r =-0.51, P <0.01). Scatterplots of three exposure variables, i.e., household headship, living with descendants, recipient of pension, and log-transformed suicide age ratios were presented in Appendix Figs. A1-A3, respectively.
+Fig. 2 presents the forest plots of suicide age ratios between countries with the higher versus the lower percentages of household heads aged 60 and above. Regions with higher elderly headship percentages had the lower suicide age ratio (1.69), whereas regions with lower percentages of elderly headship had the higher suicide age ratio (2.73). There was a significant difference between the higher and lower subgroups (Qb = 7.57, P = 0.01). In terms of the gender-specific analyses, the impact of household headship on suicide age ratios was only found in men (ratios = 1.77 vs. 2.92, P = 0.02) but not in women (ratios = 2.10vs. 2.54, P = 0.55).
+Fig. 3 shows the forest plots of suicide age ratios among countries with higher versus lower percentages of co-residence of the elderly with their descendants. As to the overall population, regions with higher percentages of co-residence of the elderly with their descendants had the higher suicide age ratio (2.72), whereas regions with lower percentages of co-residence of the elderly with their descendants had lower suicide age ratio (1.39). There was a significant difference between the higher and the lower subgroups (Qb = 12.14, P<0.01). In addition, the impact of co-residence of the elderly with their descendants on suicide age ratios could be observed in men (ratios = 2.83 vs. 1.56, P = 0.01) but not in women (ratios = 1.88vs. 1.57, P = 0.16).
+Referring to the forest plots of suicide age ratios in countries with higher versus lower percentages of older adults receiving a pension (Fig. 4), regions with higher percentages of the elderly receiving a pension had lower old to non-old suicide rate ratios, whereas regions with lower percentages of the elderly receiving a pension had higher old to non-old suicide rate ratios. The significant impact of pension on suicide age ratios could be observed in the overall population (ratios = 1.42vs. 2.76, P<0.01), men (ratios = 1.56vs. 2.91, P<0.01), and women (ratios = 1.64vs. 2.66, P<0.01).
+As seen from Models 1-3 in Table 1, the crude regression analyses showed that the lower status of the elderly within a family in terms of the loss of household headship, dependent dwelling and having no pension, were significantly associated with higher suicide age ratios in overall population and both genders (P <0.01). Since the factors were correlated with each other, adjustments had also been made to the regression analyses. In the adjusted model (Model 4), receiving no pension remained to be a significant determinant for both overall population (P = 0.01) and men (P<0.01) but not for women (P = 0.29), and loss of household headship was only significant for men (P = 0.05) but not for either overall population (P = 0.22) or women (P = 0.55), whereas the elderly living with their descendants was no longer significant for either overall population (P = 0.60) or both genders (men: P = 0.72; women P = 0.11).
+4.	Discussion
+The present study reveals that worldwide variations in suicide age ratios were associated with constructs reflecting the socioeconomic status of the older adults within families. Relatively higher suicidal risks in later life were linked to loss of domestic headship/authority, living with their descendants, and receiving no pension. In the case of the absence of pension provision, it showed robust effects on higher suicide age ratios worldwide. The culture-based indicator of intra-family status revealed that household headship was more sensitive in men than in women. The impact of co-residence with the younger generations on suicide age ratios was however controlled by the elderly economic status and household headship.
+Many previous ecological studies have examined how certain factors such as mental health funding and mental health service provision (Shah and Bhat, 2008), life expectancy and markers of socioeconomic
+status and health care (Shah et al., 2008), and elderly dependency ratios (Shah et al., 2008) are specifically associated with suicide rates in later life but without considering the relative suicide rate ratios. The present findings suggested that certain socioeconomic factors may lead to higher suicide rates in older adults and consequently higher suicide age ratios. Based on the present research, for cross-national comparison, the ratio was more robust than the rate itself. Suicide age ratio was used as it was a better-chosen indicator to compare the prevalence rates of two specific population groups, namely, the older adults versus the non-older adults. In addition, as the quality of GBD suicide rates data was not so reliable in low- and middle-income countries, presuming that there is no differential underreporting by age, exploring suicide age ratios may be better able to address concerns about potential underreporting of absolute age-specific suicide rates.
+The family life cycle theory has placed the nuclear family as a group with its regular patterns of expansion, transition, and contraction (Mattessich and Hill, 1987). The present findings can well be understood from the family developmental perspective. Families in later life are facing the graying transitions and challenges. With the structural contraction of a family from a multi-generational household to an elderly couple or single parent, changes brought about by retirement, grandparenthood, illnesses, deaths, widowhood and so on, alter complex relationships within a household, often requiring family support, adjustments to losses, reorientations, and reorganizations (Walsh, 2016). Many disturbances such as mental problems are associated with losses in family adaptation and moving to the stages of “empty nest” and “aging families” such as loss of household headship, loss of independent residence and loss of financial security.
+This study is the first to detect that loss of household headship, the culture-based indicator of domestic status, was significant to the suicide age ratios in men but not in women. The potential explanation lies in the cultural construction of traditional gender roles in a family, namely, the patriarchal authority of masculinity, but there was no such cultural expectation for women. According to traditional culture, patriarchy was mainly based on the construction of the norm of the male as breadwinner (Seccombe, 1986). Other family members may also have difficulty with the retirement of the male head, accompanied by losses of his job-related status and social network (Walsh, 2016). The loss of the elderly male as the household head signified loss of authority, and with it his self-esteem, and replaced by his adult children within the family. The role of the male gender tends to emphasize greater levels of strength and independence, and reinforcement of this gender role often deters the males from seeking help in suicidal thoughts, feelings, as well as depression (Zhang, 2014). The present study suggests the need for further research on how to enhance the resilience of the males in later life and how to renegotiate their relationships to achieve a new balance with other family members after their loss of headship.
+This study further illustrates that the impact of the suicide risk of the elderly living with the younger generations was relatively mitigated by the recipient of pension and household headship of the elderly. Predominantly, recipient of pension showed robust effects on higher old to non-old suicide rate ratios in both men and women in the world. In other words, depending on the late-life status of the elderly, especially on whether in receipt of pension, living with their descendants has dual effects on the well-being of the elderly within the family. If the seniors have economic independence such as a pension, they do not have to become a financial burden on their children, and can even provide better grandparenthood, which in turn benefits their health. On the other hand, having no pension can significantly strain relationships with cohabited descendants. Therefore, those older adults who have lost their jobs and benefits should find new work or make contributions to the family such as housework and caring for the grandchildren, but they should always be aware of facing age discrimination. Owing to the stigma of dependency in the dominant culture, based on the Interpersonal Theory of Suicide, perceived burdensomeness could increase the suicide risks for the older adults (Jahn et al., 2011).
+According to previous studies, living arrangement for the elderly also yielded the most inconsistent and mixed effects on late-life suicidal risks (Chang et al., 2017). Thus, the present observation on the associations between the elderly living with their descendants and late-life suicide risk should be highly context dependent. For instance, living with children in the Chinese community is more than an indication of dependency, however, it is usually a cultural expectation that children take responsibility to support their older parents and show their filial piety. This seems to contradict the intention to measure loss of residential independence. Another problem is that our measurement of living arrangement would include households with only older adults and children, which indicates that grandparents supporting grandchildren with the absence of parents, in contrast to older adults with a loss of residential independence as intended to be measured as well. Moreover, as the Interpersonal Theory of Suicide posits, perceived burdensomeness and thwarted belonging are both powerful drivers of suicide in later life (Jahn et al., 2011). Hence, there should be more future studies to examine the effects of suicide on the older adults living with family members.
+The present ecological findings suggest that strategies to enhance the socioeconomic status of older adults may be important to prevent suicides in later life both within and across countries on a grand scale. At the base of a 5-tier health impact pyramid, interventions with the greatest potential impact are efforts to address the socioeconomic determinants of health (Frieden, 2010). The present findings provide important evidence to highlight the substrata role of socioeconomic factors in public health as well as late-life suicide prevention across countries. Although the exact mechanisms by which socioeconomic
+status exerts its effects are not always apparent, lower status such as the elderly losing the domestic headship/authority, dependently living with their descendants and being without receiving pension, could ostensibly increase exposure to environmental hazards (Wood, 2003). Moreover, it should be noted that social policies to enhance the late-life socioeconomic status are highly context dependent. For example, the present results revealed that higher suicide age ratios could be found in Western Pacific and African regions rather than other places in the world. According to the collected data, the proportions of the population above retirement age receiving a pension were especially low in most Western Pacific and African countries, as opposed to nearly 100% in most European countries. Therefore, in many middle- and low-income countries, priorities in social policies may include concentrating on alleviating late-life poverty and keeping financial security. By contrast, whereas in the well-off regions, eliminating socio-cultural ageism by education and legislation would be more imminent.
+Nonetheless, the present ecological findings could have important implications for suicide research and prevention on older adults at individual and family levels. Firstly, in future research and interventions, both qualitative and quantitative investigations need be made on how particular risk factors such as loss of headship, living with their descendants and receiving no pension increase the likelihood of older adults at some point displaying suicidal behaviors, and how protective factors such as family support build resilience against suicidal behaviors and thoughts. In addition, as is well known, the first driver of decreased suicide mortality is early detection of individuals at risk. With the benefit of this study, risk factors relating to lower domestic status such as loss of headship and receiving pension ought to be the main target of early detection efforts in the prevention of suicides in older adults. Thirdly, this study strongly highlights the gatekeeper role of family in late-life suicide prevention. Relational resilience can be strengthened as family members pull together to reshape the elders’ lives, plan their financial security, and explore new interests to provide meaning and satisfaction for them (Walsh, 2016).
+However, it is worth noting that this present study has several limitations. Firstly, there is the issue of the quality of the data on suicides, which is often lower in developing countries and may lead to underestimation of suicide deaths (WHO, 2014). Estimates from GBD for many countries, particularly locations in sub-Saharan Africa, have uncertain validity because there are limited vital registration data in these countries, and thus available data from a few neighboring countries may be used to impute the missing data, leading to similar estimates in these sub-Saharan African countries. Therefore, sensitivity analyses were conducted to check the robustness of our findings by excluding countries without vital registration as indicated in 2014 WHO report of suicide (WHO, 2014). The results of the sensitivity analyses showed that higher suicide age ratios were significantly found in countries with lower percentages of the elderly being heads of households (ratios = 1.50vs 2.62, Qb = 9.10, P = 0.003), higher percentages of co-residence of the elderly with their descendants (ratios = 2.50 vs 1.38, Qb = 8.77, P = 0.003), and lower percentages of the elderly receiving a pension (ratios=1.41 vs 2.58, Qb=10.27, P = 0.001). The findings of the sensitivity analyses were generally the same as the analyses with overall countries, which indicated the robustness of our findings. Secondly, as in prior GBD studies, the accuracy of the estimates depends on the availability of data for each age-sex-year-location. Due to delays in data reporting, estimates for more recent years rely on additional data and trends from prior years. Thus, the GBD data for 2015 in the present study may in some instances reflect rates from earlier years as well. Thirdly, due to the cross-sectional study design, caution should be exercised in the attribution of causal relationships. Fourthly, the cutoff point at 60 years old may have different implications in different countries where life expectancies and cultural formulations of ‘old age’ differ so much. Last but not least, it should be borne in mind that the three independent indices were the latest available estimates from 1990 to 2015 (United Nations, 2017), which
+had data deficiencies and limitations in validity. However, to date, these data are the best available data for consistent comparisons across the nations worldwide.
+5. Conclusion
+Socioeconomic factors have important impacts on public health as well as late-life suicide prevention. The present study suggests that a set of negative transitions of socioeconomic status that the older adults frequently experience, such as loss of the domestic headship, dependently living with their descendants, and receiving no pension, may lead to higher elderly suicide rates. The present ecological findings suggest that strategies to enhance the socioeconomic status of older adults may be important to prevent suicides in later life both within and across countries on a grand scale. Therefore, priority ought to be given to facilitate efforts of older adults, families, and societies to reposition their roles in a household, enhance financial independence, and explore new meanings and expectations in the elderly people's later life.
+Q. Chang, et al
+Journal of Affective Disorders 256 (2019) 618-626
+Aged 60+ and Children under 15 (%)
+Fig. A.2. Scatter plot of percentages of households with both older adults aged 60+ and children under 15 and log-transformed suicide age ratios.
+Fig. A.3. Scatter plot of the proportions of the population above retirement age receiving a pension and log-transformed suicide age ratios.
+625
+Q. Chang, et al
+Journal of Affective Disorders 256 (2019) 618-626
+Table A.1
+Pearson's correlations among the variables used in the analyses.
+Pearson's Correlation	1. Suicide age	2. Percentages of headship, 3. Percentages of co-residence of the elderly with both 4. Percentages of the population above ratios	60 + ,%s	their children and grandchildren,%s	retirement age receiving a pension,%s
+1 2 3 4	1.00 -0.36*	1.00 0.37*	-0.48*	1.00 -0.51*	0.59*	-0.55*	1.00
+■ P <0.05.
+References
+Almeida, O.P., Draper, B., Snowdon, J., Lautenschlager, N.T., Pirkis, J., Byrne, G., et al., 2012. Factors associated with suicidal thoughts in a large community study of older adults. Br. J. Psychiatry 201 (6), 466-472.
+Blenkner, M., 1965. Social work and family relationships in later life with some thoughts on filial maturity. In: E., Shanas, G.F., Streib (Eds.), Social structure and the family: generational relations. Englewood Cliffs, NJ: Prentice Hall, pp. 117-130.
+Chan, S.M.S., Chiu, F.K.H., Lam, C.W.L., Wong, S.M.C., Conwell, Y., 2014. A multidimensional risk factor model for suicide attempts in later life. Neuropsychiatr. Dis. Treat. 10.
+Chang, Q., Chan, C.H., Yip, P.S., 2017. A meta-analytic review on social relationships and suicidal ideation among older adults. Soc. Sci. Med. 191, 65-76. https://doi.org/10. 1016/j.socscimed.2017.09.003.
+Conwell, Y., 2014. Suicide later in life: challenges and priorities for prevention. Am. J. Prev. Med. 47 (3), S244-S250.
+Duberstein, P.R., Conwell, Y., Conner, K.R., Eberly, S., Caine, E.D., 2004. Suicide at 50 years of age and older: perceived physical illness, family discord and financial strain. Psychol. Med. 34 (01), 137-146.
+Frieden, T.R., 2010. A framework for public health action: the health impact pyramid. Am. J. Public Health 100 (4), 590-595.
+Global Burden of Disease Collaborative Network, 2016. Global Burden of Disease Study 2016 (GBD 2016) Results 2017.
+Gordon, M., Whelan, B., Vaughan, R., 1981. Old age and loss of household headship: a national Irish study. J. Marriage Fam. 43, 741-747.
+Jahn, D.R., Cukrowicz, K.C., Linton, K., Prabhu, F., 2011. The mediating effect of perceived burdensomeness on the relation between depressive symptoms and suicide ideation in a community sample of older adults. Aging Ment. Health 15 (2), 214-220.
+Ju, Y.J., Han, K.T., Lee, H.J., Lee, J.E., Choi, J.W., Hyun, I.S., Park, E.C., 2017. Quality of life and national pension receipt after retirement among older adults. Geriatr.
+Gerontol. Int. 17 (8), 1205-1213.
+Kline, R.B., 2005. Principles and Practice of Structural Equation Modeling, 2nd ed. Guilford Press, New York, NY.
+Lloyd-Sherlock, P., Barrientos, A., Moller, V., Saboia, J., 2012. Pensions, poverty and wellbeing in later life: comparative research from South Africa and Brazil. J. Aging Stud. 26 (3), 243-252.
+Mattessich, P., Hill, R., 1987. Life cycle and family development. In: M.B., Sussman, S.K., Steinmetz (Eds.), Handbook of marriage and the family. Plenum Press, New York, NY, US, pp. 437-469.
+Park, S.M., Moon, S.S., 2016. Elderly Koreans who consider suicide: role of health care use and financial status. Psychiatry Res 244, 345-350. https://doi.org/10.1016/_j. psychres.2016.04.055.
+Phua, V.C., Kaufman, G., Park, K.S., 2001. Strategic adjustments of elderly Asian Americans: living arrangements and headship. J. Comp. Fam. Stud. 32, 263-281.
+Purcell, B., Heisel, M.J., Speice, J., Franus, N., Conwell, Y., Duberstein, P.R., 2012. Family connectedness moderates the association between living alone and suicide ideation in a clinical sample of adults 50 years and older. Am. J. Geriatr. Psychiatry 20 (8), 717-723.
+Rubenowitz, E., Waern, M., Wilhelmson, K., Allebeck, P., 2001. Life events and psychosocial factors in elderly suicides-a case-control study. Psychol. Med. 31 (07), 1193-1202.
+Seccombe, W., 1986. Patriarchy stabilized: the construction of the male breadwinner wage norm in nineteenth-century Britain. Soc. Hist. 11 (1), 53-76.
+Shah, A., Bhat, R., 2008. The relationship between elderly suicide rates and mental health funding, service provision and national policy: a cross-national study. Int.. Psychogeriatr. 20 (3), 605-615.
+Shah, A., Bhat, R., MacKenzie, S., Koen, C., 2008a. A cross-national study of the relationship between elderly suicide rates and life expectancy and markers of socioeconomic status and health care. Int. Psychogeriatr. 20 (2), 347-360.
+Shah, A., Padayatchi, M., Das, K., 2008b. The relationship between elderly suicide rates and elderly dependency ratios: a cross-national study using data from the WHO data bank. Int. Psychogeriatr. 20 (3), 596-604.
+Snowdon, J., Phillips, J., Zhong, B., Yamauchi, T., Chiu, H.F., Conwell, Y., 2017. Changes in age patterns of suicide in Australia, the United States, Japan and Hong Kong. J. Affect. Disord. 211, 12-19.
+United Nations, 2017. Department of Economic and Social Affairs, Population Division. Household Size and Composition Around the World 2017 -Data Booklet. http:// www.un.org/en/development/desa/population/publications/pdf/ageing/ household_size_and_composition_around_the_world_2017_data_booklet.pdf accessed August 10, 2018.
+United Nations Statistics Division, Department of Economic and Social Affairs (2017). Sustainable development goals indicators-proportion of population above retirement age receiving a pension. https://unstats.un.org/sdgs/indicators/database (accessed August 10, 2018).
+Van Orden, K.A., Wiktorsson, S., Duberstein, P., Berg, A.I., Fassberg, M.M., Waern, M., 2015. Reasons for attempted suicide in later life. Am. J. Geriatr. Psychiatry 23 (5), 536-544.
+Walsh, F., 2016. Families in later life: challenges, opportunities, and resilience. Expand. Fam. Life Cycle 261-277.
+Williams, L., Mehta, K., Lin, H.S., 1999. Intergenerational influence in Singapore and Taiwan: the role of the elderly in family decisions. J. Cross Cult. Gerontol. 14 (4), 291-322.
+Wood, D., 2003. Effect of child and family poverty on child health in the United States. Pediatrics 112 (Supplement 3), 707-711.
+World Health Organization (“WHO”), 2014. Preventing suicide: a Global Imperative. World Health Organization.
+Zhang, J., 2014. The gender ratio of Chinese suicide rates: an explanation in confucianism. Sex Roles 70 (3-4), 146-154.
+626

A systematic review of the predictions of the Interpersonal-Psychological Theory of Suicidal Behavior.txt

@@ -0,0 +1,86 @@
+1.	Introduction
+Suicide is a phenomenon that bears a significant public health impact worldwide. Each year it is estimated that approximately 800,000 people die by suicide, ranking it as the second leading cause of death in 15-29 year olds globally (WHO, 2014). Though preventable, suicidal thoughts and behaviors are complex phenomena influenced by several interacting factors, including personal, social, psychological, cultural, biological, and environmental (Goldston et al., 2009; King et al., 2001; Mann, 2003; O'Connor, 2011). As such, there is no singular underlying explanation as to why a person may attempt suicide, resulting in a highly contextual and varied picture of the barriers and facilitators to help seeking.
+Recently, the Interpersonal Psychological Theory of Suicide (IPTS) (Joiner, 2005; Van Orden et al., 2010) was developed with the aim of providing a theoretical model of suicide behavior. The theory consolidates a broad range of suicide risk factors, and provides testable predictions of who will develop desire for suicide (i.e., ideation), and from these, who will go on to attempt. As such, the theory holds much promise in regards to bettering our understanding of how certain suicide risk factors interact, and where prevention and intervention efforts may be best focused.
+According to the IPTS, suicidal desire is caused by the simultaneous presence of two proximal, causal risk factors: (1) thwarted belongingness and (2) perceived burdensomeness, and hopelessness (i.e., “this will never change”) about these states (Joiner, 2005; Van Orden et al., 2010). Thwarted belongingness refers to the experience that one is alienated from friends, family, or other valued social circles. It is said to comprise of two facets, loneliness (i.e., “I feel disconnected from others”) and the absence of reciprocal care (i.e., “I have no one to turn to and I don't support others”). It is viewed as a dynamic cognitive-affective state that is influenced by inter and intrapersonal factors such as experiencing family conflict, living alone, possessing few social supports, and being prone to interpret others'behavior as rejection (Van Orden et al., 2010). Perceived burdensomeness, on the other hand, refers to the view that one's existence is a burden on friends, family members, and/or society, and comprises of two facets, self-hate (i.e., “I hate myself") and feelings of liability (i.e., “my death is worth more than my life to others”). Like thwarted belongingness, perceived burdensomeness is conceptualised as a dynamic cognitive affect state, where risk factors such as homelessness, unemployment, physical illness, and feelings of low-self-esteem and being unwanted are said to contribute to its development (Van Orden et al., 2010). Though it is hypothesized that experiencing either perceived burdensomeness or thwarted belongingness alone will elicit passive suicidal ideation, it is their
+interaction coupled with the view that they are stable and unchanging (i.e., hopelessness) that will cause active suicidal desire.
+The development from active suicidal desire to suicidal intent is said to only result through the presence of an additional third construct: (3) acquired capability. Acquired capability refers to one's ability to overcome the inherent drive for self-preservation and engage in lethal self-injury (Joiner, 2005). This is hypothesized as being possible due to a lowered fear of death resulting from repeated exposure and habituation to physically painful and/or fear-inducing experiences, and an elevated tolerance of physical pain. It is viewed as a continuous construct that accumulates over time, with risk factors such as family history of suicide, previous suicide attempt, exposure to combat, and childhood maltreatment contributing to its development (Ribeiro &Joiner, 2009; Van Orden et al., 2010). Thus, individuals who have high levels of all three constructs, thwarted belongingness, perceived burdensomeness, and acquired capability, are said to be at most risk for lethal suicidal behavior, as they possess both the desire for and capability to attempt suicide. See Fig. 1.
+Since the development of the IPTS in 2005, a growing body of research has emerged testing different aspects of the theory across a range of populations. In 2009, an article on the current status and future directions of the IPTS stated that the theory has stood up to 20 direct empirical tests, with results generally substantiating the theory's main predictions (Ribeiro &Joiner, 2009). Since then, two systematic reviews
+on the IPTS have been published, one reporting on the role of perceived burdensomeness on suicide-related behavior within clinical samples (Hill & Pettit, 2014), and another examining support for the IPTS from studies published between 2002 and 2011 (Wachtel &Teismann, 2013).
+In their systematic review of 27 empirical studies testing the association between perceived burdensomeness and suicide ideation, suicide attempts, or suicide within clinical samples, Hill & Pettit (2014) found perceived burdensomeness to have statistically significant bivariate associations with both suicide ideation and past suicide attempt. Perceived burdensomeness was also found to be a predictor of suicidal ideation beyond the effects of other well established risk factors, and played a role as both moderator and mediator between suicide-related behaviors and other risk and protective factors. The authors noted that the majority of studies conducted focused on the relationship between perceived burdensomeness and suicide ideation, with results highlighting the role of perceived burdensomeness as a potential route for suicide intervention in clinical populations. A limitation of this review, however, is that it focused exclusively on the role of perceived burdensomeness within clinical samples, to the exclusion of the theory's more critical interaction predictions and applicability within other sample types.
+The other systematic review, by Wachtel & Teismann (2013), was more comprehensive, in that it reviewed the results of 29 studies (published between 2002 and 2011) that examined support for all three interpersonal risk factors in relation to suicide-related behaviors. The authors found perceived burdensomeness, thwarted belongingness, and acquired capability to be associated with different facets of suicidality, concluding that there was a lack of studies investigating the interrelation of the theory's constructs. However, this review was published solely in German with its findings being inaccessible to non-German readers in the field. Additionally, the review was limited to articles published up to 2011, with a considerable proliferation of IPTS studies since that time.
+Thus, the aim of the present review was to provide the first English systematic review of the full set of predictions of the IPTS across multiple populations. To assess the predictive power of the IPTS constructs independently of the contribution of other major suicide risk factors, the review focused specifically on the results of studies that adjusted for the presence of other IPTS variables (i.e., thwarted belongingness, perceived burdensomeness, and acquired capability) and/or mental health-related measures (e.g., depression, anxiety, hopelessness) to provide a rigorous test of these predictions. In doing so, the current review aims to identify whether empirical research supports the theory, and to highlight critical gaps in the evidence base by reviewing what populations and what aspects of the theory have been most tested and supported.
+2.	Methods
+On the 8th of July 2015, the Medline and PsycInfo databases were electronically searched for English-language, human, peer reviewed articles published from January 2005 up to July 2015 using the search terms: “Interpersonal psychological OR interpersonal-psychological OR Joiner* OR thwarted belong* OR perceived burden* OR acquired capability AND suicid*.” With limits imposed, 315 records were identified through database searching, and two additional articles from reference list searches. After duplicates were removed, 207 records were screened by the primary author for relevance to the systematic review. Sixty-three articles were excluded based on content (i.e., articles that were topically unrelated), and type of publication (i.e., review and scale development articles). The remaining 144 articles were considered for full-text review.
+Full-text articles were coded by the primary author (JM) and one of three independent reviewers (PJB, ALC, JH). Potential discrepancies in double coding were resolved by reaching a joint consensus between the two authors, or by assent of a third author where consensus could not be reached. Articles were included in the systematic review if they met all of the following criteria: (i) included a direct predictor measure of IPTS components (i.e., either thwarted belongingness, perceived
+burdensomeness, or acquired capability), (ii) included a direct outcome measure of suicidal thoughts or behaviors (i.e., either suicide ideation, attempt, or a composite measure), and (iii) reported on original, quantitative data. The exclusion criteria were as follows: (i) the study did not adjust for the presence of other IPTS variables/and or mental health-related measures, (ii) the article was not in English, (iii) no original data were reported, (iv) the study was a case-control design, (v) the study was qualitative, (vi) the study was not published after 2005, and (vii) the study was not published in a peer-reviewed journal. In the case where analysis was repeated on the same samples across articles, the most comprehensive and/or recent article was chosen for analysis, with the other being excluded.
+In total, 58 articles, comprising of 66 studies, adhering to the inclusion and exclusion criteria were included in the present review (see Fig. 2). Where sufficient data was available, effect size estimates were calculated based on formulas from “Practical Meta-analysis” by Lipsey & Wilson (2001). Odds Ratios were converted to Cohen's d (Cohen, 1988) for comparability between continuous and dichotomous outcomes using formulas outlined by Hasselblad & Hedges (1995). According to Cohen (1988), an effect size of 0.20 is considered small, 0.50 moderate, and 0.80 large. Where an effect size was not calculable, analyses of results relied on number of tests significant, using an alpha level of p < 0.05. Due to the heterogeneity of the studies (range of settings), the lack of effect size data, and the insufficiency of available data on interaction effects, we were unable to conduct a meta-analysis.
+3.	Results
+A total of 66 studies were identified that tested the IPTS constructs in relation to suicide ideation or attempt (See Appendix A for study characteristics). In order to present the results categorically under either suicide ideation or suicide attempt, composite measures such as “suicide risk”, “suicide potential”, “suicide proneness”, “suicidal symptoms,” “suicide behavior”, “future likelihood of behavior”,and “suicidality” were classified under suicide ideation, as they all encompassed a measure of suicide ideation. Eleven studies were found to include a composite measure, operationalised by the measurement scale used. The most commonly used composite measurement scale was the 4-item Suicidal Behaviours Questionnaire Revised (SBQ-R; Osman et al., 2001). The SBQ-R comprises of 4 items that measure suicidal ideation and attempt (“Have you ever thought about or attempted to kill yourself’); suicide ideation in the past year (“How often have you thought about killing yourself in the past year”); communication of intent (“Have you ever told someone that you were going to commit suicide, or that you might do it”); and likelihood of future attempts (“How likely is it that you attempt suicide someday”). Other composite measures used were similar in that they comprised of items or subscales that combined current suicidal ideation, suicide plans and preparation, and communication or threats of suicide.
+Across the 66 studies, 206 tests adjusted for the presence of other IPTS variables (i.e., thwarted belongingness, perceived burdensomeness, and acquired capability) and/or mental health-related measures (e.g., depression, anxiety, hopelessness). The largest number of tests was on the main effect of perceived burdensomeness on suicide ideation (33.4%), followed by thwarted belongingness on suicide ideation (22.6%). Tests on the main effect of acquired capability on suicide attempt (4.3%), and the two-way (5.8%) and three-way interactions (3.3%) proposed by the IPTS were scant in comparison. Table 1 summarises the results of the adjusted tests across the various IPTS constructs.
+3.1.	Suicide ideation
+3.1.1.	IPTS critical interaction effect: Thwarted belongingness and perceived burdensomeness on suicide ideation
+Twelve tests of the interaction between thwarted belongingness and perceived burdensomeness on suicide ideation were found, 8 (66.6%) of
+which were significant, and 4 (33.3%) non-significant. Significant study sample sizes ranged from 115 to 6133, with a mean of 1033.4, and median of 239. Non-significant study sample sizes ranged from 60 to 293, with a mean of 147, and median of 88. Only two studies reported an effect size, with effect sizes ranging from 0.46 to 0.61, with a mean of 0.53, considered a moderate effect.
+The interaction of thwarted belongingness and perceived burdensomeness was found to predict suicide ideation across hospital, primary care, school, and community populations. In one of the largest studies testing this interaction in a community sample, Christensen, Batterham, Mackinnon, Donker, & Soubelet (2014) found that after adjusting for gender, age, and the IPTS main effects, the combination of high levels of thwarted belongingness and perceived burdensomeness significantly contributed to suicide ideation in a cross-sectional sample of 1167 participants aged between 32 and 38 years old. This effect was also observed in studies that used proxy measures, such as social support (proxy for thwarted belongingness) and mattering (proxy for perceived burdensomeness). In their study on 815 young adults, Joiner et al. (2009) found that those low in both mattering and family social support reported the highest levels of suicidal ideation, controlling for the effects of six-month and lifetime histories of depression.
+Some studies showed that the interaction between thwarted belongingness and perceived burdensomeness on suicide ideation was only significant at high levels of perceived burdensomeness (Van Orden, Witte, Gordon, Bender, &Joiner, 2008(1)), high levels of thwarted belongingness (Kleiman, Riskind, et al., 2014; O'Keefe et al., 2014), or by age group (Christensen, Batterham, Soubelet, & Mackinnon, 2013). In their community-based study of 6133 participants aged between 28 to 72 years of age, Christensen et al. (2013) found that the interaction between thwarted belongingness and perceived burdensomeness was significant in a model including the main effects of thwarted belongingness, perceived burdensomeness, hopelessness, and the two-way and three-way interactions between the constructs only when the analyses was stratified by age, as opposed to when
+analyzed in the full sample. Here, the interaction between thwarted belongingness and perceived burdensomeness became non-significant in the full sample when the three-way interaction between thwarted belongingness, perceived burdensomeness, and hopelessness was included, suggesting that hopelessness plays an important role as a suicide risk factor. Studies reporting on this interaction effect were typically limited by cross-sectional designs and focus on samples with low base rates of suicidal ideation.
+3.1.2.	IPTS main effect: Thwarted belongingness and suicidal ideation
+Fifty-five tests were conducted on the effect of thwarted belongingness on suicide ideation. Of these, 22 (40%) were significant, and 33 (60%) were non-significant. Sample sizes among significant studies ranged from 38 to 6133, with a mean of 721.6, and median of 335. Non-significant study sample sizes ranged from 60 to 994, with a mean of 328.4, and median of 208. Only three studies reported an effect size, with effect sizes ranging from 0.49 to 0.74, with a median of 0.57, considered a moderate effect.
+Thwarted belongingness was found to predict suicide ideation, suicide risk, and suicidality across the mental health clinic, primary care, school, community, and detainee populations. One study conducted on a sample of 129 undergraduates found that thwarted belongingness contributed to 6% of the variance in suicide ideation (Davidson, Wingate, Rasmussen, & Slish, 2009). The effect of thwarted belongingness on suicide ideation was also reflected in studies using proxy measures, such as distress in interpersonal relations (Wilson, Kowal, Henderson, McWilliams, & Peloquin, 2013), detachment/estrangement (Davis, Witte, & Weathers, 2014), family belongingness (Ploskonka & Servaty-Seib, 2015), social support (Christensen et al., 2013), social relations (Joiner et al., 2009(1)), and interpersonal conflict and belongingness (You, Van Orden, & Conner, 2011). Some of the studies used proxy measures because they undertook secondary analysis of an existing dataset, and thus had to examine the IPTS interpersonal risk factors as post-hoc constructs. Others did so to compare different facets of thwarted belongingness. For instance, Ploskonka & Servaty-Seib (2015) explored the relationship between three domains of belongingness (family, peer, and academic institution) and suicide ideation in a sample of 249 undergraduates. They found that the only domain that significantly contributed to suicide ideation was family belongingness, suggesting that it may be one of the most important sources of belongingness.
+In regards to the non-significant tests, many studies that included measurements of both perceived burdensomeness and thwarted belongingness found that only perceived burdensomeness was a significant predictor of suicide ideation within hospital, mental health clinic, and school settings. In one undergraduate sample, the effect of thwarted belongingness on suicide ideation became non-significant after adjusting for depressive symptoms (Hill & Pettit, 2013). Additionally, in an online sample, thwarted belongingness was only significant after accounting for mediation by hopelessness (Kim & Yang, 2015).
+3.1.3.	IPTS main effect: Perceived burdensomeness and suicidal ideation
+Sixty-nine tests were conducted on the effect of perceived burdensomeness on suicide ideation. Of these, 57 (82.6%) were significant, and 12 (17.3%) were not significant. Significant study sample sizes ranged from 47 to 6133, with a mean of 419.6, and median of 245. Non-significant study sample sizes ranged from 38 to 815, with a mean of 286.8, and median of 205. Only six studies reported an effect size, with effect sizes ranging from 0.61 to 12.60, with a median of 1.42, considered a large effect.
+Perceived burdensomeness was found to predict suicide ideation and suicide risk across the hospital, mental health clinic, primary care, school, community, and online populations. Some of the studies indicated that perceived burdensomeness contributed substantial additional variance (36% and 41%) to suicide ideation, above and beyond the contribution of depressive symptoms and hopelessness (Davidson et al., 2009; Van Orden, Lynam, Hollar, & Joiner, 2006). However, these studies were limited by their cross-sectional design and use of primarily Caucasian samples. The effect of perceived burdensomeness on suicide ideation was also reflected in studies using proxy measures, such as whether people's lives would be positively impacted by one's death (Kanzler, Bryan, McGeary, & Morrow, 2012). For instance, in a sample of 103 patients experiencing chronic pain recruited from a mental health outpatient clinic Kanzler et al. (2012) found perceived burdensomeness to be the sole predictor of suicidal ideation, even after controlling for age, gender, depressive symptoms, and pain severity. However, this study was limited by its use of a non-validated single-item assessment for perceived burdensomeness and low base rate of suicidal ideation.
+Most of the studies that did not find a significant effect for perceived burdensomeness on suicide ideation also found no significant effects for other IPTS variables and covariates. For example, perceived burdensomeness alongside the three-way interaction of thwarted belongingness, perceived burdensomeness and hopelessness (Cukrowicz, Jahn, Graham, Poindexter, & Williams, 2013), and the three-way interaction of direct combat exposure, depression, PTSD, and hopelessness (Bryan,
+Ray-Sannerud, Morrow, & Etienne, 2013) did not significantly predict suicide ideation in the mental health clinic and primary care settings. These studies were limited by their cross-sectional design and lack of power to detect moderate effect sizes.
+3.1.4.	Acquired capability and suicide ideation
+There were 21 tests of the relationship between acquired capability and suicide ideation, with 12 found to be (57.1%) significant, and 9 (42.8%) non-significant. Significant study sample sizes ranged from 38 to 1208, with a mean of 324.4, and median of 168. Non-significant study sample sizes ranged from 55 to 1167, with a mean of 374.5, and median of 327.5. No effect size data was available. Acquired capability was found to predict suicide ideation, suicide risk, suicide potential, suicidal symptoms, and suicidality across the mental health clinic, school, and community populations (including military and detainee samples). It has been found to explain a significant portion of variance in suicidal ideation beyond the contribution of prior suicide attempt, stress, depression, and hopelessness in a military sample (Shelef, Levi-Belz, & Fruchter, 2014), and in one study using an undergraduate sample, contributed to 4% of the variance in suicide ideation (Davidson et al., 2009). In one of the few studies conducted on acquired capability conducted outside of the United States, Shelef et al. (2014) found that in a sample of 168 soldiers recruited from the Israel Defence Forces, suicide attempters were found to have significantly higher levels of dissociation and acquired capability compared to psychologically treated and healthy control groups, where depression and acquired capability were found to explain a significant portion of variance in suicide ideation.
+3.2.	Suicide attempt
+3.2.1.	IPTS full model: Three-way interaction of thwarted belongingness, perceived burdensomeness, and acquired capability on suicide attempt
+Seven tests of the interaction between thwarted belongingness, perceived burdensomeness, and acquired capability on suicide attempt were found, 3 (42.8%) of which were significant, and 4 (57.1%) nonsignificant. Significant study sample sizes ranged from 313 to 6133, with a mean of2312.6, and median of492. Non-significant study sample sizes ranged from 181 to 376, with a mean of 278.5. Only one study reported an effect size, that of 1.01, considered a large effect.
+In a cross-sectional study of 313 patients recruited from outpatient and inpatient facilities affiliated with a major U.S. Army medical centre (one of the first studies to assess the full model) the three-way interaction of thwarted belongingness, perceived burdensomeness, and lifetime number of suicide attempts (proxy for acquired capability) was found to predict recent suicide attempt and current suicide status controlling for the covariates of depression, hopelessness, and borderline personality disorder symptoms (Joiner et al., 2009(2)). It was noted that the strength of this effect was similar to other traditionally strong predictors such as family history of suicide. However, like many of the other studies, this study was limited by its cross sectional design and use of proxy measures to assess the IPTS constructs. For instance, lifetime number of suicide attempts was used as a proxy for acquired capability, neglecting other experiences of physically painful or fearinducing experiences which also contribute to the development of acquired capability.
+In another cross-sectional study conducted on 492 patients seeking treatment at a mental health clinic, Anestis & Joiner (2011) found that the three-way interaction predicted participant's lifetime number of suicide attempts, controlling for depression and participant sex. In one of the largest studies on the full model, the interaction between suicide ideation and acquired capability, but not the main effect of acquired capability, was found to predict suicide attempt in a community sample of 1167 adults (Christensen et al., 2014).
+A non-significant effect for the three-way interaction was observed in in-patient settings. For instance, Monteith, Menefee, Pettit, Leopoulos, & Vincent (2013) found that only the two-way interactions of perceived
+burdensomeness and acquired capability, and thwarted belongingness and acquired capability predicted suicide attempt cross-sectionally. Here, the only variable that was found to distinguish participants who reported no suicide attempts in the past from those who reported one suicide attempt was recent suicidal ideation.
+3.2.2.	IPTS main effect: Acquired capability and suicide attempt
+Nine tests were conducted on the effect of acquired capability on suicide attempt. Of these, 5 (55.5%) were significant, and 4 (44.4%) were non-significant. Significant study sample sizes ranged from 44 to 376, with a mean of 177.7, and median of 145.5. Non-significant study sample sizes ranged from 52 to 6133, with a mean of 1659.2, and median of 226. Only three studies reported an effect size, with effect sizes ranging from 0.51 to 1.09, with a median of 0.76, considered a moderate to large effect.
+Acquired capability was tested across the hospital, mental health clinic, community, and detainee populations. In one of the three longitudinal studies included in the review, baseline history of suicide attempt (a proxy for acquired capability) was found to predict suicide attempt at 12 months after hospitalisation in an in-patient, primarily Caucasian hospital sample (Czyz, Berona, & King, 2015). Another study conducted in the UK by Ireland & York (2012) found that in a sample of 191 detainees, engagement in a range of self-damaging behaviors (proxy for acquired capability) significantly predicted self-injurious behavior (proxy for suicide attempt) cross-sectionally.
+Of the non-significant studies, acquired capability was found to not be significantly associated with past suicide attempt, nor differentiate individuals in the suicidal behavior group from individuals in the non-suicidal behavior groups. One cross-sectional study conducted in a community sample, found that the main effect of acquired capability was only a significant predictor among the middle-aged (44-48) age group (Christensen et al., 2013).
+3.2.3.	Thwarted belongingness and suicide attempt
+Eleven tests were conducted on the effect of thwarted belongingness on suicide attempt. Of these, 4 (36.3%) were significant, and 7 (63.7%) non-significant. Significant study sample sizes ranged from 131 to 1167, with a mean of 704. Non-significant study sample sizes ranged from 181 to 6133, with a mean of 1185, and median of 376. Only three studies reported an effect size, with effect sizes ranging from 0.51 to 0.89, with a median of 0.54, considered a moderate effect.
+Thwarted belongingness was found to predict suicide attempt in studies set in hospital, mental health clinic, school, and community populations. In one cross-sectional study of 131 patients in treatment for opiate dependence, Conner, Britton, Sworts, & Joiner (2007) found that in a model including the effects of drug use severity, aggression, depression, hopelessness, thwarted belongingness, and perceived burdensomeness, only scores on belonging were associated with lower probability of having a history of attempted suicide. The effect of thwarted belongingness on suicide attempt was also reflected in studies using proxy measures such as belongingness (reverse proxy) (You et al., 2011) in a sample of 814 patients in a substance use treatment program.
+3.2.4.	Perceived burdensomeness and suicide attempt
+There were 13 tests of the relationship between perceived burdensomeness and suicide attempt, 3 (23%) significant, and 10 (76.9%) non-significant. Significant study sample sizes ranged from 215 to 1167, with an average of 554.2, and median of 417.5. Nonsignificant study sample sizes ranged from 52 to 6133, with an average of 1110.1, and median of 313. Only two studies reported an effect size, with effect sizes ranging from 0.52 to 1.70, with a median of 1.11, considered a large effect. The significant studies were conducted in mental health clinic and community populations. For instance, in a crosssectional study of 215 mental health out-patients, Hawkins et al. (2014) found that perceived burdensomeness was significantly
+associated with past suicide attempt, adjusting for depression, although effect sizes were small. In another cross-sectional study, perceived burdensomeness significantly predicted suicide plans/attempts, alongside thwarted belongingness and acquired capability, adjusting for gender, age, and the two-way interaction between thwarted belongingness and perceived burdensomeness in a sample of 1167 community-based participants (Christensen et al., 2014).
+3.3.	Alternative relationships
+3.3.1.	Mediation & moderation effects
+When undertaking the systematic review, the authors came across many studies that tested the effect of thwarted belongingness, perceived burdensomeness, and acquired capability as mediators across the hospital, primary care, mental health clinic, school, and community settings. The following factors were found to significantly mediate the relationship between constructs of the IPTS and suicidal ideation or behaviors:
+•	Thwarted belongingness: attachment security, agreeableness, parental displacement
+•	Perceived burdensomeness: anger, depression, post traumatic disorder symptoms, childhood emotional abuse, sexual orientation victimisation, sexual identity, body mass index, negative cognitive style, maladaptive perfectionism, basic need satisfaction
+•	Both thwarted belongingness and perceived burdensomeness: neuroticism, extraversion, forgiveness of self and others, family discrepancy, discrimination
+•	Acquired capability: over-exercise
+3.3.2.	Other two-way interactions
+Other two-way interactions among the IPTS risk factors were found to be significant in the literature. These were conducted across the hospital, mental health clinic, school, and community settings and included the interactions between thwarted belongingness and acquired capability in predicting suicidality, current risk for suicide, and suicide attempt; perceived burdensomeness with individuals' reproductive potential, health, and romantic relationship satisfaction in predicting suicide ideation; thwarted belongingness and optimism, and perceived burdensomeness and optimism in predicting suicide ideation; and acquired capability with agitation, and over-arousal on suicidality and suicidal symptoms.
+3.3.3.	Other three and four-way interactions
+Other significant three and four-way interactions among the IPTS risk factors were reported in the literature. These were conducted across the mental health clinic, school, and community settings and included: the three-way interaction of thwarted belongingness, perceived burdensomeness, and acquired capability on suicide ideation; the three-way interaction of age, combat exposure, and belongingness on suicide ideation; and the four-way interaction of thwarted belongingness, perceived burdensomeness, acquired capability and negative urgency on suicide attempt.
+4.	Discussion
+4.1.	Overview of the support for the Interpersonal Psychological Theory of Suicide's main predictions
+The current review aimed to systematically examine current evidence testing the effects of thwarted belongingness, perceived burdensomeness, and acquired capability on suicide ideation and attempt. Contrary to our expectations, the studies provided mixed support across the theory's main predictions. The main effect of perceived burdensomeness on suicide ideation was the most tested and supported relationship, with over three-quarters (82.6%) of the studies found to be significant across hospital, mental health clinic, primary care, school,
+community, and online populations. It was found to contribute a considerably larger amount of variance (36% to 41%) in suicide ideation compared to the contribution of thwarted belongingness, and in some cases overrode thwarted belongingness as the only significant effect. The main effect of thwarted belongingness on suicide ideation, on the other hand, though found to be significant across a range of settings, was tested less frequently than perceived burdensomeness, and was less supported, with over half (60%) the tests being non-significant due to the stronger effects of perceived burdensomeness and other covariates. In cases where it was found to be significant, thwarted belongingness seemed to contribute a smaller amount of variance in suicide ideation (6%) compared to perceived burdensomeness, and had a moderate median effect size, compared to the large median effect size reported for perceived burdensomeness. Contrary to the IPTS prediction that thwarted belongingness and perceived burdensomeness would be specific to suicide desire, approximately a third of the tests of thwarted belongingness on suicide attempt, and a quarter of perceived burdensomeness on attempt were significant, with a moderate median effect size for the former, and a large median effect size for the latter.
+In comparison to the main effects of perceived burdensomeness and thwarted belongingness, the main effect of acquired capability on suicide attempt was tested considerably less, with results providing only partial support. Just over half of the studies found a significant effect for acquired capability on suicide attempt across hospital, mental health clinic, and community populations, with a moderate to large median effect size. Additionally, contrary to the theory's predictions of acquired capability being specific to suicide attempt, half of the tests on acquired capability and suicide ideation were significant. However, it is important to note that this percentage may have been influenced by the reclassification of composite outcomes under suicide ideation.
+Studies testing the IPTS predictions regarding the interaction effects were scant in comparison to those testing the main effects of thwarted belongingness and perceived burdensomeness, and showed mixed results. Two thirds (66.6%) of the tests on the interaction between thwarted belongingness and perceived burdensomeness in predicting suicide ideation were found to be significant, with a moderate mean effect size. The specificity of their interaction contributing to suicide ideation only was supported by the literature. Moreover, only three (42.8%) out of the seven tests on the interaction between thwarted belongingness, perceived burdensomeness, and acquired capability on suicide attempt were significant, with over half of the tests on the full model found to be non-significant across the hospital, mental health clinic, and community populations. However, given that these nonsignificant effects were found in studies with samples sizes ranging from 181 to 376, these findings may be the product of too many low-powered studies to detect an effect for the full IPTS model, as a large effect size was found in one of the significant studies. Nevertheless, studies that did identify significant interaction effects tended to have similar sample sizes compared to those that did not find an effect.
+Overall, these results suggest that, at this point in time, the IPTS may not be as clearly defined nor supported as initially thought. Some of the conflicting findings across thwarted belongingness, acquired capability, and the two-way, and three-way interactions provoke a number of questions, including: (a) whether the interpersonal risk factors have different relationships on suicide ideation and attempt than stipulated by the theory (i.e., alternative interactions), (b) whether the measures commonly used across the studies adequately capture the constructs, (c) whether the theory is only accurate in predicting suicidal outcomes for a subset of suicidal individuals, and (d) whether there are other crucial variables that may help to better predict suicide ideation and attempt, which are not accounted for in the theory. In relation to (a), it may be that perceived burdensomeness is a more robust interpersonal risk factor for suicide ideation, in comparison to thwarted belongingness, which seems to also have associations with suicide attempt. However, in relation to (b), it may be the case that the measures used to assess
+thwarted belongingness, particularly the thwarted belongingness subscale on the Interpersonal Needs Questionnaire (INQ; Van Orden, Cukrowicz, Witte, & Joiner, 2012), do not fully capture the construct. This is an issue that has been raised by other researchers who have observed thwarted belongingness to have non-significant effects on suicide ideation when measured directly, as opposed to when measured using a proxy (Bryan, Clemans, & Hernandez, 2012). As research may privilege testing the relationship of perceived burdensomeness over thwarted belongingness, due to the conflicting findings of the latter, future research could look at validating broader proxy measures for thwarted belongingness, and examining what components may be missing from existing measures in order to balance out the evidence base.
+In relation to (c), whether the theory predicts suicidal outcomes for a subset of individuals, recent work using latent class analysis indicates that there are subclasses of individuals experiencing suicide ideation or attempt who display different symptom patterns and risk trajectories over time (Logan, Hall, & Karch, 2011). As suicidality is a heterogeneous outcome, it may be the case that the theory has more explanatory power for certain subsets of individuals. For example, in the case of acquired capability, studies that found a non-significant effect for the role of acquired capability on suicide attempt tended to have larger sample sizes (i.e., had greater statistical power) than those which found a significant effect. This suggests that other factors, such as sample characteristics and study setting may play a role in detecting a relationship. Future research testing the IPTS risk factors across different sub-sets of individuals would help to further specify the generalizability and explanatory strength of the IPTS predictions.
+In relation to (d), whether there are other crucial variables of interest not accounted for in the theory, studies have begun to examine the integration of the IPTS with other models of depression and suicide-related behavior, such as Hopelessness Theory (HT; Abramson, Metalsky, & Alloy, 1989) and the weakest link theory of suicidal ideation (Kleiman, Law & Anestis, 2014; Kleiman, Riskind, et al., 2014). Research is also being conducted on counterpart theories, such as the Integrated Motivational-Volitional Model of Suicidal Behavior (IMV; O'Connor, 2011), which builds upon the IPTS through the incorporation of thwarted belongingness, perceived burdensomeness, and acquired capability as moderators with other constructs, such as defeat and humiliation appraisals and entrapment; the work of which is essential to furthering theoretical endeavors within the field.
+In relation to clinical implications, these remain unclear due to the disparity in the number of studies focusing on the different IPTS constructs, and in particular, the lack of studies testing the critical interaction effects. Though work has been undertaken to outline how the IPTS can be used as a framework for identifying pernicious risk factors and tailoring assessments and interventions to address these factors (Stellrecht et al., 2006), further research elucidating the strength of the critical interaction predictions is needed to aid in the development of interventions that are able to specifically target the IPTS constructs to reduce suicidal ideation and suicide attempt. On a preliminary note, the results of the systematic review suggest that intervention-based efforts focused on identifying and decreasing levels of perceived burdensomeness in patients may be a more potent pathway for minimising risk of suicide-related behavior compared to that of thwarted belongingness. There is also evidence suggesting that interventions based on reducing levels of the three interpersonal risk factors may act to reduce different aspects of suicide-related behavior than initially stated by the IPTS, the pathways of which could be influenced by additional presenting risk or protective factors. Here, given the focus of the theory on identifying interpersonal risk factors, patients may feel more comfortable talking about feelings of belonging and burden with a clinician, as opposed to discussing suicidal behaviors. Focusing clinical discussions on risk factors, rather than suicidal behaviors, may help to increase engagement with clinical services and circumvent the potential stigma of discussing suicide (Calear, Batterham, & Christensen,
+2014; Gulliver, Griffiths, & Christensen, 2010). This interpersonal focus may also promote clinician empathy by highlighting the clinician's role as an important source of social support in the suicide risk factor framework, and could provide flow-on effects in improving the therapeutic alliance and patient outcomes (Baldwin, Wampold, & Imel, 2007; Lambert & Barley, 2001).
+4.2.	Strengths and limitations
+4.2.1.	Study strengths and limitations
+A major strength of the studies included in the current review was that they examined the IPTS across a large range of settings, and were not limited to testing the theory's main predictions. Many explored other interactions between the IPTS interpersonal risk factors and related constructs, contributing to our understanding of how distal risk factors influence suicide-related behavior through the IPTS proximal risk factors. However, many studies were limited by their cross-sectional design (63 out of 66), largely relying on retrospective reporting of suicidal ideation or behaviors), use of undergraduate samples with a low level of suicide ideation and attempt that were primarily Caucasian and female, use of self-report measures, evaluation of suicide ideation only (where suicide attempt was often underpowered), small sample sizes, and in some cases, small effect sizes for significant findings. Additionally, though the present review provides coverage of four additional years of publications on the IPTS, the same limitations regarding the lack of studies investigating the interrelation of the theory's constructs remain from previous systematic reviews. More high powered studies testing these critical interactions are needed to more comprehensively evaluate support for the theory.
+4.2.2.	Systematic review strengths and limitations
+To our knowledge, this is the first systematic review on the IPTS that examines the English-language literature on validation studies covering the full theory across multiple populations. By specifically analyzing the results of studies that adjusted for the presence of other IPTS variables and/or mental health-related measures, the review was able to robustly examine the strength of the theory's predictions. Additionally, the inclusion of studies using proxy measures of the IPTS variables
+highlighted alternative measurement pathways that may aid in better operationalisation of the IPTS constructs.
+Although comprehensive, a limitation of the present review was that it did not include articles that used non-standard terminology, nor articles published in languages other than English. The reclassification of suicide composite measures as suicide ideation, though helping to clarify the IPTS risk factor relationships with either suicide ideation or attempt, may also have inadvertently obscured more complex discussion of concurrent suicide-related behaviors. Here, it is important to note that the suicide composite measures that were reclassified as suicide ideation may not have been directly comparable, and should thus be interpreted with caution. Additionally, due to the lack of available data reported by the reviewed studies, the review relied primarily on summarising the results of significance tests, as opposed to effect sizes, limiting estimation of the magnitude of the relationships across studies. Moreover, when effect sizes were reported, Odds Ratios were converted to Cohen's d for comparability between continuous and dichotomous outcomes, which relied on the assumptions about the underlying distributions. Lastly, due to the comprehensiveness of the review, resulting in heterogeneity of studies, and the lack of reporting of effect size data, we were unable to conduct a meta-analysis.
+5.	Conclusions
+The review indicates that the relationship between perceived burdensomeness and thwarted belongingness on suicide ideation, and their interaction with acquired capability on suicide attempt appears to be less straightforward than originally stated in the IPTS. There is a need for more high powered studies examining the two-way and three-way interactions of the theory's constructs, use of longitudinal designs, and further tests of alternative interaction and mediation effects identified by some studies, highlighting potential for re-thinking the relationships as predicted by the IPTS. Future research focused on expanding the availability of valid measurement approaches for the interpersonal risk factors, and further elaborating upon their mixed relationships with suicide ideation and attempt across multiple populations is important to advance both theoretical and clinical progress in the field.

A-systematic-assessment-of-smartphone-tools-for-suicide-preventionPLoS-ONE.txt

@@ -0,0 +1,93 @@
+harmful content were also identified. Despite the number of apps available, and their varied purposes, there is a clear need to develop useful, pragmatic, and multifaceted mobile resources for this population. Clinicians should be wary in recommending apps, especially as potentially harmful content can be presented as helpful. Currently safety plan apps are the most comprehensive and evidence-informed, for example, “Safety Net” and “Mood-Tools—Depression Aid”.
+Introduction
+Rationale
+Suicide is a leading cause of death globally, particularly amongst young people [1]. Although immediate help during a crisis is critical, those who may be experiencing suicidal ideation or crisis experience barriers to help-seeking, such as not perceiving a need for professional help, lack of time, preference for informal help, access to and cost of services, and fear of stigma and disclosure [2]. With the increasing ubiquity of mobile phones, health applications (apps) have the potential to improve access and availability of evidence-based support to this group, as apps are low-cost, convenient, and discreet. Apps may be especially suited to deliver suicide prevention interventions with their ability to deliver support and intervention in situ and at the time of crisis. As suicide ideation and suicide risk change rapidly, access to high quality mobile resources may save lives.
+Consumers are rapidly embracing apps, proactively seeking apps to manage their personal health. Recent data suggest 85% of young people in the USA own a smartphone, three quarters of whom have used their device to access health information [3]. In a survey in the psychiatric out-patient setting, 69% of respondents and 80% of those aged 45 years or younger indicated a desire to use a mobile app to track their mental health [4].
+This consumer enthusiasm for apps to manage mental health has spurred the development of numerous apps for suicide prevention. Many of these offer digitised versions of tools and strategies common in mental health. However, to our knowledge the content of these apps has not been investigated. There is also an absence of efficacy data for apps related to suicide prevention, although the publication of designs [5], proof-of-concept results [6], and protocols for evaluation studies [7] are indicative of the future research direction. Assessment of content is vital, as the Android [8] and iOS [9] app stores do not have guidelines specifically related to the restriction of pro-suicidal content, or app content quality. Therefore, we currently do not know whether apps provide potential harmful content which promotes suicide or encourages suicidal behaviour [10], nor whether their content is consistent with clinical and population based policy.
+In previous work, Donker et al. found that mental health apps evaluated in randomised controlled trials [11] were not publicly available, while those with no research evidence were. Reviews of apps for other mental and physical disorders support this, reporting low adherence to clinical best practice, or the provision of unreliable, unsuitable tools [12-14]. In the present review, in view of the absence of published efficacy data for existing apps, we use the corpus of extant research trial evidence to address whether the features of publicly available health apps for suicide prevention are consistent with the research evidence. Suicide prevention strategies identified within the apps were ranked according to the strength of the evidence, as indicated by inclusion in the World Health Organization report by Scott and Guo [15], inclusion in other published systematic reviews [16-18], or inclusion in the Suicide Prevention Resource Centre Best Practices Registry [19].
+The findings from this review will inform clinicians and consumers of the content quality of suicide prevention apps currently available in the marketplace. An examination of the evidence base of app components will assist clinicians in recommending particular apps as part of adjunctive care and those promoting apps through the web to prioritise those most consistent with current evidence. Ultimately, this review will assist consumers to find apps consistent with best practice, and developers to consider the evidence-base of content during app design.
+Objectives
+Using descriptive methodology, the primary aim of this study was to compare evidence-based strategies undertaken for suicide prevention with the content of publicly available apps providing tools for suicide prevention.
+Methods
+Eligibility criteria
+Huckvale et al. highlighted the difference between informational and tool-based apps in a review of apps for asthma [14]. Tool-based apps are “active” or “interactive” as defined by De Jaegere et al. [20], specifically requiring active involvement from the user, or allowing users to interact with one another. Meanwhile, “passive” apps are those that solely present content, which could be in a variety of formats such as text or video, but require no user input or interaction beyond navigating through the content. In this review, only “active” or “interactive” apps were included due to the previously identified challenges in identifying the provenance of information contained within suicide prevention apps [21].
+In the current review, free and paid-for apps containing content related to suicide were included if they could be downloaded via the official Android and iOS stores. Apps were excluded if they: contained no “active” or “interactive” suicide prevention content; referred to suicide non-literally, for example in branding, or music titles; referred specifically to self-harm with non-suicidal intent; were related exclusively to depression, bipolar disorder, or other mental health conditions, unless suicidality was explicitly mentioned; or were not in English, or included character sets which did not display correctly.
+Information sources
+Apps were identified by searching the Australian Google Play store (Android) via its web interface, and the Australian iTunes store (iOS) using its search application programming interface (API). Results were limited by the search engines to a maximum of 250 (Android) or 200 (iOS) apps, and all of these search results were screened.
+Search
+A set of core search terms related to suicide was created: suicid*; parasuicid*; kill me/myself/ yourself; take my/your [own] life; self[-]harm*; DSH. To ensure consistency in the stemming of search terms across the two app stores, these keywords were manually expanded to create a comprehensive set of terms (see S2 Text). The terms related to deliberate self-harm (DSH) allowed the initial identification of apps where self-harming behaviours are with, without, or with unclear suicidal intention. Search terms on the Google Play store were surrounded by quotation marks to ensure the exact phrases were matched-this functionality was performed automatically by the iTunes search API. The unique identifier, title, description, and price of each app were retrieved from the app store, and apps which appeared in the results for multiple search terms were de-duplicated.
+App selection
+During the screening stage, two reviewers independently assessed the title and description of each app against the inclusion and exclusion criteria. The reason for each exclusion was recorded. Results of the screening were compared, and discrepancies were resolved by discussion until consensus was achieved. All apps that were identified as being eligible for inclusion were downloaded and installed on a Samsung Galaxy S4 mini (Android version 4.2) or iPhone 5s (iOS version 7.1) for full content review.
+Data collection process
+Following download, each app was opened and assessed independently by the two reviewers to confirm eligibility. The content and features of the apps were then independently reviewed for both harmful and suicide prevention content. The reviewers used a custom coding scheme (see Data Items section), and coded the interventional components directly into a database created for the review. Discrepancies were resolved by discussion until consensus was achieved.
+Data items
+Suicide prevention strategies are broad, ranging from population based activities (such as restricting access to means) to specific treatment interventions (such as dialectical behaviour therapy, DBT). Moreover, suicide is commonly experienced in a range of mental and physical health conditions. Apps may also contain potentially harmful content, and may be targeted at different user groups with different purposes (for example consumers, or clinicians). To accommodate these multiple facets, we developed four broad categories on which data were extracted for each app, as described below.
+App characteristics. The download cost for each app, whether free or paid-for, was recorded. Apps which contained any suicide prevention content were broadly categorised based on their primary function. The following distinct foci were determined: primarily or solely suicide prevention; depression (for example, an app which mentions suicide in the context of depression); deliberate self-harm; physical health or other mental health (for example, an app may include suicide as one of many health topics); or setting-based psychological or general support (for example, a university information app which mentions suicide prevention as part of its welfare programme). This categorisation allowed us to compare apps with a specific focus on suicide prevention and those in which suicide prevention is embedded within a wider context.
+Harmful content. Harmful information was coded using a synthesis of schemes used by Biddle et al. [10], Tam et al. [22], and Westerlund et al. [23] in their reviews of suicidal content on the internet. The harmful categories were: describing or facilitating access to lethal means; providing encouragement to people to end their life; portraying suicide in a fashionable or appealing manner; or an open category for other harmful content.
+App quality. In line with previous eHealth and suicide prevention reviews [10,20, 21,24] we rated broader app quality indicators. These quality-related features included: the type of developer or provider of the app (Q1); whether the provider name or contact details was explicitly stated within the app (Q2); whether references for the source of app content was included (Q3); whether a privacy policy was included within the app or app store description (Q4); whether the app could be protected with an account login, password or personal identification number (Q5); and whether bugs or reliability issues were apparent through use of the app (although the reviewers did not seek to exhaustively test the app for reliability; Q6).
+Suicide prevention tools. The spectrum of suicide prevention strategies is wide, spanning public health interventions associated with prevention in the general population, those targeted
+at higher risk groups, and mental health interventions for treatment and maintenance. In this review, strategies were coded based on the spectrum initially presented by Mrazek and Haggerty [25], and as reported by Scott and Guo in their report for the World Health Organization [15]. For convenience, we divided the strategies into the following five categories: public health, screening, accessing support, mental health/treatment strategies, and follow-up strategies.
+Public health strategies. Suicide prevention strategies include public health techniques targeting: information about legislation and policies restricting access to lethal means (S1); guidelines for media reporting of suicides (S2); and material about organisational, regional, or national suicide prevention strategies and policies (S3). These public health strategies were expected to be largely information-based and unlikely to be delivered through an interactive mobile app, however they were retained in the coding scheme for completeness.
+Screening strategies. This category consisted of strategies to improve screening and detection of suicidal risk, with apps targeted at physicians (S4); those in gatekeeper roles (S5); or for individuals to self-screen (S6).
+Accessing support strategies. Content to encourage or facilitate getting access to help included: requesting help and support from peers or family (S7); accessing help via a gatekeeper (S8); accessing non-crisis support services (S9); and access to crisis support and helplines (S10). For those apps which provided crisis support details, an additional data item was recorded to assess whether the crisis contacts were always visible within the app (for example, through a “get help now” button; S10a), as suggested by De Jaegere et al. [20].
+Mental health/treatment strategies. Mental health strategies focussed on preventing suicide either before or after an attempt. These strategies included: psychotherapy (S11); pharmacotherapy (S12); non-drug physical therapies (S13); the use of safety plans (S14); and postvention support for those bereaved by suicide (S15). As suicide safety plans contain multiple components and address a number of suicide prevention strategies, we performed a separate sub-analysis of the content of these apps. As with the public health category, we did not expect all of these strategies to be deliverable through an app (for example, drug or electroconvulsive therapy), however they were retained for completeness, and to record possible inclusion, for example, as part of a treatment diary.
+Follow-up strategies. Additional longer-term strategies focussed specifically on follow-up support after a suicide attempt. These strategies included: ongoing outreach and contact (S16); adherence management (S17); and peer support for those who have made a suicide attempt (S18).
+Evidence quality. After coding the apps’ suicide prevention strategies into the five categories described above, the quality of evidence for each strategy was rated from the extant literature for reducing suicide. As noted previously, we developed a coding scheme based upon the WHO report by Scott and Guo [15], and prevention strategies were coded as having strong evidence (E1) if they were consistent with findings in this report. Supplementary evidence was gathered from reviews by Mann et al. [16], Leitner et al. [17], and Shekelle et al. [18], and strategies with some degree of evidence from these reviews were coded as E2. Finally, if there was a lack of evidence within these reviews, a final check was made with the Suicide Prevention Resource Centre Best Practices Registry [19] to check if the strategy was at least consistent with expert ratings of best practice (E3). Otherwise, we coded a strategy as containing no relevant evidence (E4).
+Summary measures
+The number and percentage of apps satisfying each of the coding elements are reported. For each suicide prevention strategy contained within the apps, the coded evidence quality is also
+Fig 1. PRISMA flowchart showing the app search, screening, and review.
+doi:10.1371/journal.pone.0152285.g001
+reported. The number of strategies and apps containing recognised evidence is reported, along with the number of suicide prevention strategies per app.
+Results
+App selection
+The PRISMA flowchart for the review is shown in Fig 1. From the original 1271 search results, the descriptions of 856 unique apps were screened, and 123 apps were downloaded for review. Seventy-four apps were excluded following download, including 13 that did not contain any suicide prevention content. Eight of these 13 apps were games with the aim of killing or inflicting harm upon the character, including Russian Roulette. One of the excluded apps suggested risky behaviour, including deliberate self-harm or taking drugs, as an alternative to a suicide attempt. These suggestions contained disclaimers relating to the risk, possible legal consequences, and the lack of concordance with professional medical advice.
+Fifty-two of the reviewed apps were excluded as they did not provide interactive features to support suicide prevention. As expected, these informational apps described a wider range of suicide prevention strategies than could be incorporated into a tool-based app. In addition to components identified in the evidence review (described in the following sections), information was provided about: media reporting guidelines (one app; S2), national suicide prevention guidelines (one app; S3), gatekeeper screening for suicide (seven apps; S5), gatekeeper access to support (eight apps; S8), pharmacotherapy (four apps; S12), and non-drug physical therapies (one app; S13). None of these excluded apps contained information on adherence management (S17), or peer-support for those who attempt suicide (S18). The remaining 49 apps (Android: 20, iOS: 29) were included in the evidence review (S1 Dataset).
+App characteristics
+All 20 of the reviewed Android apps were free, while seven of the iOS apps required payment to download (four at AU$1.29, one at AU$2.49, one at AU$3.79, and one at AU$18.99).
+Approximately half of the apps downloaded and reviewed were suicide-specific (n = 24, 49.0%). Of those apps with a wider context, five apps had a focus on depression (10.2%), four on deliberate self-harm (8.2%, none with a specific focus on self-harm with suicidal intent), six on general health information (12.2%), and 10 on general support (20.4%).
+Harmful content. In addition to the potentially harmful content described for the excluded apps, two additional apps contained a list of means of instant death, although this was presented as suggestions for removing access to means. The risks of presenting lethal means have been discussed in previous work looking at the presentation of suicide on the internet [10,22].
+App quality. The majority of apps were developed (Q1) by academic/healthcare institutions, or commercial organisations (20 apps, 40.8%, each). Five apps (10.2%) were privately developed by individuals, and the type of provider was not clear for four apps (8.2%). Despite this range of developers, only 34 (69.4%; 13 suicide-specific apps, 54.2%) included contact details of the provider within the app (Q2).
+Although the apps were interactive, delivering a resource to users, just six (12.2%; one suicide-specific app, 4.2%) referenced the source of the content (Q3). While many apps prompted users to enter personal data, less than a half (19 apps, 38.8%; seven suicide-specific apps, 29.2%) included a privacy policy (Q4). Fewer still offered the option to protect the app with an account login, password, or personal identification number (eight apps, 16.3%; three suicidespecific apps, 12.5%; Q5). Nineteen apps (38.8%; eight suicide-specific apps, 33.3%) demonstrated obvious bugs or reliability issues during the content review (Q6).
+Suicide prevention tools. Table 1 shows the suicide prevention tools that were found in the 49 reviewed apps. The 24 apps which pertained primarily to suicide prevention are shown separately. Accessing peer support and safety plans were the most common features in the apps with a suicide prevention focus, as well as those with a broader focus. Follow-up strategies were least commonly identified within both groups of apps.
+Public health strategies. The only public health strategy identified within the apps related to means access restriction (S1), although this was at the level of an individual, rather than a population. Seven apps (six suicide-specific) allowed the user to identify lethal means that should be removed from their environment in the context of a safety plan (S14). No apps contained interactive features relating to media reporting guidelines (S2), or suicide prevention policies (S3).
+Screening strategies. Sixteen apps provided interactive screening tools for depression or suicidality: three for mental health professionals (S4), and 13 for individuals to self-screen (S6). No apps provided gatekeeper screening tools (S5).
+Two of the self-screening apps were suicide-specific, whereas none of the professional screening apps were primarily focussed on suicide. Nevertheless, one of the professional screening apps contained customised instruments to assist mental health professionals in screening for both suicidality and depression. The second app contained the Hamilton Rating Scale for Depression [26], and the third app included an extended version of the PHQ-9 [27] with additional questions related to suicide, paranoia, hallucinations, and mania.
+Of the 13 self-screening apps, two (both suicide-specific) contained a custom screening tool for detecting suicidality, including items on suicidal thoughts, social withdrawal, denying responsibilities, and other warning signs. The remaining self-screening apps focussed on depression: four presented an extended DSM scale, four used a modified or expanded PHQ-9 scale [27], and one which was designed specifically for post-natal depression, reproduced the Edinburgh post-natal depression survey [28]. The remaining two apps provided custom screening tools, one app specifically for depression based on a list of symptoms, and the other app provided multiple custom tools to screen for anxiety, depression, substance use, and suicide.
+Eight of the self-screening apps directed users to seek support from health or mental health professionals, or provided crisis support information when users screened high on depression or suicidality measures. Two apps additionally suggested users might be suitable for psychotherapy or antidepressant treatment, but did not directly suggest seeking help. Three further apps did not direct users to help-seeking options, however two were designed to be used as a checklist prior to an appointment with a health professional, and the final app used the screening results to populate a list of suggested tasks, including seeking help, in another section of the app.
+Accessing support strategies. Apps enabling access to support directed users to either peer support networks, non-crisis support, or crisis support services. None of the reviewed apps provided interactive access to gatekeeper services (S8). Of the apps providing access to help, 27 (16 suicide-specific) apps allowed users to access support from their peers, friends, or family (S7). Approximately half of the apps providing this function did so as part of a safety plan (n = 14; 13 suicide-specific). Eight of the non-safety-plan apps (three suicide-specific) allowed the user to nominate people as supporters and facilitated easy contact during a crisis. The remaining five apps (all non-suicide-specific) additionally allowed users to interact with one another within the app-users could share and discuss common experiences, and support others. This interaction and support took many forms with users interacting through photo sharing in one app, and by video sharing in another. One app also included a personal peer-to-peer support function where users could request support, or nominate times throughout the week when they were available to provide support. Of the five apps which offered peer interaction, four offered some degree of content moderation. Two of these apps specifically indicated that discussion of dangerous, unsafe, or violent acts would be removed; one other app included a function for alerting the service provider about worrying posts; and in one other app all content was centrally approved before being made publicly available.
+Twelve apps (10 suicide-specific) also provided access to non-crisis support services (S9), with 11 apps (10 suicide-specific) doing so within a safety plan. The remaining app was developed specifically for a clinical psychology practice and provided active access to the practice via a direct text message.
+A further 13 apps (10 suicide-specific) provided access to crisis support services (S10), seven of which (four suicide-specific) were independent of safety plans. Of these, five apps (four suicide-specific) contained interactive crisis support or helpline components. These features included: finding the nearest crisis centre based up location/GPS data (four apps, which were localised versions of the same app); and the ability for the user to enter their own crisis support contact (one app). Two apps also offered features for users to interact with other people. Both apps allowed users to initiate contact with an organisation-affiliated support service, either by text message or online chat capabilities. None of the 13 apps which provided access to crisis support services ensured that this access was visible at all times within the app (S10a).
+Mental health strategies. Safety planning (S14) was a prevalent mental health strategy contained with the reviewed apps, and is reported separately in a following section. Two apps also provided some degree of interactive psychotherapeutic content (S11), and another two provided postvention support for those bereaved by suicide (S15). No apps provided interactive features related to pharmacotherapy (S12) or physical therapies (S13).
+Both of the apps which delivered psychotherapy were based on cognitive therapy-one in the context of depression, and the other for deliberate self-harm. Both apps used thought challenging techniques: the depression app provided a tool that assisted users in identifying and challenging negative thoughts, while the DSH app provided a space to think about negative thoughts and to reframe them positively. The psychological content of both apps was selfguided, with no personalised input from a mental health professional. However, the DSH-ori-entated app did offer advice based on user-entered responses to motivations behind the current urge to harm, and suggested strategies or activities to distract users until the urge subsided.
+The two postvention apps provided an interactive plan for those bereaved by suicide. This was similar to a safety plan, in which the user completed sections of their plan after watching short videos that discussed different aspects of suicide bereavement. Elements of the plan included information, thoughts, and feelings associated with the event, and coping strategies and long term goals. The apps also highlighted that those bereaved by suicide may be at increased risk for suicide themselves, and encouraged seeking support if suicide ideation was present.
+Follow-up strategies. Finally, one app contained content specifically targeted at supporting someone who survived a suicide attempt. In addition to a safety plan, this app provided an appointment reminder function, which was coded as ongoing contact/outreach (S16). No apps addressed adherence management (S17) or peer support (S18) following a suicide attempt.
+Safety plans. As discussed above, many suicide prevention tools were incorporated into apps as part of a safety plan (S14). Safety plans were one of the most prevalent app features, with 14 apps (13 suicide-specific) enabling users to create a plan.
+Half of these safety plan apps (seven apps, six suicide-specific apps) allowed users to identify lethal means that they should remove from their environment in a crisis (S1). All safety plan apps allowed users to identify peer supporters who could be contacted in a crisis situation (S7). Additionally, 10 of the apps connected to the user’s contact/address book, enabling users to contact peers from within the app (all 10 apps), and facilitating the input of contacts by importing their details from the address book (five apps). In addition to peer support, 11 safety plan apps (ten suicide-specific) included details of non-crisis support services (S9) including psychiatrists, psychologists, mental health organisations and service providers, and general practitioners. All of these apps allowed users to enter their own contacts, and one app additionally assisted users in finding the nearest mental health resource based on location data obtained
+from the phone handset. Crisis support information (S10) was available within six of the apps (all suicide-specific) and was similar to non-crisis support, allowing users to input relevant crisis support line information. However, one app additionally prompted the user to call a national crisis support centre if pre-nominated warning signs were selected.
+Safety plans also contained components not otherwise coded in the description above. Just over half of the safety plan apps (eight apps, seven suicide specific) included a section for users to identify their individual warning signs for a crisis, with two apps (both suicide-specific), also allowing users to actively identify personal triggers. All 14 apps had a section for users to record coping strategies to ameliorate these factors, either allowing the user to enter their own strategies (12 apps), or allowing the user to listen to music or meditation tracks as a means of relaxation (two apps).
+In addition to specific means access restriction, an additional six apps featured sections for making the user’s environment generally safer and more comfortable, for example by not being alone. Four apps allowed users to nominate distracting places to go to in a crisis, such as social environments. Finally, seven apps encouraged users to record details associated with medium and longer term life planning, or reasons to live.
+Synthesis of results
+Ten distinct suicide prevention strategies were identified within the reviewed apps, one of which was associated with good evidence at level E1 (see Table 1). Five strategies were associated with secondary evidence (E2), three with concordance with best practice guidelines (E3), and one for which no relevant evidence could be identified (E4).
+Within the 49 apps, 94 individual interactive components were identified. Thirteen components (13.8%) were coded with E1 evidence, 46 (48.9%) were coded as E2, 34 (36.2%) were coded as E3, and one component (1.1%) had no relevant evidence (E4). Sixty individual components were identified in the 24 suicide-specific apps: 10 (16.7%) at E1,28 (46.7%) at E2, 21 (35.0%) at E3, and one (1.7%) at E4.
+Aggregating these results to the app-level, accounting for multiple components within each app and the evidence-based components within safety plans, 13 apps (26.5%; 10 suicide-specific, 41.7%) contained at least one element with some degree of evidence from the WHO report (E1). A further 28 apps (12 suicide-specific) contained at least one component with evidence from the literature reviews (E2), and the remaining eight interactive apps (two suicide-specific) contained at least one element which follows best-practice guidelines (E3). None of the reviewed apps were completely absent of components consistent with evidence or best practice.
+Excluding the safety plan apps, a mean of 1.1 (range: 1-2) identified suicide prevention strategies were found in each app. Safety plan apps, which inherently contain multiple components, contained a mean of 3.9 (range: 2-6) components.
+Within the 13 apps that contained the one identified high quality strategy coded as E1, the most comprehensive app was a safety plan app [29]. Overall, the most comprehensive app was also a safety plan app [30]. Both these apps were available for Android only. Outwith the safety plan apps, no apps which contained crisis contacts (E1) contained any other coded suicide prevention strategies.
+Discussion
+Summary of evidence
+This review has examined app store descriptions for 856 unique apps, the in-app content of 123 apps, and evidence for interactive suicide prevention strategies within 49 apps. Overall, providing access to crisis support services was the only strategy included within apps with E1 level evidence, with approximately a quarter of apps providing this feature. A further half of
+the apps were consistent with strategies identified in previous evidence reviews, and all apps contained elements consistent with at least best practice guidelines. A small number of apps with potentially harmful content were also identified during the review process.
+Twenty-four apps focussing specifically on suicide prevention were identified, all of which included features broadly concordant with the evidence base. This degree of concordance is higher than observed in reviews of other physical and mental health apps, and possibly reflects a higher degree of involvement from professional institutions in app development. For example, Nicholas et al. found that only 4% of apps for bipolar disorder were developed by institutions [12], and similarly Shen et al. reported that universities and institutions accounted for only 4.2% of developers of depression-related apps [13]. In contrast, institutions accounted for approximately half of developers of the reviewed suicide prevention apps. This potentially accounts for the difference in the proportion of apps that are evidence-informed between the current study and other mental health areas. This may also explain the reasonable provision of duty of care embedded within the relevant suicide prevention strategies. Most apps which offered self-screening tools alerted users towards help seeking options if risk of suicidality was detected, although the suggestion was not always direct or immediate. Apps which allowed users to interact with one another also contained content moderation, which is important considering the potential for sharing potentially harmful content.
+Despite the involvement of academic and healthcare institutions in their development, relatively few suicide prevention apps contained broader markers of app quality, such as referencing of source material. Indeed, a review by Aguirre et al. [21], specifically of suicide prevention apps, sought to review the evidence base of the content, however found it not possible due to the lack of information within apps indicating the provenance of the content. Apps also suffered from a lack of privacy policies, locking and protecting of apps, and reliability. These deficiencies may influence consumer and professional confidence in these apps.
+The components contained within the reviewed apps covered a broad range of suicide prevention content, with the strongest emphasis on safety planning and getting help in a crisis. However, the vast majority of apps only featured one interactive component. Given that the WHO report indicates good evidence for multifaceted suicide prevention strategies, the lack of comprehensive app-based support via the inclusion of numerous tool-based components represents a missed opportunity. Therefore, there is considerable scope for increasing the comprehensiveness of apps for suicide prevention. This could include targeted crisis support for individuals, including immediate access to support services through the app, and an active safety plan (despite the lack of clear evidence for this, it remains best practice and a prudent inclusion). Secondarily, non-crisis tools could include identifying suicide risk factors and triggers, and the delivery of psychological interventions.
+In addition to increasing the number of components offered, there is also a need for greater coverage of specific suicide prevention strategies that were missing or under-represented in the reviewed apps. While it may not be feasible to deliver large, public health strategies, pharmacotherapy, or physical therapy through an app, there is room for development of apps to deliver psychotherapy specifically for suicide prevention, improved physician-led screening for suicidality and wider risk factors, and for assertive follow-up following a suicide attempt. Although these strategies lacked the highest grade of evidence, there was some evidence in the literature.
+It is perhaps reassuring that there are a number of suicide prevention apps already publicly available to support individuals who may be in crisis, and that the interactive components generally follow best practice guidelines and strategies for which there is at least some degree of evidence. The identification of these good-quality apps, however, remains a challenge. Just under 90% of the apps identified in the app stores contained no suicide prevention strategies, and some contained potentially harmful content. With no regulation in the app marketplace, it
+currently falls on clinicians and consumers to delineate app quality. Therefore app developers have a challenge in not only creating suicide prevention apps with evidence informed content, but in dissemination strategies so that the app is identified and used by the target audience.
+Limitations
+There are a number of possible limitations with the current review. App stores allow publishers to restrict distribution to particular territories, and therefore not all apps may be available globally. As app store searches are localised to one particular country, it is possible that some suicide prevention apps were not found in the search of the Australian stores. However, an ad-hoc search of the term “suicide” on the American, Australian, British, Canadian, French, and German iOS app stores found 100% concordance, and no apps that were not available in each territory. This provides confidence in this review reflecting the global app market.
+Unlike searches of literature databases, app store search results provide a static snapshot of a dynamic marketplace. Apps can be updated at any time, removed entirely from the app stores, or disappear from the search results due to decreasing popularity. As an illustration, an ad-hoc search of the Australian iOS store at the time of writing found seven of the 149 apps originally identified through the “suicide” search term were no longer available. This illustrates a methodological challenge inherent in such reviews, as the results can only provide a snapshot into the current offering of available apps. This can also be a challenge for clinicians recommending an app, as there is no guarantee that an app will continue to be available.
+Mapping of app components to the evidence base has been, to some degree, hampered by the extent to which both apps and suicide prevention programmes adopt a multifaceted approach. Identifying which features or individual components are effective is therefore a challenge. This is further exacerbated by a relative paucity of good quality evidence for specific suicide prevention interventions that would be appropriate for inclusion into an app. As reported by Leitner et al.: “the research literature has adopted a ‘scattergun’ approach... The evidence base for any single form of intervention is therefore very limited” [17]. As there is a lack of a gold standard for effective suicide prevention interventions, we adopted the WHO report as that standard. Implementation of quality suicide prevention strategies, whether in app form or indeed wider policy implementation, could benefit from standard guidelines.
+Conclusions
+Despite a lack of evidence in the literature, there are a growing number of apps publicly available for suicide prevention. Many of these provide no interactive features, representing a lost opportunity to engage users in suicide prevention programmes. There are also a small number of apps which, to varying degrees, present potentially harmful content-of greatest concern is the encouragement to engage in risky behaviours such as drugs and deliberate self-harm to manage a crisis. Of those that do provide interactive prevention content, there was limited concordance with high quality evidence-based practice. However, all apps contained at least one component that was broadly consistent with either known evidence or best practice guidelines. While this represents a promising first step in harnessing apps to compliment suicide prevention awareness and strategies, there is a need for suicide-prevention apps to move beyond best practice into the delivery of genuine evidence-based practice with apps supported by empirical data on their effectiveness at reducing suicidal behaviours.
+Supporting Information
+S1 Dataset. Review Data.
+(CSV)

A-systematic-review-of-mortality-in-schizophrenia-Is-the-differential-mortality-gap-worsening-over-timeArchives-of-General-Psychiatry.txt

@@ -0,0 +1,62 @@
+IT IS NOW WIDELY ACKNOWLedged that schizophrenia contributes substantially to the global burden of disease.1,2 It is also well known that schizophrenia is associated with elevated suicide rates.3 Less widely appreciated is the fact that people with schizophrenia are at increased risk for premature death associated with comorbid somatic conditions.4 Apart from adverse effects related to medication, schizophrenia can trigger a cascade of socioeconomic and lifestyle factors that, in turn, can translate into adverse physical health outcomes. These comor-bid physical conditions contribute to increased mortality risks among people with schizophrenia.
+The association between severe mental illness and increased mortality rates has long been recognized.5 With respect to the
+group of disorders now labeled schizophrenia, increased mortality rates have been the object of scrutiny since the early 20th century.6-8 The quality of research on this topic has improved greatly in recent decades, with access to larger, better-characterized samples and the availability of high-quality mortality data for the general population. Access to these data allows the calculation of the standardized mortality ratio (SMR), which compares mortality in people with schizophrenia vs the general population. The SMRs are calculated by dividing the observed mortality rates in a given population (eg, the number of deaths in a group of individuals with schizophrenia) by the expected mortality rates in that same group as predicted by age- and sex-specific mortality rates for a standard population. Thus, an SMR of 2.0 would indicate that people
+with schizophrenia are twice as likely to die compared with the general population. The SMRs can be calculated for overall mortality (all-cause) or for more specific, widely used categories (eg, cancer, cardiovascular disease, endocrine disorders, or suicide).
+In recent years, several scholarly reviews4,9-11 have noted higher mortality in schizophrenia compared with the general population. A meta-analysis,4 based on 18 studies published between 1969 and 1996, reported an all-cause SMR for people with schizophrenia of 1.51. Another meta-analysis,11 based on 20 studies published between 1973 and 1995, reported a similar SMR for people with schizophrenia (1.57). Although these 2 systematic reviews agreed on the size of the pooled SMR associated with schizophrenia, there were discrepancies in the sex difference of overall mortality ratios. Brown4 found a small but significant male excess in the overall mortality ratio, whereas other studies12,13 reported either no sex dif-ference11 or higher mortality ratios in females compared with males.
+In collating data from different sites, systematic reviewers need to appreciate the structure of the underlying data. In light of the differing population age structure and disease profile among sites,1,14 we would expect substantial variation in mortality ratios among sites. For example, one would predict that SMRs for people with schizophrenia would differ between developed and developing nations, where the profiles of disease and the access to services vary markedly.
+Because of the increased focus on mental health care seen in many countries during the last few decades, one might predict that SMRs associated with disorders such as schizophrenia should be decreasing over time.15,16 However, several authors have suggested that SMRs in schizophrenia have been increasing during recent decades. For example, Osby et al17 found a linear increasing trend of mortality during 5-year periods from 1976 to 1995 among people with schizophrenia. The meta-analysis by Brown4 also reported significantly higher mortality in the 1980s compared with the 1970s. Deinstitutionalization may have influenced recent secular changes in mortality rates in schizophrenia. Although deinstitutionalization started in the 1950s, findings on its relationship to mortality have been inconsistent.10,11,18
+The aims of this study were to undertake a systematic review of mortality in schizophrenia and to examine a limited number of planned sensitivity analyses. In keeping with our previous systematic review of the in-cidence19 and prevalence20 of schizophrenia and considering that variability is to be expected in systematic reviews of SMRs,4,21 we sought to preserve the expected variation in the data rather than to focus only on pooled values derived from meta-analysis. Thus, for the main analyses, we present distributions of mortality estimates with measures of central tendency (eg, median or means) and quantiles (10% and 90% quantiles). On the basis of all-cause SMR, we predicted that the SMRs of males and females would not differ significantly. We also predicted that SMRs from the developed world would differ from those from the developing world (nondirec-tional hypothesis). We wished to explore the impact of study quality on SMRs. With the assumption that higher-
+quality studies would be more likely to identify deaths in schizophrenia, we predicted that SMRs derived from such studies would be higher compared with those from lower-quality studies. On the basis of previous systematic reviews and commentaries, we predicted that SMRs would increase over time.
+METHODS
+DATA SOURCES
+Most mortality studies are based on record linkage. People with schizophrenia are identified via psychiatric case registers and then subsequently linked to registers of cause of death. Some studies13,22 report mortality ratios based on hospital inpatient cohorts. Other studies23,24 have used community-based follow-up data for people with schizophrenia who are first identified through community surveys and then followed up for extended periods.
+IDENTIFICATION OF STUDIES
+Guidelines outlined by the Meta-analysis of Observational Studies in Epidemiology25 were followed to identify and collate mortality studies. The broad search string of (schizo* or psych*) and (mortality or outcome orfollow-up) was used in MEDLINE, PsychINFO, Web of Science, and Google Scholar to identify all research studies that investigated mortality in schizophrenia. Potentially relevant articles (in all languages) were accessed to review the full text. Citations from significant articles and review articles were scrutinized to locate additional relevant articles, book chapters, and conference papers. The Web of Science Cited Reference Search system was also used to locate relevant articles. Finally, letters or e-mails were sent to the senior authors of articles that met the inclusion criteria. These authors were provided with an interim list of included studies and asked to nominate missing studies.
+INCLUSION AND EXCLUSION RULES
+Studies were included if they met all the following criteria: (1) published and/or available between January 1, 1980, andJanu-ary 31, 2006, (2) reported deaths in people with schizophrenia as diagnosed by any criteria, (3) studied a population 15 years and older, (4) reported primary data on all-cause mortality and/or cause-specific mortality, and (5) reported SMRs and/or data on observed and expected deaths sufficient to calculate SMRs. Studies were excluded if they (1) involved people with a diagnosis other than schizophrenia (ie, studies that reported on broader categories of psychosis were excluded), (2) reported duplicate data, (3) reported SMRs solely attributable to suicide (this was the focus of a recent systematic review and meta-analysis3), and (4) reported mortality in subgroups of the population (eg, homeless people,26 twins,27 and those involved in clinical trials).
+DATA ABSTRACTION
+Once a study was included, data were extracted and entered into a 3-level, normalized database that included study-level variables (eg, authors, year of publication, and site), middlelevel variables (eg, age group, recruitment duration, casefinding method, and diagnostic criteria), and estimate-level variables (eg, general and specific-cause SMRs for all persons, males, or females). Two or more of the authors checked all data used in the analysis. When disagreements arose, these were re
+solved by consensus. If required, we contacted the original authors for clarification of issues. The full data set is available from the authors (www.qcmhr.uq.edu.au/epi).
+To assess the impact of overall quality of the distribution of SMRs, we devised a quality score. On the basis of operationalized criteria, this score rewarded studies that (1) used superior research design features (eg, more thorough case ascertainment, published diagnostic criteria, methods to confirm diagnosis, and longer periods of follow-up) and (2) provided comprehensive reporting of the study results (eg, provision of numerator, denominator, SMRs, details of subject attrition, and description of the completeness of the data source). Full details of the quality score used in this review are available from the authors (www.qcmhr.uq.edu.au/epi).
+In systematic reviews, it is important to avoid double counting of the index variable (deaths) by the same or different studies. Thus, a key feature of this review is the application of sequential filters to identify discrete mortality estimates. We applied a similar sorting algorithm to that used in our previous reviews of schizophrenia.19-20 Briefly, the mortality estimates were sorted into different causes of death. Study-level and middlelevel filters were applied to isolate data from multiple studies that overlapped in both time and place. The third filter was used to select 1 representative mortality estimate for inclusion in the cumulative distribution using the “most informative” rule. For example, if 1 study presented multiple overlapping ratios, the ratios based on the largest sample were preferred (ie, the widest age range was preferred over narrower age strata).
+The highest-order (and most reliable) category of death, allcause mortality, can be further subdivided according to rules such as those codified by the International Classification of Diseases, Ninth Revision (ICD-9) .28 Almost all included studies in this review were coded with the ICD-9. Although death can result from the combination of many different health problems, in circumstances in which several codes may be suitable, emphasis is given to the underlying cause of death. More specific causes of death can be allocated to categories according to organ systems (eg, cardiovascular or gastrointestinal) or nature of disease (eg, cancers are coded together). Apart from codes for these specific domains, studies occasionally report SMRs for middle-level categories such as all-unnatural (ICD-9 codes E800-E999) (which includes codes for suicide, accident, and homicide) and all-natural (ICD-9 codes 001-799; the remainder from all-cause when all-unnatural cause is excluded).
+The SMRs were extracted from the publications or calculated by dividing the sum of observed deaths by the sum of expected deaths (when sufficient data were available to calculate these). The distributions of SMRs were assessed in cumulative plots, with every SMR contributing to the distribution. The distribution of the data was assessed in rank order for SMRs (lowest to highest ranks) with the cumulative percentage of SMRs shown on the vertical axis. Key features of these distributions are presented (eg, median, mean, geometric mean, standard deviation, and quantiles at 10%, 25%, 50%, 75%, and 90%).
+For all-cause death, we were often able to extract data on case fatality rate (CFR). The CFR is calculated by dividing the number of deaths in people with schizophrenia during a certain period by the number of people with that disorder at the beginning of the period. An annualized CFR was derived to allow comparisons among studies of different durations.14
+In keeping with definitions from our previous systematic reviews of schizophrenia,20-29 we divided studies according to the per capita gross national product of the study site (based on 2004 data)30 and used a standard World Bank definition of country status31: (1) least developed countries,mean income of less than US $2995; (2) emerging economy countries,mean income between US $2995 and $9266; and (3) developed coun-tries,mean income of greater than US $9266.
+To assess secular trends, we used meta-regression to examine the relationship between the midpoint of the follow-up period and all-cause SMR for persons. Study quality scores were divided into tertiles, and the distribution of all-cause SMR for persons were compared according to these 3 levels.
+We performed statistical analyses for the test of significance between distributions of different SMRs. These analyses take into account (1) the need to control for within-study variation (estimates drawn from the same study tend to be more alike than SMRs drawn from different studies) and (2) the use of a log transformation to analyze distributions that are often positively skewed. Analyses were performed with SAS statistical software, version 9.2 (SAS Institute Inc, Cary, North Carolina).
+We also undertook a secondary analysis based on conventional meta-analytic techniques. Because SMRs are known to vary widely among sites because of population and disease frequency differences, we adopted a random-effects model to estimate a pooled SMR for all-cause mortality for persons.21 When necessary, 95% confidence intervals (CIs) were generated according to the formula detailed by Rothman and Greenland.21 Heterogeneity among the studies was tested using the Cochran heterogeneity statistic.32 Apart from the specific analyses related to sex differences, we restricted the analyses to persons to limit the number of planned comparisons. The funding source played no part in the design, analyses, writing, or submission of this study.
+RESULTS
+The electronic search identified 1726 articles, whereas manual reference checking identified an additional 26 references. We received responses from 16 authors, who provided an additional 11 references. Four articles from languages other than English were included after translation. Eleven studies33-43 were excluded because they completely overlapped with other included studies. Further details of the results of the search strategy and key features of the included studies are available from the authors (www.qcmhr.uq.edu.au/epi).
+The systematic review identified 37 studies9,12,13,18,22-24,44-73 that provided data on 561 SMRs for different causes of deaths drawn from 25 different countries: Australia (n = 2),59,68 Brazil (n=1),61 Bulgaria (n=1),53 Canada (n = 3),50,51,65 China (n = 1),53 Columbia (n = 1),53 Czech Republic (n=1),53 Denmark (n=2),63,64 Finland (n=3),18,22,23 France (n=2),46,48 Germany (n = 1),57 Hong Kong (n = 1),53 India (n = 2),12,53 Indonesia (n = 1),58 Ireland (n = 2)53,62 Israel (n= 1),73 Italy (n = 2),60,67 Japan (n=3),53,69,71 Norway (n=1),52 Russia (n=1),53 Sweden (n = 2),9,66 Taiwan (n = 1),49 the Netherlands (n = 1),13 the United Kingdom (n = 5),44,47,53,54,56 and the United States (n = 6).2445,53,55,70,72 The SMRs were based on an estimated total of 22 296 discrete deaths. Thirty-seven studies9,12,13,18,22-24,44-73 provided SMRs for all-cause mortality for either all persons, males, or females.
+Figure 1 shows the distribution for all-cause SMRs for all persons, males, and females. The median all-cause SMR for all persons (based on 38 SMRs) was 2.58, with 10% and 90% quantiles ranging from 1.18 to 5.76 (Table 1). In other words, people with schizophrenia had 2.5 times the risk of dying compared with the general population, and the central 80% of all SMRs varied over a 4-fold range. The median annualized all-cause CFR for all persons was
+95.4 per 10 000 population, with 10% and 90% quantiles ranging from 57.2 to 301.7 (5-fold range).
+The median all-cause SMR for males (3.02) was slightly higher than females (2.37); however, these 2 distributions were not statistically significantly different (F 1,18=0.0003; P =.99). For all persons, the median SMR for natural causes of death was 2.41, and the 10% and 90% quantiles ranged from 0.99 to 4.10 (Table 1). Elevated median SMRs were found in all of the specific causes of death apart from cerebrovascular diseases.
+Seven studies18,47,49,51,56,65,66 published data for the summary category of unnatural causes of death for all persons, males, or females. Table 1 gives the distributions of SMRs for unnatural causes of death. People with schizophrenia had 12 times the risk of dying of suicide compared with the general population (median SMR, 12.86).
+Twenty-two studies* were identified that contributed 28 SMRs for developed countries, 3 studies53,58,61 contributed 6 SMRs for emerging economy countries, and 1 study53 contributed 4 SMRs for least developed countries. When divided according to this criterion, the allcause SMR distributions were not significantly different (F2,34=0.30; P = .74); the median all-cause SMRs for least developed, emerging economy, and developed countries were 2.02, 2.19, and 2.79, respectively (Table 2).
+When the all-cause SMRs for all persons were divided into study quality score tertiles, no significant differences were found between SMR distributions (F2,24=0.61; P =.55). On the basis of follow-up periods, we identified 8 studies24,45,51,54,55,63,71,72 with SMRs from the 1970s, 10 studies47,53,57-60,65-67,73 with SMRs from the 1980s, and 7 studies22,46,48,53,61,62,68 with SMRs from the 1990s. Concerning secular change, meta-regression confirmed a significant positive association between the follow-up period midpoint year and all-cause SMR (slope coefficient, 0.06; 95% CI, 0.01-0.11; z = 2.21; P = .03). The median SMRs for the 1970s, 1980s, and 1990s were 1.84, 2.98, and 3.20, respectively. Concerning CFRs, the median rates per 10 000 population (all-cause mortality) were 162.2,
+95.4, and 108.3 for the 1970s, 1980s, and 1990s, respectively. The CFRs for the 3 decades were not statistically significantly different (F2,23 = 0.38; P =.38).
+The 38 studies that report all-cause SMRs for all persons are shown in a traditional forest plot with a pooled estimate based on a random-effects model in Figure 2. Using traditional meta-analytic techniques, we found that the pooled random-effects all-cause SMR (based on 37 SMRs with finite standard errors) for all persons was 2.50 (95% CI, 2.18-2.83). The Cochran Q test found a marginally acceptable level of heterogeneity ( Q36=50.72; P =.06). We undertook several post hoc analyses to explore potential sources of variation (eg, published vs unpublished diagnostic criteria, cohorts based on first-episode patients vs all patients, cohorts based on inpatient and/or outpatient samples, sites clustered according to World Health Organization regions, and SMRs attributable to suicide sorted by decade). However, none of the post hoc comparisons resulted in significantly different SMR distributions (data not shown).
+COMMENT
+People with schizophrenia have a substantially increased risk of death compared with the general population. Overall, people with schizophrenia have 2.5 times the risk of dying. This review was able to extract data from 37 studies that were conducted in 25 countries. As predicted, the distribution of all-cause SMRs showed prominent variability.
+Confirming the hypothesis that the relative mortality risk associated with schizophrenia is increasing, we found that SMRs have increased in a linear fashion during the 3 decades examined in this study. This finding is consistent with earlier studies.4,17 Considering that (1) CFRs for schizophrenia did not significantly differ among the decades and (2) age-standardized mortality rates are generally decreasing in most nations,74-76 these findings suggest that people with schizophrenia have not fully benefited from the improvements in health outcomes available to the general population. The SMRs are ratio measures and thus reflect differential mortality. If mortality rates in the general population decrease over time at a faster rate than those for people with schizophrenia, then SMRs for people with schizophrenia will increase over time. The evidence from the current study suggests that this differential mortality gap has widened over time.
+Mental health services have advanced in many parts of the world during the past few decades. Apart from a different mix of community-based care, the introduction of the second-generation antipsychotic medications in the early 1990s was initially found to be associated with better quality of life and reduced risk of relapse.77-79 More recent trials have questioned the clinical superiority of second-generation antipsychotic medication,80,81 and concern is now widespread about the adverse effects associated with these medications.82 In particular, compared with typical antipsychotics, several of the second-generation antipsychotics are more likely to cause weight gain and metabolic syndrome.83 Because the metabolic syndrome is associated with a 2-
+to 3-fold increase in cardiovascular mortality and a 2-fold increase in all-cause mortality,84 these adverse effects would be expected to contribute to even higher SMRs in the next few decades.85,86 Unfortunately, we are unable to explore the role of atypical medications as a contributing factor for the increasing SMRs associated with schizophrenia (eg, deaths related to clozapine-induced agranulocytosis or deaths related to atypical antipsychotic-induced weight gain). Adverse health outcomes associated with weight gain and/or metabolic syndrome (eg, myocardial infarction, cerebrovascular accidents, or cancer) may take decades to fully emerge. Thus, it seems likely that studies undertaken in the 1990s (ie, the most recent studies included in this review) would capture only a small fraction of the eventual burden of mortality associated with the adverse effect profile of the second-generation antipsychotic medications. In light of the rising secular trends in SMRs already identified by this review, the prospect of further increases in mortality risks for schizophrenia is alarming.
+In keeping with the findings of Harris and Barra-clough11 and Simpson,10 we found no significant sex difference in all-cause SMRs. Thus, although many well-documented sex differences exist in the epidemiological features of schizophrenia,19,87,88 the increased risk of mor
+tality associated with schizophrenia affects men and women equally.
+Of the specific-cause SMRs, suicide was associated with the highest estimate: 12 times greater than expected from the general population. In keeping with previous reviews, the SMRs associated with many different types of natural causes of death were elevated in people with schizophrenia. Curiously, the category neoplastic disorder had one of the lowest median SMRs (1.37). Although the median was still greater than 1, several record linkage studies89 have suggested that cancers may be significantly less prevalent in people with schizophrenia. The current review examines only mortality, and studies that examine morbidity would be better able to explore this issue.90
+We found no significant difference in SMRs among sites when sorted by economic status. However, this metaanalysis identified just 3 studies53,58,61 that provided discrete SMRs from the least developed and emerging economy countries; thus, caution should be exercised in the interpretation of this finding. Furthermore, a single derived variable was used to define economic status, which was applied at the ecological level.
+What factors have contributed to the differential mortality risk associated with schizophrenia? Many demo-
+graphic, clinical, political, and cultural factors mediate pathways and barriers to health care in general (eg, availability of services, stigma, and disease profiles).91 With respect to schizophrenia, the onset of the illness can result in a cascade of unhealthy lifestyle factors that elevate the risk of various somatic diseases and consequently increase the risk of death. People with schizophrenia are thought to be less inclined to seek health care, to consume less medical care, to engage in high-risk behaviors, and to be less compliant with their treat-ments.82,90,92 However, in addition to factors that operate on the pathway to care, schizophrenia and its
+associated comorbid somatic conditions may be downstream expressions of common genetic or environmental factors.92,93 For example, it is feasible that polymorphisms in genes may increase the susceptibility to both schizophrenia and diabetes94 or that de novo germline mutations across many generations could result in an increased risk of schizophrenia95 and a wide range of adverse health outcomes. Prenatal nutritional disruptions may equally affect brain development and general metabolic functioning.96,97 Although the current review cannot address these issues directly, the worsening SMRs associated with schizophrenia noted in recent decades
+suggest that this already disadvantaged group is not benefiting from the improved health of the community in an equitable fashion. A systematic approach to monitoring and treating the physical health needs of people with schizophrenia is clearly warranted.98
+Several important caveats to this review should be noted. Publication bias is always an issue in systematic reviews. We endeavored to address this by obtaining data from all available sources, including those from electronic databases, citations and authors, and publications in languages other than English. Factors such as the reliability of psychiatric diagnoses and admission practices (between sites and across time) could contribute to the variability identified in this systematic review. The reliability of the categorization of cause of death is also a cause for concern. With respect to specific-cause mortality, changes in the coding rules for the ICD-9 and be-tween-site variability in the application of these rules also need to be taken into account.99,100 However, these issues do not affect all-cause SMRs (which were used for the main analyses in this review). The current study found a higher all-cause SMR (median SMR, 2.58; pooled metaanalysis SMR, 2.50) compared with the 2 previous reviews, which reported all-cause SMRs of 1.514 and 1.57.* 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 The 2 previous systematic reviews were based on studies published before 199511 and 19964 compared with the current systematic review, which included 18 additional studies published after 1995.
+In conclusion, compared with the general population, people with schizophrenia have a 2- to 3-fold increased risk of dying. Suicide contributes to the increased mortality associated with schizophrenia; however, people with schizophrenia have increased mortality risks attributable to a wide range of somatic conditions. The increased mortality risk affects both sexes equally. Substantial variation occurs in all-cause SMRs among sites. In recent decades, the differential mortality gap associated with schizophrenia has been increasing. It is sobering to reflect on this paradox of schizophrenia treatment. As we become better at detecting and treating the core symptoms of schizophrenia, patients have worsening SMRs. Given the potential for an even greater disease burden as a result of the introduction of second-generation antipsychotic medications, research aimed at optimizing the physical health of people with schizophrenia needs to be undertaken with a sense of urgency.
+Submitted for Publication: November 4, 2006; final revision received January 16, 2007; accepted March 12, 2007.
+Correspondence: John McGrath, MD, PhD, FRANZCP, Queensland Centre for Mental Health Research, The Park Centre for Mental Health, Wacol Q4076, Australia (john _mcgrath@qcmhr.uq.edu.au).
+Author Contributions: Mr Saha has full access to all of the data in the study and takes responsibility for the integrity of the data.
+Financial Disclosure: None reported.
+Funding/Support: The Stanley Medical Research Institute supported this project.
+Additional Information: The following additional ma
+terial is available at www.qcmhr.uq.edu.au/epi: Figure S1: Flow Diagram (Selection Strategy) of Included Studies in the Mortality of Schizophrenia; Table S2: Quality Reporting Scale; Table S3: Summary Table of All-Cause Mortality and Standardized Mortality Ratio for Schizophrenia (1980-2006); Table S4: Standardized Mortality Ratios (SMRs) for Schizophrenia by Different Causes of Death for Males and Females; Table S5: Standardized Mortality Ratios for 3 Quality Score Tertiles of All-Cause Death; Table S6: Standardized Mortality Ratios for Schizophrenia of All-Cause Mortality for Various Post Hoc Analyses (for All Persons); and Microsoft Excel spreadsheet of the primary data for this systematic review, plus associated labels and formats.
+Additional Contributions: Dozens of researchers from around the world assisted in locating the data for this systematic review, and the staff of the Queensland Centre for Mental Health Research assisted in extracting the data and preparing the original manuscript.

ASSOCIATION OF RELIGIOSITY.txt

@@ -0,0 +1,47 @@
+The crude suicide rate for individuals aged 18-30 years has increased, and in 2015 the rate was 14.87 suicides per 100,000 people.1 Although the suicide rate among sexual minority young adults is unknown, suicide ideation and attempt occur more frequently among lesbian, gay, bisexual, and questioning (LGBQ or sexual minority) individuals than heterosexual people.2-7 Specifically, gay men, bisexual men, and lesbian women have a greater risk for suicide attempts than heterosexual adults.8 In general, religiosity is regarded as protective against suicidal thoughts and behaviors; yet, religion can be either a source of support or stress for LGBQ individuals.4,9-12 Consequently, it is
+From the 1Department of Psychiatry, University of Rochester Medical Center, Rochester, New York; 2Injury Control Research Center, West Virginia University, Morgantown, West Virginia; 3Center for Health Equity Research and Promotion, VA Pittsburgh Medical Center, Pittsburgh, Pennsylvania; 4Division of General Internal Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania; 5School of Social Work, University of Texas at Austin, Austin, Texas; 6Population Research Center, Austin, Texas; 7Counseling and Mental Health Center, University of Texas at Austin, Austin, Texas; and 8Department of Educational Psychology, College of Education, Austin, Texas
+Address correspondence to: John R. Blosnich, PhD, MPH, Injury Control Research Center, West Virginia University, 3606 Collins Ferry Road, Research Ridge, Suite 201, Morgantown WV 26508. E-mail: jblosni1@hsc.wvu.edu.
+0749-3797/$36.00
+https://doi.org/10.1016/j.amepre.2018.01.019
+unclear whether religiosity is protective against suicide ideation and attempt among LGBQ individuals.
+The mechanisms through which religiosity diminishes suicide risk are unclear.13-16 Particularly, moral objections (e.g., that suicide is an unforgiveable sin) may protect against suicidal behaviors,15 and religion may serve as a proxy for connections to community or social support.17 Thus, scholars have started differentiating among religious importance, seeking spiritual guidance, and religious attendance to determine whether these factors may serve as mechanisms of suicide prevention. Among the few longitudinal studies examining religion and suicidal behaviors, adults who attended religious worship at least once a month had lower odds of attempting suicide over the next 10 years compared with those who did not attend, and individuals who sought spiritual comfort had lower odds of suicide ideation for 10 years compared with people who were not spiritual.18 Similarly, there are inverse relationships between suicide ideation and religious attendance, religious well-being, and spiritual well-being among college students.16
+Religious groups’ perceptions vary about LGBQ individuals. High levels of individual religiousness are often associated with negative attitudes towards LGBQ peo-ple,19 and the link between internalized homonegativity and religiously based stigma is well documented, especially among non-affirming religious environments.9,10,12 Despite the fraught relations between religion and sexual orientation, many LGBQ individuals are religious, view religion as important, or have sought religious support after attempting suicide.9-11,20-22 Thus, the association between religion and suicidal behavior among LGBQ individuals have been mixed.
+Religiosity among LGBQ individuals and their parents have direct relationships to suicide attempts.12 For example, a study of LGB individuals in Austria with a religious affiliation had lower odds of attempting suicide than LGB adults who were not affiliated, and those who felt a greater sense of belongingness to their religious organization were less likely to endorse suicide ideation.9 Within a religiously diverse sample, the prevalence of passive (e.g., wish life would end) and active (e.g., considered suicide attempt) suicide ideation was greater among atheist/agnostic, Christian, non-religious, and other religiously affiliated LGB students than heterosexual students.4 Relatedly, LGB individuals who left their religion to resolve the conflict between their sexual orientation and religious affiliation had greater odds of attempting suicide than those with unresolved conflict.11
+LGBQ individuals may experience alienation and distress from religion or attempt to negotiate their intersecting religious and sexual identities.23,24 Consequently, the association between religiosity and suicidal
+behaviors is complicated for LGBQ individuals. Religion may not confer protection against suicidal behaviors or may be positively associated with suicidal thoughts and behaviors. Because few data sets contain information about sexual orientation, religiosity, and suicide ideation and attempt, there is a paucity of studies examining the association between religiosity and suicidal behavior among LGBQ individuals. The present hypothesis is that religiosity is negatively associated with suicide ideation and attempt among heterosexual individuals, but positively associated with suicide ideation and attempt among LGBQ individuals. Further, LGBQ status is associated with greater odds of suicide ideation and attempt among individuals endorsing greater religiosity.
+METHODS
+Study Sample
+Data are from the National Research Consortium of Counseling Center in Higher Education at the University of Texas at Austin. The Consortium conducts national studies on mental health among college students. In 2011, the Survey of Distress, Suicidality, Student Coping was conducted among probability-based samples from 74 higher education institutions and aggregated into a national data set made available to researchers. This survey was self-administered through a web-based questionnaire, the combined response rate between the undergraduate and graduate students was 26.3% and 26,292 students completed the survey. Because this study focused on young adulthood, the sample was restricted to individuals aged 18-30 years (n=21,247). Approximately 2.1% (n=550) were excluded for missing data about age, along with 4,495 individuals (17.1%) who were aged >30 years. Additional information about the methodology have been published.25 This study was approved by the University of Texas at Austin’s IRB.
+Measures
+The main outcome measures were suicide ideation in the past year, suicide attempt in the past year, and lifetime suicide attempt. Respondents were asked: Have you ever seriously considered attempting suicide at some point in your life? Individuals who answered yes were presented questions about suicidal behaviors. Those who answered no did not receive follow-up inquiries and were recoded as no on further suicide ideation and attempt questions. People who indicated lifetime suicide ideation were asked: During the past 12 months, have you seriously considered attempting suicide? Affirmative responses were defined as recent suicide ideation.
+People who indicated lifetime suicide ideation were asked: How many times in your life have you attempted suicide? Response options ranged from zero to five or more. All non-zero responses were defined as lifetime suicide attempt. Those who indicated a nonzero response were asked: How many of those attempts occurred in the past 12 months? Response options ranged from zero to five or more. All non-zero responses were defined as recent suicide attempt.
+Religiosity was operationalized as: How important are your religious or spiritual beliefs to your personal identity? Individuals responded on a Likert-type scale ranging from 1 (not at all important) to 5 (very important). Although the survey included a
+question about religious affiliation (e.g., Buddhist, Jewish), this variable was not included because: (1) it was not mutually exclusive, making it impossible to discern a dominant religion among those who endorsed multiple affiliations; and (2) despite overarching doctrine, many individuals seek alternative or affirming places of worship within an otherwise unwelcoming doctrine (e.g., a Baptist church that officiates same-sex marriages).26 The survey did not include measures of religious activities (e.g., frequency of worship).
+For sexual identity, respondents were asked: How would you describe your sexual orientation? Response options included: bisexual, gay or lesbian, heterosexual, questioning, and other. Among the 286 (1.3%) who indicated other, 268 supplied open responses. Although some of the other respondents could be included in the main sexual orientation groups (e.g., 59 respondents indicated straight), the majority of the responses (e.g., asexual, pansexual, queer) did not align with the existing categories. Thus, one respondent was recoded as lesbian/gay, 124 were recoded as heterosexual, and 143 were excluded from analyses. Because young people who are unsure of their sexual identity often report selfdirected violence, the questioning category was maintained.2
+Multivariable models were adjusted for sociodemographic characteristics. Gender identity was coded as female, male, or transgender and age was included as a continuous variable. Race and ethnicity was recoded into mutually exclusive groups of white, black, Asian, Hispanic, and other; for multivariable models, race/ ethnicity was dichotomized into white and racial/ethnic minority. International student status (yes/no) and partnership status were included. Respondents were asked: What is your current relationship status? (Select all that apply). The response options were: single and not currently dating, casually dating, in a steady dating relationship, partnered or married, separated or divorced, and widowed. Because respondents could indicate multiple categories, the variable was dichotomized into individuals who only endorsed single and not currently dating versus all other responses as a conservative definition of partnership status.
+Statistical Analysis
+Chi-square tests of independence were used to examine differences by sexual orientation in sociodemographic characteristics, religious importance, and prevalence of suicide ideation and attempt. Two sets of multivariable models were conducted to explore the relations of religious importance and sexual orientation with suicidal behavior. In the first set, recent suicide ideation was regressed on religious importance (as a continuous variable), stratified by sexual identity and adjusted for sociodemographic variables; this modeling was repeated for recent and lifetime suicide attempt. In the second set, recent suicide ideation was then regressed on sexual orientation, stratified by religious importance and adjusted for sociodemographic variables, and this analysis was repeated for recent and lifetime suicide attempt. Because of small cell sizes across the five Likert categories of importance of religion, this variable was recoded into a 3-category variable, 1-2 were merged (not important), 3 (moderately important), and 4-5 were combined (very important). Because of differences in self-directed violence among men and women, models were also stratified by gender identity.1,28 All estimates are reported as AORs with corresponding 95% CIs. Listwise deletion of all included dependent and independent variables was used for all analyses. All analyses were conducted using Stata/SE, version 12.
+RESULTS
+Among the analytic sample, 2.3% (n=485) individuals identified as lesbian/gay, 3.3% (n=696) identified as bisexual, and 1.1% (n=233) identified as questioning. All sociodemographics differed between sexual orientation groups (Table 1). Compared with heterosexuals, significantly greater proportions of sexual minorities reported that religion was not important. Notably, questioning individuals had the highest prevalence of recent suicide ideation (16.4%) and bisexual students had the highest prevalence of lifetime attempts (20.3%).
+In multivariable analyses stratified by sexual orientation, religious importance was not significantly associated with suicide ideation and attempt among bisexual individuals, but was significantly protective among heterosexual individuals (Table 2). Among lesbian/gay and questioning individuals, religious importance was associated with increased odds of recent suicide ideation, which seemed driven primarily by women. For example, among lesbian/gay individuals, increasing religious importance was associated with 38% increased odds of recent suicide ideation and for lesbian/gay women, specifically, was associated with 52% increased odds of recent suicide ideation. Additionally, for questioning individuals, increasing religious importance was also associated with increased odds of recent suicide attempt (AOR=2.78, 95% CI=1.14, 6.78). For lifetime suicide attempt, there was a negative association of religious importance among heterosexual women (AOR=0.90, 95% CI=0.85, 0.95), but weak positive associations for lesbian women (AOR=1.34, 95% CI=0.97, 1.85) and questioning men (AOR=1.53, 95% CI=0.98, 2.37).
+In multivariable analyses stratified by religious importance, there were mixed findings (Table 3). For example, lesbian/gay sexual orientation was not associated with greater odds of recent suicide ideation among individuals who reported religion was unimportant and moderately important; however, it was significantly associated with recent suicide ideation among individuals who reported religion as very important (Table 3). Conversely, bisexual and questioning sexual orientations were significantly associated with recent suicide ideation across all strata of religious importance; however, the patterns seemed to indicate the strongest effects were among the group for whom religion was very important.
+Because of the rarity of recent suicide attempt, some estimates in Table 3 could not be generated for all sexual orientations across all religious importance strata; those that were estimable were unstable and should be
+interpreted with caution. Among individuals who reported religion was unimportant, lesbian/gay sexual orientation was not associated with recent suicide
+attempt, but it was significant among the group for whom religion was very important. Bisexual sexual orientation was significantly associated with recent
+suicide attempt across all religious importance strata, but again the pattern of results suggested the strongest effects among the group for whom religion was very important.
+Lastly, LGBQ groups overall had greater odds of lifetime suicide attempt than heterosexual individuals (Table 3). In gender-stratified analyses, compared with heterosexual people, all sexual minority groups
+had greater odds of lifetime attempt, aside from gay men who viewed religion as very important, lesbian women who viewed religion as moderately important, and questioning men who viewed religion as unimportant.
+Data from Table 3 were also summarized in post-hoc analyses that estimated the adjusted prevalence of recent
+suicide ideation and lifetime suicide attempt in Appendix Figures 1 and 2 (available online). Results from recent suicide attempt could not be graphed because of suppression of some estimates across sexual orientation.
+Post-hoc analyses were also conducted to include a 3-item scale of social connectedness (i.e., how understood by others do you feel, how cared for by others do you feel, and how much do you feel that you can count on others). Each item had a 5-point Likert-type response that ranged from 1 (lower values) to 5 (greater values); reliability was acceptable (a=0.78). Overall, the adjustment of social connectedness did not change the pattern of findings for LGBQ respondents (Appendix Tables 1 and 2, available online); however, it did seem to account for many of the protective associations between religiosity and suicide ideation and attempt among heterosexuals (Appendix Table 1, available online).
+DISCUSSION
+The results partially supported the hypothesis that LGBQ groups do not experience the benefits of religiosity’s protective association against suicide ideation and attempt. Conversely, greater religious importance was significantly protective against both suicide ideation and attempt among heterosexuals in this sample. Moreover, these findings corroborate that gender differences in the association between religiosity and suicidal behaviors are minimal,16 suggesting that other factors, such as connectedness, may play a stronger role. For example, the change in results after adjusting for social connectedness suggests how religiosity confers protection against suicide ideation and attempt among heterosexuals; the lack of change among LGBQ individuals suggests other religious factors (e.g., antigay messaging and internalized homophobia) may be involved. In fact, among individuals with the strongest religiosity, LGBQ people seemed to have the greatest odds of suicide ideation and attempt; however, there was considerable heterogeneity among them.
+The positive associations among LGBQ groups are not surprising, given the relations between religion and LGBQ individuals, which are complicated at best and toxic at worst. For example, it is common knowledge that two of the world’s most common religions, Christianity and Islam, largely condemn homosexuality as a sin. However, significant positive associations were not consistent among all sexual minority groups. One potential explanation for this may be that different individual approaches are used to negotiate the intersection of sexual and religious identities. For example, some sexual minority individuals may withdraw from religion or seek affirming communities, whereas others
+may immerse themselves in religion.24,29 Thus, the heterogeneity in the results may speak to the potential nuanced ways that sexual minority communities navigate religious milieus.
+Moreover, religious-based conflict over sexual identity is often associated with conversion therapy (i.e., trying to change/suppress one’s sexual orientation),30 a practice that is denounced by the American Psychological Asso-ciation,31 among other professional organizations. This historic persecution of non-heterosexuality as well as more modern interpretation of scripture may have driven some religious institutions toward broadening their dogmatic practice to actively affirm and welcome LGBQ individuals.32 Yet, further research is needed about whether religions that are LGBQ-affirming may confer protective effects against suicidal behaviors among LGBQ individuals.
+More importantly, the present results have direct implications for mental health services, suicide prevention, and help-seeking efforts. Specifically, efforts that are built around faith-based organizations (FBOs) may not be appropriate for LGBQ individuals in distress, especially when religion may be a contributing element of distress for LGBQ individuals.33-37 This conundrum seems to have been overlooked in the suicide prevention literature, perhaps because of the paucity of quantitative studies, such as the present investigation. For example, the 2012 National Strategy for Suicide Prevention suggests FBOs be a major partner in suicide prevention and that, by promoting connectedness, FBOs may aid in suicide prevention.28 But to whom does this connectedness extend when ample literature suggests LGBQ people experience ostracism from their faith communities?24,38,39 Further, it is unclear whether enhanced training in suicide prevention for clergy and FBOs would serve LGBQ individuals if they perceived religious institutions as unwelcoming, thus undercutting help-seeking behaviors. Consequently, these findings, paired with the endorsement of FBOs as partners in suicide prevention, warrants research in several areas. For example, do LGBQ individuals actively avoid FBOs for mental health-related services? To what extent do FBOs serve LGBQ individuals, and do outcomes of service provision differ between heterosexual and LGBQ clients?
+Limitations
+There are a number of advantages to this study. Specifically, this large and diverse sample allowed investigating the differences among LGBQ individuals as well as rigorous adjustment for covariates (e.g., social connectedness). Despite the strengths of this research, there are several limitations. The data did not include
+questions about religious practice (e.g., religious attendance) or whether the associated religion espoused stigmatizing beliefs about sexual minorities; therefore, it was not possible to explore more nuanced relationships between religiosity and self-directed violence among LGBQ individuals. Although there is a religious affiliation variable, it was not included because it cannot account for the significant variation between denominations (e.g., Catholics, Protestants). With a sample from higher education institutions, these findings may not generalize to the broader population of LGBQ individuals. Although religious beliefs typically are instilled early in life by parents, because this is a cross-sectional analysis, it is not possible to ascertain any causal inferences between religiosity and suicidal behavior or if this relationship evolved over time. Although the response rate is similar for other large studies of young adults,40-42 the response rate was relatively low, which limits generalizability. The estimates for some outcomes, primarily recent suicide attempt, were unstable because of small sample size of the LGBQ groups. Finally, the measure of sexual identity did not allow for nuanced categorization (e.g., mostly heterosexual).
+CONCLUSIONS
+This study begins to address an important gap in the literature by exploring the association between religiosity, suicidal behaviors, and sexual orientation. Previous literature suggested that religiosity may protect against suicidal behaviors, yet those protective benefits were not observed among LGBQ individuals in this sample. In fact, the results suggested that, among people who regarded religion as very important, sexual minority status was more strongly associated with suicide ideation and attempt than the associations observed among people who regarded religion as unimportant. Suicide prevention efforts that partner with religious-based services should be aware of potential conflicts between religion and LGBQ individuals. Faith-based partners in public health suicide prevention and intervention services should be willing and equipped to assist all people who seek their services, regardless of sexual orientation. Moreover, this study opens a more general question about how and if faith-based public health partnerships benefit LGBQ populations.

Access to means of lethal overdose among psychiatric patients with co-morbid physical health problems Analysis of national suicide case series data from the United Kingdom.txt

@@ -0,0 +1,111 @@
+1.	Background
+Restricting access to lethal means is the suicide prevention intervention with the best evidence for effectiveness (Zalsman et al., 2016). Means restriction has most public health impact in relation to common, high-lethality suicide methods. After hanging, poisoning is the second commonest method of suicide in England, Scotland and Wales;
+accounting for 18% of male and 36% of female suicides in 2016 (Office for National Statistics, 2016a). Restricting means of overdose entails impeding access to the medication load available to at-risk persons to a level that, even if taken in one dose, will not pose serious harm (Hawton et al., 2013). Usually this involves adjusting the frequency (and therefore volume) of medication prescribed or available over-the-counter. The value of this approach is exemplified in the
+significant reduction in fatal paracetamol overdoses associated with UK legislation restricting pack size of over-the-counter analgesics (Barber and Miller, 2014). Where methods are not easily substituted by others, means restriction does not necessarily prompt means substitution (Sarchiapone et al., 2011). Indeed, the UK withdrawal of co-proxamol was associated with a significant reduction in deaths involving co-proxamol poisoning but no corresponding increase in deaths involving analgesics (Hawton et al., 2009). Physical disorders such as cancer (Henson et al., 2019; Ahmedani et al., 2017), osteoporotic fracture (Chang et al., 2018; Webb et al., 2012), back pain (Ahmedani et al., 2017), diabetes (Ahmedani et al., 2017; Webb et al., 2014), and heart disease (Ahmedani et al., 2017; Wu et al., 2018) are associated with an increased risk of suicide, and may provide affected individuals access to potentially lethal doses of prescribed medication Gorton et al., 2016). In a Swedish sample, 9% of patients diagnosed with diabetes who died from fatal poisoning had taken overdoses of diabetic drugs (Webb et al., 2014). For people with pain conditions, particularly chronic pain (Petrosky et al., 2018), opioids are a key target for means restriction, especially as the association of non-cancer pain and suicide risk is independent of psychiatric illness (Ilgen et al., 2013). In 2016 opioids accounted for 54% of all fatal drug poisonings (suicides and accidental overdoses) in England & Wales (ONS, 2016b). The most common opioid responsible was heroin and/or morphine (ONS, 2016b), although available data do not indicate what proportion involved ‘street’ opioids or those prescribed for chronic pain. With approximately 6000 people dying by suicide in the UK annually (ONS, 2016c), there is great interest among both clinicians and policymakers in the potential to restrict the volume of potentially lethal medication available to patients with physical illnesses. However, an improved understanding is needed regarding the role of access to these medications in pathways to suicide.
+Our research question was whether a greater proportion of psychiatric patients also diagnosed with physical illnesses who die by suicide poison themselves compared to individuals without physical comorbidities, and whether they are more likely to self-poison using medication prescribed to treat their physical health problems. We thereby aimed to explore the potential for means restriction interventions in a sub-group of psychiatric patients with co-morbid physical illnesses. Using national suicide case series data on psychiatric patients who died by suicide in England & Wales during 2004-2015, we aimed to describe the sociodemographic and clinical characteristics of psychiatric patients with a diagnosis of a co-morbid physical illness. We tested the hypotheses that:
+•	a greater proportion of deceased patients diagnosed with co-morbid physical illness fatally poisoned themselves than such patients without co-morbidity
+•	a greater proportion of deceased patients diagnosed with physical illness who fatally poisoned themselves overdosed on medication used for physical health problems versus such patients who died by intentional self-poisoning without co-morbidity
+•	among deceased patients with co-morbid physical illness who fatally poisoned themselves with medications prescribed to treat these conditions, a higher proportion had been prescribed the medication taken in overdose versus those without physical health disorders
+•	among deceased patients diagnosed with cancer, diabetes, and pain conditions the proportion who fatally self-poisoned using physical health medications was greater than among such patients diagnosed with other physical illnesses. These conditions have been linked with elevated suicide risk (Henson et al., 2019; Ahmedani et al., 2017; Webb et al., 2012), whilst also providing access to medications that are highly toxic in overdose (Gorton et al., 2016).
+2.	Methods
+2.1.	Study dataset
+Questionnaire data were collected as part of the National Confidential Inquiry into Suicide and Safety in Mental Health (Appleby et al., 1999). This database provides a national case series of patients under the care of mental health services who have died by suicide across the UK (i.e. England, Scotland, Wales and Northern Ireland). A detailed description of the National Confidential Inquiry's methodology is available elsewhere (Windfuhr et al., 2008). In brief, firstly, data on all deaths in England & Wales receiving a verdict of suicide or unnatural death of undetermined intent (‘open’ verdict) at coroner's inquest were received from the Office for National Statistics (ONS). Suicide research conducted in the UK conventionally includes open verdicts to avoid underestimating the number of suicide deaths (Linsley et al., 2001). Second, administrative contacts at NHS Trusts or Health Boards in the deceased person's district of residence identified whether contact had been made with secondary mental health services in the 12 months prior to death. Third, for those individuals with psychiatric contact, detailed data were collected via a questionnaire sent to the clinicians who had been responsible for that psychiatric patient's care. The questionnaire captured information on suicide method, demographic details, clinical characteristics, including any major physical illness at the time of death, aspects of care and treatment received.
+2.2.	Ethical approvals
+The National Confidential Inquiry has research ethics approval from North West - GM South REC (reference: ERP/96/136) and Section 251 Approval under the NHS Act 2006 (reference: PIAG 4-08(d)/2003), allowing collection of patient identifiable data for medical research.
+2.3.	Measures
+We defined physical health conditions on the basis of responses to the questionnaire item: “Did the patient have a major physical illness at the time of death? (include conditions even if well controlled by treatment)”. Free text responses to a further specifier permitted categorisation of conditions into those corresponding to International Classification of Diseases (ICD-10, 1992) categories (diseases of the musculoskeletal system, circulatory system, nervous system, digestive system, and endocrine disease). We used clinician-derived search terms to identify conditions with heterogeneous descriptors. For our sub-analyses we defined a specific diabetes category and overlapping categories for pain conditions and cancer.
+We categorised the substances used in self-poisoning on the basis of fixed-choice responses to the questionnaire item: “If self-poisoning, specify substance (if more than one substance, select most likely cause of death)”, to develop a categorical measure of whether or not these drugs are prescribed to treat physical illnesses. This was coded by a psychiatrist (AP), including free text responses to the “Other drug (please specify)” category. Categories within the physical illness treatment group were: opioids (morphine, codeine and methadone), paracetamol/ opioid compounds, other analgesics, insulin, cardiac medications, and other specified drugs for physical conditions (Box 1).
+We categorised the source of substances used in self-poisoning cases using fixed-choice responses to the relevant questionnaire item (prescribed for the patient; prescribed for someone else; not prescribed). For data collected from 2012, where opioids were reported in self-poisoning cases, further detail was available on whether these were prescribed for the patient for treatment of pain or for the treatment of drug misuse, prescribed for someone else, obtained illicitly, or obtained over-the-counter. We analysed drugs used for physical health conditions in cases of self-poisoning, irrespective of whether they had been prescribed for
+98%. We excluded 6% (1014 cases) with missing data for presence/ absence of physical co-morbidities, leaving a final dataset for analysis of 14,648 patients. Of these, 3525 (24%) had a recorded diagnosis of one or more co-morbid physical illness, most commonly diseases of the musculoskeletal (884, 25%); circulatory (822, 23%); endocrine (646, 18%); nervous (608, 17%); and digestive systems (580, 16%). Overall, 66% had a condition from a single major category of physical illness, 25% from two major categories, and 9% from three or more. Overlying these diagnostic categories, 16% (546 patients) had a pain condition and 9% had a cancer diagnosis.
+3.2.	Patient characteristics of those with a co-morbid physical illness
+The median age of psychiatric patients who died by suicide and had a co-morbid physical illness was 53 years (interquartile range (IQR) 43-64); significantly older than those without a physical health condition (median age 44, IQR 33-54; p < .001). Patients with a physical illness were more likely to be female, white, widowed, and to live alone than other patients (Table 1). They were less likely to be unemployed, unmarried or homeless. Whilst the proportions with a history of selfharm did not differ (around 68% in both groups), those with a physical health condition less often had a history of violence (19% v. 22%; p < .001) or of alcohol (42% v. 46%; p < .001) or drug misuse (27% v. 35%; p < .001). Patients with a physical illness were more likely than those without to have a primary psychiatric diagnosis of affective disorder, and less likely to have schizophrenia (including other delusional
+the patient or for someone else, or obtained illicitly.
+2.4. Statistical analysis
+Chi-square tests (with a 2-sided p-value threshold of < 0.05) were used to compare proportional distributions of sociodemographic and clinical characteristics between psychiatric patients with versus without diagnosed physical illness. We fitted logistic regression models to estimate the strength of these associations, with and without adjustment for age, gender, ethnicity, and presence of a primary drug dependence/ misuse disorder (which may itself be associated with chronic pain conditions). Odds ratios (ORs) and their 95% confidence intervals (CIs) were presented. Pairwise deletion was applied to address missing data; ie. if an item of information was unknown, the case was removed from the analyses of that variable. All analyses were conducted using Stata version 15.0 (StataCorp, 2017).
+2.5. Sensitivity analyses
+We conducted sensitivity analyses to assess robustness of findings when using a more stringent definition of medications that may have been prescribed to treat physical health conditions. This excluded drugs that can be used to treat psychiatric conditions (e.g. gabapentin and pregabalin for anxiety) or to address the side effects of psychotropics (e.g. metformin for antipsychotic-induced weight gain). We also repeated our analysis for data from 2012-2015 excluding opioids not prescribed for pain, medications prescribed for someone else, and nonprescribed medications (including over-the-counter paracetamol/ opioid compounds). In a post hoc sensitivity analysis we tested whether our findings were accounted for by the older age of those with co-morbid physical illness, and their greater prevalence of affective disorder.
+3. Results
+3.1. Descriptive statistics and prevalence of physical illnesses
+Between 1st January 2004 and 31st December 2015 inclusive, the National Confidential Inquiry was notified of 57,863 suicides in England & Wales (43,539 cases with a suicide verdict; 14,324 with an open verdict). Of these, 15,934 (28%) people had been in contact with secondary mental health services in the 12 months before they died. Questionnaires were returned on 15,662 patients, a response rate of
+disorders) or personality disorder (Table 1). They were less likely to have been a psychiatric in-patient at the time of death, to have recently (<3 months) been discharged from psychiatric in-patient care, or to have been under the care of a crisis resolution/home treatment team. They had more often attended their last contact with mental health services and were more likely to have been adherent with medication treatment compared with patients with mental illness alone. Nearly half (47%) had been in contact with services in the week before death, which was significantly fewer than for patients without a physical condition (51%; p < .001), with 68% exhibiting psychiatric symptoms at this appointment, proportionally more than other patients (63%; p < .001). However, these differences were unlikely to be clinically significant.
+3.3.	Method of suicide and substances used in self-poisoning
+A significantly greater proportion of psychiatric patients who had been diagnosed with a physical illness died by self-poisoning compared to those without physical co-morbidity (37% v. 20%, p < .001; AOR 2.47, 95% CI 2.26-2.70; Tables 2 & 3). The proportions who died by hanging/strangulation (33% v. 47%; p < .001), jumping/multiple injuries (12% v. 16%; p < .001), and gas inhalation (1% v. 3%; p < .001) (Table 2) were significantly lower in the physical co-morbidity group,
+although some of these differences were unlikely to be clinically significant.
+It was possible to classify the specific drugs used in cases of selfpoisoning in 3283 (86%) of cases; in 445 patients (12%) the data were missing and in 77 (2%) the substances were described as “multiple toxicity”. More patients with a physical illness were described as using multiple drugs in the overdose compared to those without a physical illness (37, 3% v. 37, 2%; p = .02), although this difference was unlikely to be clinically significant. Opioids were the most common type of drug used in all cases of self-poisoning, but particularly for those with a physical illness, nearly a third (30%) of whom died by opioid overdose compared with those with mental illness alone (22%; p < .001) (Table 2). Patients with physical illness were also more likely to use paracetamol/opioid compounds (11% v. 7%; p < .001) and insulin (4% v. 1%; p < .001) and less likely to use SSRIs/SNRIs (7% v. 11%; p < .001) or antipsychotics (8% v. 13%; p < .001) in self-poisoning.
+Overall, half (586, 50%) of psychiatric patients with a co-morbid physical illness who died by self-poisoning had used medications for a physical health disorder (i.e. opioids, paracetamol/opioid compounds, other analgesics, insulin, cardiac medications, and other specified drugs for physical conditions). This compared to a third (680, 34%) of those without a physical illness (p < .001) (AOR 2.10, 95% CI 1.80-2.46; Table 3). The majority (436; 64%) of this latter group had used opioids in overdose.
+3.4.	Sub-group analyses
+3.4.1.	Method of obtaining medication
+Details of how the substances were obtained were available for 2097 (55%) of the 3805 patients who died by self-poisoning, before excluding cases without data on physical illness. For the 1306 with a physical illness who died by overdose with any medication, data were available on how they obtained the drugs in 727 (56%), of whom 523 (72%) were prescribed those drugs, 20 (3%) used drugs prescribed for someone else, and 184 (25%) used unprescribed drugs.
+Focussing specifically on non-psychotropics, of the 586 patients with a comorbid physical illness who overdosed using a medication for a physical disorder, 246 (74% when excluding unknowns) had been prescribed this medication (Table 2). This compared to 102 (27%) of those without a documented physical illness who overdosed using prescribed non-psychoptropics (p < .001) (AOR 7.14, 95% CI 4.98-10.24; Table 3). The main substances used in the 102 cases without documented physical illness were opioids (52%), paracetamol/ opioid compounds (24%), other substances, e.g. propranolol (15%), and other analgesics (6%). A minority (14%) of this group had a diagnosis of drug dependence/misuse, and 44% had a history of drug misuse; these patients may have been prescribed opioids for drug misuse. Others may have been prescribed medication for a health condition not viewed by the clinician completing the questionnaire as a major physical illness.
+A quarter of patients with comorbid physical illness who overdosed using a physical health medication had not been prescribed it. A clinically significant minority had overdosed on prescription-only medications not prescribed for them. Insulin had been prescribed to 32 (86%) of the 37 patients with diabetes who self-poisoned using insulin. Of the 12 patients diagnosed with cardiovascular conditions who selfpoisoned using cardiac medications, these were prescribed for 8 (67%). However, it was more common for patients without a documented co-morbid physical health problem to have used medications for a physical disorder prescribed for someone else (13% v. 5%; p < .001) or obtained elsewhere (60% v. 21%; p < .001), presumably over-the-counter or illicitly.
+3.5.	Sub-analyses: patients with cancer, diabetes, and pain conditions
+When repeating the analysis for patients diagnosed with cancer
+compared to those with other physical illnesses, there was no association of death by self-poisoning with medication used for treating physical disorders (49% v. 50%; p = .973) (Table 3). Substances used most commonly in overdose in patients with cancer were: opiates (29%), paracetamol/opiate compounds (16%), and paracetamol (12%).
+Similarly, there was no association of death by self-poisoning with substances for physical disorders for patients with diabetes (54% v. 49%; p = .203) compared to those with other physical illnesses. Substances used most commonly in overdose among patients with diabetes were: insulin (21%), opiates (18%), and tricyclic antidepressants (11%).
+However, patients with a pain condition (the largest sub-group) were significantly more likely to overdose with drugs for non-psy-chiatric conditions compared to other patients with a physical condition (63% v. 46%; p < .001; AOR 2.12, 95% CI 1.56-2.88). The majority (67%) of substances used in overdose in patients with a pain condition were pain medications (opioids 46%; paracetamol/opiate compounds 12%; paracetamol 6%; any other pain meds 3%), whilst 9% used tricyclic antidepressants.
+3.6.	Sensitivity analyses
+The above associations remained unchanged in sensitivity analyses using a more stringent definition of drugs that could have been prescribed for treating physical health problems (Supplementary file). A post hoc sensitivity analysis to test whether our findings partly reflected the older age of those with co-morbid physical illness and their greater prevalence of affective illness, we found no association between older age or affective disorder and self-poisoning.
+4.	Discussion
+4.1.	Main findings
+We found that almost a quarter of psychiatric patients who died by suicide over the period 2004 to 2015 had a co-morbid physical health condition, and that over a third of this group died by self-poisoning. Our findings of an association between physical health problems and fatal overdose among psychiatric patients suggest that access to means is a key explanation. We found striking differences in the suicide methods used by psychiatric patients with and without physical health
+problems. Hanging (followed by overdose) was the most common method used by those with no physical co-morbidities; matching the national picture for psychiatric patients (NCISH, 2017), and the general population (ONS, 2016c). However, self-poisoning (followed by hanging) was the leading method used by patients with physical health problems, suggesting that overdose is the most accessible approach for this patient group if contemplating suicide. Restricting access to this method is more feasible than for hanging.
+The substances used in overdose by patients with a co-morbid physical health condition were more likely to be medications prescribed to treat physical health problems, and less likely to be psychotropics, even though these patients probably had access to both. Nearly half of those with a co-morbid physical health condition who died by selfpoisoning did so using a medication for such a condition. Of specific sub-groups, patients with pain conditions, for whom chronic pain is itself a risk factor for suicide (Racine, 2018) were most likely to overdose with drugs for physical disorders. This was likely due to a high proportion of this group using toxic pain medications in overdose. The tendency of patients with physical co-morbidities to overdose using non-psychotropics rather than psychotropics may relate to perceived lethality of non-psychotropics, potentially greater lethality of non-psychotropics, or to prescribers being more primed to consider overdose potential when issuing and monitoring potentially cardiotoxic psychotropic drugs (Hawton et al., 2010) than medications used for physical health problems. Whilst acknowledging the poor predictive value of suicide risk classification scales (Steeg et al., 2018), our findings suggest that needs-based assessments of psychiatric patients with physical health problems should focus on addressing modifiable risk factors such as reviewing the need for more toxic medications, particularly opioids (Ilgen et al., 2016), considering safer transdermal routes for opioid administration (Coplan et al., 2017), and addressing inadequately-treated pain (Yarborough et al., 2016). Guidelines on safe prescribing aim not to compromise on optimal pain management, but to reduce the potential for opioid addiction, diversion and fatalities (Volkow et al., 2019).
+4.2.	Findings in the context of other studies
+No other studies have sought to investigate this research question among psychiatric patients. More widely, a systematic review of studies investigating the association between non-psychotropic medications
+and attempted suicide found cardiovascular medications not to be associated with any increased risk, but concluded that associations with other medications remained inconclusive (Gorton et al., 2016). Separately, two studies of US veterans with non-cancer pain found an association between dose of opioids and risk of suicide (Ilgen et al., 2016), presumably with dose a marker of pain severity, but no clear excess risk of overdose in these patients over other methods (Ilgen et al., 2013).
+4.3.	Strengths and limitations
+We examined a national, comprehensive case series of all suicides amongst patients with recent contact with psychiatric services over a 12 year period. Consultants completing the questionnaire were unaware of the study's hypotheses, so it was unlikely that clinicians’ recall bias for overdose using physical health medications might explain our findings. Our categorisation of physical illnesses was systems-based but also acknowledged the overlapping categories of cancer and pain conditions. We adjusted our models for variables identified as potential con-founders a priori, such as drug dependence/misuse. Alternative explanations for associations identified are the under-identification of drug dependence/misuse, and the assumption that opioids used in overdose were obtained for a physical health problem rather than for abuse or intentional overdose. We had access to data on how medications were obtained for only 55% of the case series, but findings were similar in a sensitivity analysis confined to those who died from 2012-2015.
+The study's main limitation is that its use of survey data captured only those co-morbid physical health problems and overdose medications of which the responding consultant was aware. Under reporting of physical health problems is likely to have occurred where the patient was only briefly under their care (particularly in liaison settings), where clinical notes were unclear regarding physical health conditions or medications, or where the clinician did not judge the condition to be a ‘major physical illness’. This may have excluded conditions like acne that contribute significant clinical distress and for which medications prescribed to treat it have been linked with suicide risk (Sundstrom et al., 2010). Under reporting of specific physical health medications used in overdose is likely to have occurred where the completing clinician's response denoted multiple unspecified drugs. Over half of all general population drug poisoning deaths involve more than one drug and/or alcohol and the substance primarily responsible for the death is not identifiable (ONS, 2016b). We could also not be certain that medications used in overdose had been specifically issued to treat that patient's physical health problem, as opposed to being obtained specifically to attempt suicide. We did not have data specifying whether onset of physical illness had preceded psychiatric illness or vice versa, and it was possible in some cases that patients had been diagnosed with a physical health problem some time before their psychiatric illness commenced. This preceding physical illness may have also influenced some patients in their choice of self-poisoning agent.
+Detailed data on how opioids and paracetamol/opioid compounds were obtained were only available from 2012 onwards, but we addressed this in our sensitivity analysis. This extra analysis also ruled out the older age of those with co-morbid physical illness, and their greater prevalence of affective illness, as an explanation for our findings. Finally, by examining a national case series design, without living controls, we could estimate proportional contrasts between the groups but not incidence, or absolute/relative risks.
+4.4.	Clinical and policy implications
+These findings provide evidence to suggest that access to means of lethal overdose may contribute to suicide risk in psychiatric patients with physical co-morbidities, particularly those with chronic pain. Such patients would be more likely than other psychiatric patients to have supplies of prescribed non-psychotropics at home, particularly patients
+in chronic pain. Such availability creates the potential for suicide attempts with high lethality, particularly during a flare-up of a physical condition. All clinicians involved in the care of these patients should ensure careful prescribing for this patient group, with clear risk management. This could include regular reviews to check that indications remain, referral to pain clinics to consider transdermal opioid administration, and raised frequency of issuing pain medication prescriptions, although the latter may compromise patient convenience and therapeutic alliance. Assertive pain management is critical because inadequately-treated pain is itself a risk factor for suicide (Yarborough et al., 2016). Future research should seek to evaluate the effect of improved pain management pathways and prescribing guidelines on risk of overdose among psychiatric patients.
+Restricting access to non-prescribed medications has been partly addressed at the population level (Hawton et al., 2013, 2009) with a restriction on analgesic pack size, but there is also a role for community pharmacists in responding to customers trying to purchase over-the-counter analgesics above recommended limits (MHRA, 2014). A non-confrontational approach that responds to distress, and shows awareness of local service provision is more likely to be acceptable to patients. Our findings also suggest that access to medications prescribed for household members should be considered for psychiatric patients with or without physical illness. Carers have a role in safeguarding their own medications, as well as those of a psychiatric patient at risk.
+Finally, our findings show that opioids are a substance commonly used in lethal overdose among psychiatric patients, whether they have physical health problems (30%) or not (22%). Access to naloxone for carers and professionals, accompanied by training, is a high-risk intervention worth considering among some psychiatric patients (Ashrafioun et al., 2016). Qualitative work is needed with carers regarding their attitudes towards such a safeguarding role.
+5.	Conclusions
+Overdose, rather than hanging, is the leading method of suicide in the 24% of psychiatric patients who die by suicide and have co-morbid physical health problems; accounting for over a third of cases. In such patients, particularly for those in chronic pain, the medications used in overdose are more likely to be those for a physical health disorder; primarily opioids. Psychiatric patients with physical health co-mor-bidities therefore require careful needs-based risk assessment, with clinicians reducing access to the means of overdose where possible. Optimal care includes addressing inadequately-treated pain, reviewing the need for more toxic medications, considering transdermal routes, and involving carers in safeguarding household medications.
+A. Pitman, et al.
+Journal of Affective Disorders 257 (2019) 173-179
+Wales, the Scottish Government Health and Social Care Directorate, the Northern Ireland Department of Health, the States of Guernsey and the States of Jersey. AP is supported by the University College London Hospitals & National Institute for Health Research (UCLH NIHR) Biomedical Research Centre (BRC). None of these funders had any role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.
+Data availability
+The National Confidential Inquiry case series database is not publically available, but requests to conduct analyses in collaboration with the Centre for Mental Health and Safety team are granted, subject to internal peer review.
+Limitations of the study
+Use of survey data may have resulted in under-reporting of physical health problems and/or overdose medications.
+Supplementary materials
+Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jad.2019.06.027.
+References
+Ahmedani, B.K., Peterson, E.L., Hu, Y., Rossom, R.C., Lynch, F., Lu, C.Y., Waitzfelder, B.E., Owen-Smith, A.A., Hubley, S., Prabhakar, D., Williams, L.K., Zeld, N., Mutter, E., Beck, A., Tolsma, D., Simon, G.E., 2017. Major physical health conditions and risk of suicide. Am. J. Prev. Med. 53 (3), 308-315.
+Appleby, L., Shaw, J., Amos, T., McDonnell, R., Harris, C., McCann, K., Kiernan, K., Davies, S., Bickley, H., Parsons, R., 1999. Suicide within 12 months of contact with mental health services: national clinical survey. BMJ 318, 1235-1239. https://doi. org/10.1136/bmj.318.7193.1235.
+Ashrafioun, L., Gamble, S., Herrmann, M., Baciewicz, G., 2016. Evaluation of knowledge and confidence following opioid overdose prevention training: a comparison of types of training participants and naloxone administration methods. Subst. Abus 37 (1), 76-81. https://doi.org/10.1080/08897077.2015.1110550.
+Barber, C.W., Miller, M.J., 2014. Reducing a suicidal person's access to lethal means of suicide. Am. J. Prev. Med. 47 (3), S264-S272. https://doi.org/10.1016Zj.amepre. 2014.05.028.
+Chang, C-F., Lai, E-C., Yeh, M-K., 2018. Fractures and the increased risk of suicide: a population-based case-control study. Bone Joint J. 100-B, 780-786. https://doi.org/ 10.1302/0301-620X.100B6.BJJ-2017-1183.R2.
+Coplan, P.M., Sessler, N.E., Harikrishnan, V., Singh, R., Perkel, C., 2017. Comparison of abuse, suspected suicidal intent, and fatalities related to the 7-day buprenorphine transdermal patch versus other opioid analgesics in the national poison data system. Postgrad. Med. 129 (1), 55-61. https://doi.org/10.1080/00325481.2017.1269596.
+Gorton, H., Webb, R.T., Kapur, N., Ashcroft, D.M., 2016. Non-psychotropic medication and risk of suicide or attempted suicide: a systematic review. BMJ Open 6 (1), e009074. http://doi.org/10.1136/bmjopen-2015-009074.
+Hawton, K., Bergen, H., Simkin, S., Brock, A., Griffiths, C., Romeri, E., Smith, K.L., Kapur, N., Gunnell, D., 2009. Effect of withdrawal of co-proxamol on prescribing and deaths from drug poisoning in England and Wales: time series analysis. BMJ 338, b2270. https://doi.org/10.1136/bmj.b2270.
+Hawton, K., Bergen, H., Simkin, S., Cooper, J., Waters, K., Gunnell, D., Kapur, N., 2010. Toxicity of antidepressants: rates of suicide relative to prescribing and non-fatal overdose. Br. J. Psychiatry 196 (5), 354-358. https://doi.org/10.1192/%2Fbjp.bp. 109.070219.
+Hawton, K., Bergen, H., Simkin, S., Dodd, S., Pocock, P., Bernal, W., Gunnell, D., Kapur, N., 2013. Long term effect of reduced pack sizes of paracetamol on poisoning deaths and liver transplant activity in England and Wales: interrupted time series analyses. BMJ 346, f403. https://doi.org/10.1136/bmj.f403.
+Henson, K.E., Brock, R., Charnock, J., Wickramasinghe, B., Will, O., Pitman, A., 2019. Risk of suicide after cancer diagnosis in England. JAMA Psychiatry 76 (1), 51-60.
+https://doi:10.1001/jamapsychiatry.2018.3181.
+ICD-10 Classifications of Mental and Behavioural Disorder:, 1992. Clinical Descriptions and Diagnostic Guidelines. World Health Organisation, Geneva.
+Ilgen, M.A., Kleinberg, F., Ignacio, R.V., Bohnert, A.S., Valenstein, M., McCarthy, J.F., Blow, F.C., Katz, I.R., 2013. Noncancer pain conditions and risk of suicide. JAMA Psychiatry 70 (7), 692-697. https://doi:10.1001/jamapsychiatry.2013.908.
+Ilgen, M.A., Bohnert, A.S., Ganoczy, D., Bair, M.J., McCarthy, J.F., Blow, F.C., 2016. Opioid dose and risk of suicide. Pain 157 (5), 1079-1084. https://doi:10.1097/j.pain. 0000000000000484.
+Linsley, K.R., Schapira, K., Kelly, T.P., 2001. Open verdict v. suicide - importance to research. Br. J. Psych. 178, 465-468. https://doi.org/10.1192/bjp.178.5.465.
+MHRA, 2014. The Blue Guide; Advertising and Promotion of Medicines in the UK. https://www.gov.uk/government/publications/blue-guide-advertising-and-promoting-medicines.
+National Confidential Inquiry into Suicide and Homicide by People with Mental Illness, 2017. Annual Report: England, Northern Ireland, Scotland and Wales. University of Manchester October 2017.
+Office for National Statistics, 2016. Statistical Bulletin. Suicides in the UK: 2016 registrations. https://www.ons.gov.uk/peoplepopulationandcommunity/ birthsdeathsandmarriages/deaths/bulletins/suicidesintheunitedkingdom/ 2016registrations#suicide-methods.
+Office for National Statistics, 2016. Statistical bulletin: Deaths related to Drug Poisoning in England and Wales: 2016 Registrations. https://www.ons.gov.uk/ peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/ deathsrelatedtodrugpoisoninginenglandandwales/2016registrations.
+Office for National Statistics, 2016. Statistical Bulletin. Suicides in the UK: 2016 registrations. https://www.ons.gov.uk/peoplepopulationandcommunity/ birthsdeathsandmarriages/deaths/bulletins/suicidesintheunitedkingdom/ 2016registrations.
+Petrosky, E., Harpaz, R., Fowler, K.A., Bohm, M.K., Helmick, C.G., Yuan, K., Betz, C.J., 2018. Chronic pain among suicide decedents, 2003 to 2014: findings from the national violent death reporting system. Ann. Intern. Med. 169 (7), 448-455. https:// doi:10.7326/M18-0830.
+Racine, M., 2018. Chronic pain and suicide risk: a comprehensive review. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 87, 269-280. https://doi.org/10.1016/j.pnpbp. 2017.08.020.
+Sarchiapone, M., Mandelli, L., Iosue, M., Andrisano, C., Roy, A., 2011. Controlling access to suicide means. Int. J. Environ. Res. Public Health 8 (12), 4550-4562. https:// doi:10.3390/ijerph8124550.
+StataCorp, 2017. Stata Statistical Software: Release 15.
+Steeg, S., Quinlivan, L., Nowland, R., Carroll, R., Casey, D., Clements, C., Cooper, J., Davies, L., Knipe, D., Ness, J., O'Connor, R.C., Hawton, K., Gunnell, D., Kapur, N., 2018. Accuracy of risk scales for predicting repeat self-harm and suicide: a multicentre, population-level cohort study using routine clinical data. BMC Psychiatry 18, 113. https://doi.org/10.1186/s12888-018-1693-z.
+Sundstrom, A., Alfredsson, L., Sjolin-Forsberg, G., Gerdén, B., Bergman, U., Jokinen, J., 2010. Association of suicide attempts with acne and treatment with isotretinoin: retrospective Swedish cohort study. BMJ 341, c5812. https://doi.org/10.1136/bmj. c5812.
+Volkow, N.D., Jones, E.B., Einstein, E.B., Wargo, E.M., 2019. Prevention and treatment of opioid misuse and addiction: a review. JAMA Psychiatry 76 (2), 208-216. https:// jamanetwork.com/journals/jamapsychiatry/article-abstract/2716982.
+Webb, R.T., Kontopantelis, E., Doran, T., Qin, P., Creed, F., Kapur, N., 2012. Suicide risk in primary care patients with major physical diseases: a case-control study. Arch. Gen. Psychiatry 69 (3), 256-264. https://doi:10.1001/archgenpsychiatry.2011.1561.
+Webb, R.T., Lichtenstein, P., Kapur, N., Ludvigsson, J., Runeson, B., 2014. Unnatural deaths in a national cohort of people diagnosed with diabetes. Diabetes Care 37 (8), 2276-2283. https://doi:10.2337/dc14-0005.
+Windfuhr, K., While, D., Hunt, I.M., Turnbull, P., Lowe, R., Burns, J., Swinson, N., Shaw, J., Appleby, L., Kapur, N., 2008. Suicide in juveniles and adolescents in the United Kingdom. J. Child Psychol. Psychiatry 49 (11), 1155-1165. https://doi.org/10.1111/ j.1469-7610.2008.01938.x.
+Wu, V.C-C., Chang, S-H., Kuo, C-F., Liu, Chen, S-W., Yeh, Y-H., Luo, S-F., See, L-C., 2018. Suicide death rates in patients with cardiovascular diseases - A 15-year nationwide cohort study in Taiwan. J. Aff. Disord. 238, 187-193. https://doi:10.1016/j.jad. 2018.05.046.
+Yarborough, B.J., Stumbo, S.P., Janoff, S.L., Yarborough, M.T., McCarty, D., Chilcoat, H.D., Coplan, P.M., Green, C.A., 2016. Understanding opioid overdose characteristics involving prescription and illicit opioids: a mixed methods analysis. Drug Alcohol Depend 167, 49-56. https://doi:10.1016/j.drugalcdep.2016.07.024.
+Zalsman, G., Hawton, K., Wasserman, D., van Heeringen, K., Arensman, E., Sarchiapone, M., Carli, V., Hoschl, C., Barzilay, R., Balazs, J., Purebl, G., Kahn, J.P., Saiz, P.A., Lipsicas, C.B., Bobes, J., Cozman, D., Hegerl, U., Zohar, J., 2016. Suicide prevention strategies revisited: 10-year systematic review. Lancet Psychiat. 3 (7), 646-659. https://doi.org/10.1016/S2215-0366(16)30030-X.
+179

Accuracy of proactive case finding for mental disorders by community informants in Nepal.txt

@@ -0,0 +1,38 @@
+More than 75% of people living with mental health problems in low- and middle-income countries (LAMI countries) do not receive treatment for mental health problems.1 In an effort to close this vast treatment gap new models of treatment provision have been developed that propose a collaborative approach to care delivery, also known as task-sharing. This entails that the bulk of direct mental health service delivery is conducted by front-line health workers, rather than restricted to the domain of mental health professionals. This approach is also advocated by the World Health Organization’s (WHO’s) Mental Health Global Action Programme (mhGAP), an initiative aimed to equip primary healthcare workers to provide mental healthcare.2 Expanding the workforce and putting mental health services in place is crucial to close the treatment gap. However, this alone is inadequate without people with mental health problems actually making use of the services. Both availability and uptake of services are required to close the treatment gap. Thus, a major barrier for the scaling up of mental healthcare is the lack of awareness and demand for care.3 Low demand can be explained in part by non-detection or under-detection of mental health problems.4 There are a number of barriers to detection methods that are effective in high-income countries including: first, low literacy rates preclude use of self-administered self-report tools such as the Patient Health Questionnaire;5 second, high levels of stigma and existing belief systems regarding mental health problems discourage endorsement of psychiatric labels;6 and third, self-report instruments of serious mental illness, such as schizophrenia, have failed to detect cases in some LAMI countries.7,8 Therefore, development of innovative methods for identification of people with mental illness are needed to address limited literacy, cultural stigma, and applicability across a range of common and serious mental disorders.
+In response, community case-finding has been proposed to increase access to care in LAMI countries.9 Patel & Thornicroft10 propose a two-staged case-finding procedure with probable cases being identified through community case-finding followed by a
+diagnostic interview by a trained health worker. In practice, active case-finding has played an important part in increasing demand for, and accessibility of, mental health services in Nigeria and India.3, Case detection by lay-workers may, especially in low-income settings, hold some population-level advantages, including greater population coverage.12 Although proactive case finding has been used before, the current approach is, to the best of our knowledge, novel in a LAMI country setting, i.e. consisting of a structured approach using pictorial vignettes and an extensive development process emphasising compatibility with the sociocultural context. Also, to date, there has been limited evaluation of proactive case-finding strategies in LAMI country settings.
+The current study will evaluate the accuracy of a newly developed procedure for proactive case finding by community informants in Nepal, with the aim of increasing help-seeking for mental healthcare. Proactive case-finding has been introduced as a strategy to increase help-seeking, as it aims to bridge the gap between people in need of mental healthcare with available services. It provides an alternative to systematic community screening that is associated with high financial and resource burden. The developed procedure is based on the premise that well-placed informants, who are intimately connected to people in the surrounding community, have good insight into its members’ well-being. The hypothesis is that such a process of identification is best achieved through emblematic recognition, i.e. broadly matching of people they encounter in daily routine onto vignettes of mental health problems that have been made context-specific. This is the community version of a prototypematching approach for clinicians, which has demonstrated comparable validity to more complex diagnostic algorithms based on dichotomous decisions on individual symptoms.13,14 This is an approach that proposes to examine a diagnostic prototype (short narrative description) taken as a whole and to gauge the extent to which a patient’s symptom presentation matches the prototype.15 The developed procedure for proactive case-finding, which entails training of selected community members in the use of a structured
+tool (Community Informant Detection Tool, CIDT), is applied within a larger programme (Programme for Improving Mental Health Care, PRIME). PRIME aims to improve the coverage of treatment for priority mental disorders by implementing and evaluating a comprehensive mental healthcare package, integrated into primary healthcare in five LAMI countries (Nepal, India, South Africa, Ethiopia and Uganda).16 The care package includes the provision of psychosocial and pharmacological interventions by non-specialised primary health workers (following the WHO mhGAP Intervention Guide)2 and community counsellors. The objective of the study is to evaluate how accurate the CIDT procedure is in identifying people with priority mental disorders.
+Method
+Setting
+The research was conducted in Chitwan, a district in southern Nepal. Nepal is a low-income country, one of the poorest countries in Asia and is categorised by the World Bank as a fragile state.1 The total population of the country is approximately 26.5 million (Central Bureau of Statistics, 2011, www.dataforall.org/dashboard/ nepalcensus) with majority (86%) of the total population living in the rural areas. The country is passing through a transition following a 10-year intra-state conflict, between government forces and Maoists insurgents, which raged between 1996 and 2006 and claimed more than 13 000 lives. Previous studies have demonstrated the impact of political violence on psychosocial well-being and mental health in Nepal.18-21 The conflict has further shattered an already weak healthcare system. The war formally ended in November 2006 with a comprehensive peace agreement between an alliance of political parties and the Maoists. The present situation is characterised by instability and political deadlock. It is against the backdrop of recent violence and ongoing poverty that PRIME was implemented in Nepal. In Nepal’s healthcare system, there are sub-health posts that provide essential healthcare services and monitor community level healthcare activities; health posts that offer the same services with additional birthing centres, as well as the responsibility of monitoring the sub-health posts activities, and primary healthcare centres (PHCC) a higher level healthcare institution that serves as the first referral point for each electoral area. Currently, no mental health services are systematically available in primary healthcare.22
+Procedure
+The CIDT procedure, as introduced above, was constructed following a process that encompassed several steps. This process can be summarised as follows. First, vignettes were developed for some priority disorders (depression, alcohol use disorder, psychoses, epilepsy and conduct disorder) by taking vignettes in WHO’s mhGAP Intervention Guide2 as a starting point. Subsequently, an inventory of local non-stigmatising idioms related to the vignette was made based on prior ethno-psychological research in Nepal.6 Selection of most relevant idioms was done through prioritisation by an expert panel of Nepali mental health professionals. The CIDT therefore does not use the diagnostic labels but Nepali descriptions that have found to be commonly acceptable and understandable for each of the vignettes. To facilitate the process of prompt recognition of people that potentially match the vignette, pictures were developed (see online Fig. DS1 for an example CIDT sheet). Pilot testing of the procedure was instrumental in identifying the most relevant groups of people to serve as community informants. The pilot testing, as well as formative research done for the overall PRIME initiative,23 demonstrated high levels of perceived acceptability for the proposed procedure among key stakeholders.
+The CIDT procedure is used by community informants briefly trained in the essentials of public mental healthcare, the use of the procedure and the related ethical considerations. In their routine daily activities and tasks, the community informants aim to gauge the extent to which people in their direct vicinity match paragraph-long vignettes (aided with pictures) using a simple 5-point scale. If the community informant believes that a person in the community has significant features of the description (i.e. the person fits well with, or exemplifies, the description), then the informant answers two additional questions: one on whether the identified individual is perceived to have impaired daily functioning and a second question on whether the person would want support in dealing with these problems. In case of significant matching and a positive response to at least one of the additional questions, the community informant will encourage the person (possibly through their family) to seek help in the health facility where mental health services are being offered as part of PRIME, and where caseness can be confirmed by a trained health professional. No stigmatising diagnostic labels or psychiatric terminology is used, and encouragement for help-seeking is targeted to specific observable behaviours and/or signs of distress.
+To evaluate the accuracy of the CIDT we assessed whether the people identified by the community informants were correctly detected according to a structured diagnostic assessment performed by a clinician. This was done by comparing CIDT results with results of a structured clinical assessment following the Composite International Diagnostic Interview (CIDI).24 CIDT results comprised of either ‘probable caseness’, which were respondents that met the criteria outlined above, or ‘probable non-caseness’ which are respondents that do not meet those criteria. The probable negative cases were identified by asking community informants to select people they felt confident of not matching the pictorial vignette. The group is included only for the purpose of this study (true negatives and false negatives are needed for standard analyses on psychometric properties). In keeping with the proposed real-life application of the tool, respondents were not screened systematically; rather it presents a selected sample of people proactively identified by the community informants as probable positives or negatives. The goal of the CIDT procedure is not for community informants to make a specific diagnosis, but rather to identify someone with any mental distress that would benefit from treatment. It is therefore intended as a proxy indicator for people with mental disorders. As a result, the comparisons at the heart of the study are not based on identification of specific disorders but take them as a composite concept (i.e. a possible case of any disorder).
+The community informants that were part of the study included female community health volunteers (n = 8) and members of the local mother groups (n = 4), distributed over 6 wards. Members of mother groups do not require formal education, whereas female community health volunteers should minimally be literate and between 25 and 45 years of age. After receiving the 1-day training, they were asked to start using the CIDT forms. The training was provided by a health assistant, with years of experience in mental healthcare, who has been coordinating the training and implementation of the mental health services within the PRIME programme. The training consisted of minimal didactic teaching, and emphasised group discussion and practicing through role-plays. The community informants receive a small monthly allowance for their work (approximately US$2/month, paid as a flat rate as opposed to payment per identified case). All completed forms were handed over to a research assistant who subsequently arranged for the clinical assessment to happen with the identified person. When the forms were incomplete or unclearly completed the research
+assistants immediately contacted the community informant to clarify. Community members who were identified by the informants as potential cases were then approached by research assistants to request their permission to participate in the study. Participants were only included in the study, and clinical interviews were only conducted, after obtaining informed consent with signatures provided by literate participants and acknowledgement markings made by illiterate participants. Nepali psychosocial counsellors, with over 6 months of training and more than 5 years of experience in counselling in community settings, conducted the CIDI-structured clinical interviews and were masked to CIDT results at the time of the interview. Due to the shortage of mental health specialists in Nepal, psychosocial counsellors are the cadre of service providers with the most extensive skills-based clinical training other than psychiatrists who have limited availability to participate in research interviews given the overwhelming volume of clinical need. Psychosocial counsellors using structured clinical interviews have performed effectively in prior validation studies in Nepal.25
+The study was done in the area where pilot testing of the PRIME mental healthcare package was ongoing, which made referral to treatment possible. The treatment package included both psychotherapeutic and pharmacological interventions, and was offered in the nearby health facility. The PRIME programme and the CIDT procedure are endorsed by, and implemented in partnership with, the Ministry of Health and district level health authorities. Ethical approval was obtained from the Nepal Health Research Council and the Human Research Ethics Committee of the Faculty of Health Sciences, University of Cape Town (REC Ref: 412/2011). Data were collected in the period January to March 2013.
+Instruments
+In addition to the CIDT, we used different sections from the CIDI, notably the screens for psychosis, depression, alcohol use disorder, conduct disorder and oppositional defiant disorder. In addition, we used a 9-item screening questionnaire to detect epileptic seizures.26 The Nepali-language CIDI has been validated in Nepal, (area under the curve any disorder = 0.85, area under the curve depression = 0.97).27 The psychosocial counsellors received a week of training in the Nepali CIDI including 6 h of observed administration and review of scoring. The training was conducted by an Australian psychologist with experience in conducting structured clinical interviews. Additional training and detailed review of videotaped interviews was provided by an expatriate psychologist (M.J.D.J) and an expatriate psychiatrist (B.A.K.) who are both fluent in Nepali.
+Analyses
+Using descriptive statistics the results from the CIDT and the clinical assessments were compared, and plotted as true or false positives and true or false negatives. Next, positive predictive value (PPV) and negative predictive value (NPV), positive and negative likelihood ratios were calculated. Analyses were done for the entire sample, as well as for children and adults separately. Primary analyses were done with caseness defined as identification or diagnosis of any disorder (see above). In addition, exploratory analyses were conducted for the disorders separately. Analyses were conducted using the Statistical Package for Social Sciences (SPSS version 19.0).28
+Results
+The total sample consisted of 195 people. In total 210 people were identified by the community informants, of whom 5 people
+refused to participate and 10 people were unable to complete the interviews. See Table 1 for sociodemographic details of the sample for adults and children separately. In the combined sample, the average age was 32.21 years (s.d. = 15.47) and comprised of 59.5% female participants. Interrater reliability between the two counsellors based on independently conducted clinical interviews repeated with the same patient series was found to be good (intraclass correlation = 0.92; 95% CI 0.89-0.94). All of the study participants were identified by the community informants using the CIDT procedure, either as a probable positive (n = 110) or as a probable negative case (n = 85). After clinical assessments, 70 of the CIDT positive cases and 6 of the CIDT negative cases were found to meet criteria for a clinical diagnosis for one or more of the target disorders (i.e. true positives = 70, true negatives = 79). See Table 2 for a breakdown of the diagnoses (with the total diagnoses (101) exceeding the 76 individuals due to comorbidity).
+Table 2 further summarises the results of the analyses, including PPV (for the entire sample: 0.64), NPV (0.93), positive likelihood ratio (2.71) and negative likelihood ratio (0.12). The results for children and adults differ substantially only for the PPV (0.50 among children), due to the high number of false positives among the child sub-sample. When comparing the results between both types of community informants, we see differences on all indicators, with a far smaller proportion of false positives among the mother group participants. Finally, for explorative purposes we have included the analyses for the separate diagnostic clusters, indicating how well the separate disorder-specific vignettes of the CIDT detect those disorders.
+We also analysed these results when varying the CIDT decision algorithm. Results changed only marginally when allocating positives based on matching symptoms only, or based on symptoms plus both additional criteria of functioning impairment and need for support - rather than matching of symptoms plus at least one of the two additional criteria (which is the original algorithm that was used in this study).
+Mental healthcare in LAMI countries is characterised by low resources,29,30 which goes hand in hand with low demand for,
+and utilisation of, services. To fully capitalise on efforts to increase trained human resources and evidence-based treatments, an increase in demand and utilisation is imperative. Relying solely on self/family-referrals may risk missing out on a large group of clients that are simply not identified, yet need treatment. Use of self-report written screeners is not widely feasible given low literacy rates. A context-sensitive procedure of proactive casefindings, built on the notion of using non-stigmatising idioms and matching people on vignettes (aided with visual clues and two questions), has been piloted in rural Nepal to identify need for treatment and ultimately increase the demand for and utilisation of mental healthcare.
+Community informants can assign caseness in the persons rated with pictorial vignettes with accuracy for the majority of persons rated. Given that real-life use of the CIDT only results in reporting positives, as opposed to a usual universal or community screening procedure that will result in both negatives and positives, the key indicator of accuracy of the CIDT are the PPV (i.e. the proportion of positive test results that are true positives) and the positive likelihood ratio (i.e. how much to increase the probability of disease if the test is positive), which were 0.64 and 2.71 respectively for all respondents and all disorders in this study.
+With approximately two-thirds of the expected positives having a confirmed disorder, the community informants do well overall. By comparison many common standardised symptoms-based screening instruments have a similar (or lower) PPV31 or positive likelihood ratio32 - noted of course that these are generally disorder specific, and some also have higher values. This is quite surprising, given that the CIDT informants in our study do not have any formal training or even formal education in many instances. Although a promising strategy to identify cases that might have otherwise gone unnoticed, it does present with a risk for over-identification and additional burden for the health facilities,33 but this burden is not necessarily higher than would be encountered with a standardised screening instrument. Overburdening providers is a serious concern with many health facilities already strained. At the same time, it is plausible that the ‘false positives’ do indeed require treatment but are just subthreshold or have problems that were not assessed with the five CIDI disorders covered in the clinical interviews.
+These promising results may be explained by the focus on a vignette-based structure, which like the prototype matching approach for clinicians, is a form that is more congruent with human (and clinical) cognitive processes than checking whether each of a series of symptoms is absent or present.13 The use of this approach in a community setting is quite new, and, to the best of
+our knowledge, a first in a low-income setting. In addition, incorporating locally salient manifestations of mental health problems through idioms of distress and drawings, may have contributed to good accuracy. This was also demonstrated in the cross-cultural construct validation of a brief screener for psychosocial distress for children in conflict affected settings.34
+To implement this proactive case finding strategy, it is important to know who the most adequate community informants might be. Formative research had already indicated that in Nepal female community health volunteers and mother group leaders would be most appropriate. The results further demonstrate that mother groups outperform female community health volunteers as the former have significantly lower false positives. From a perspective of potentially burdening the system, involving the mother’s groups would lead to lower burden. This is an interesting finding given that female community health volunteers are currently an extension of the healthcare system in Nepal, and are excessively used for different task-shifting roles. Based on this finding it would actually be better to not add another task on this group of community health workers, but rather chose of group of people that is less taxed and may have better knowledge of the well-being of the community members.
+For determining the overall accuracy of the CIDT, it is not necessary to see how well the procedure picks up on individual disorders. Yet, it is interesting to see that the depression module was most accurate among the five included disorders (in terms of PPV). This goes counter to the notion that depression and mood disorders generally are the more challenging to identify by lay people,9, 5 and also contrary to the formative study into the feasibility of this approach. This is a promising result for a problem that is present so ubiquitously, but with near absence of treatment at present in rural Nepal. This would mean that the aim to increase demand for mental healthcare using the CIDT will not be limited to the more ‘visible’ disorders, such as schizophrenia or alcohol use disorder. It should be noted that the numbers for these subgroup analyses are very small, so this trend will need to be confirmed with a larger sample. An issue that should also be further considered when implementing this procedure is that it is less accurate in identifying children’s mental health problems. This fits with common patterns of under-detection of children’s problems.36 It is possible that this difficulty is especially pertinent for the conduct problems, and possibly less so for problems that are more clearly demarcated as pathological (i.e. severe mental retardation) by community members. Further development is needed to fine-tune the procedure towards children’s mental health problems.
+There are important potential downsides to this sort of case detection.12 Community informants could potentially abuse their new role, however informal it is, and force people to seek treatment who may not be willing to do so. Individuals who are identified may experience stigma and discrimination, particularly if community informants are not bound by a clear code of confidentiality. The power dynamics of using this procedure is something that was addressed in the training and has been monitored throughout the pilot phase. It has not yet led to any such incidents, but this does not mean it will not, especially when the strategy is scaled up making intensive monitoring much harder. Also, possible downsides of a vignette-based approach are that its users may selectively recall certain features of the prototype, and that it allows for a lack of standardisation between users.37 It is important to emphasise that the proactive case finding strategy is not meant as a form of systematic community screening. If it were, population prevalence rates would need to be taken into consideration resulting in lower PPVs. Selection of respondents for this study was therefore pragmatic, reflecting the CIDT’s intended real-life use. Although asking community informants to identify negatives they ‘felt confident’ about is congruent with the actual process of excluding cases when engaged in proactive case-finding; this identification may have had an effect on the NPV. A strength of the approach is the extensive development process that relied on available ethnographic study for selecting idioms of the vignettes and drawings, incorporating local stakeholder perspectives and finetuned a training package that balanced utilitarian and ethical concerns. This is important to emphasise, as potential replication of this approach without such preparation might impact the accuracy and introduce ethical or clinical risks.
+Future research is needed to assess the actual effectiveness of the proactive case-finding procedure. Where the present study evaluated the accuracy, the next step is to evaluate whether the use of the CIDT also results in an increase in demand for, and uptake of, mental healthcare. Furthermore, more work is needed to make the procedure more sensitive to capture children’s mental health problems in the future.
+Implications
+The procedure shows potential to identify the right people in need of treatment and the study suggests that it provides for a good surveillance procedure. About 64% of the people that the community informants identified as probable cases using the CIDT were actually positive cases based on clinical interviews and 93% of people that community informants were confident probable non-cases, were indeed found negative. It appears that the procedure does not need to exclusively rely on already overburdened community health volunteers. Given the selected use of proactive case finding, the procedure is not a substitute for systematic community screening. Actually, the CIDT may present a pragmatic alternative approach preferable to community screening.
+The CIDT can be an important demand-side strategy to increase help-seeking for settings that are integrating mental health into primary healthcare. It can be used in conjunction with the training and implementation of WHO’s mhGAP guidelines, and can be scaled up relatively easily to a national level. From a policy point of view this is important, given the commitment that so many countries have made towards this goal.38 The use of the proactive case-finding may lead to significantly increased coverage of mental healthcare in a target area where mental health services are put in place, provided that the community informants are selected to represent a small geographical area (village or part of a village) where they are intimately connected and known. As
+stated above, the last part is at present still an assumption. Currently, research is planned to evaluate the effectiveness in facilitating referrals and reinforcing treatment-seeking behaviour.

An Empirical Investigation of Acculturative Stress and Ethnic Identity.txt

@@ -0,0 +1,61 @@
+Though experts agree that suicide is characterized by a strong cultural element (Institute of Medicine of the National Academies, 2002; Maris, Berman, & Silverman, 2000), few studies have examined culturally relevant phenomena in delineating suicide risk for diverse ethnic groups in the United States. Some studies have found evidence for increased suicide risk and depression in accul-turated Central American and Mexican adult immigrants (Hovey, 2000a, 2000b) and youth (Hovey, 1998). Other studies have examined religiosity as a culturally relevant factor in buffering suicide vulnerability in African Americans (Stack, 1998; Walker & Bishop, 2005). However, only one study to our knowledge (see Kaslow et al., 2004) has examined African American ethnic identity in predicting suicidal behavior. One other study (Joiner & Walker, 2002) considered acculturative stress as a factor in suicidal ideation in African Americans. Neither study examined the moderating capacities of either acculturative stress or ethnic identity in understanding the relation of depression to suicidal thoughts in African Americans. Though there has been an increase in the African American suicide literature (e.g., Castle, Duberstein, Meldrum, Conner, & Conwell, 2004; Harris & Molock, 2000; Ialongo, Kaslow McCreary, & Pearson, 2002; Marion & Range, 2003; Palmer, 2001; Roy, 2003; Willis, Coombs, & Drentea, 2003) much more definitive work in this area is needed.
+African American suicide remains poorly understood. Risk factors that have been identified for suicide deaths in European
+American youth and adults do not hold up for African Americans (as an example, see Garlow, Purselle, & Heninger, 2007). Abe, Mertz, Powell, and Hanzlick (2004) compared medical examiner reports for 784 White and 348 Black suicide deaths and found that Blacks were younger, less likely to have a history of depression, and less likely to have financial problems, suicide gestures, chronic disease, and substance abuse relative to Whites who died by suicide. Given a narrow understanding (i.e., exclusion of culturally relevant variables) of causal factors in suicide risk assessment, fatal suicide attempts for African Americans are inherently less predictable than those of European Americans. As an example, alcohol or cocaine use, highly cited factors in suicide attempts (in primarily European American populations) were detected for less than 18% of African American youth suicide deaths compared with more than 40% of European American youth suicide deaths (see Garlow, et al., 2007). Though studies cite protective features of African American culture that mitigate suicide risk (see Early & Akers, 1993; Gibbs, 1997), investigations of culturally relevant phenomena have been limited primarily to studies of religiosity, spirituality, or social factors (see Compton, Thompson, & Kaslow, 2005; Kaslow et al., 2004; Marion & Range, 2003). Broader investigations to contextual factors such as acculturation and acculturation stress, which have been identified for some underrepresented groups in the U.S., may provide a better understanding of African American suicide risk. We will explore the cross-cultural relationships of acculturative stress and ethnic identity to suicide ideation in a sample of African Americans and European American college students.
+Among college students, suicide is a leading cause of death (Center for Disease Control and Prevention [CDC], 1997). Risk factors are said to include depression (Lester, 1999) and also hopelessness (Heisel, Flett, & Hewett, 2003). However, studies
+very rarely explore suicidality and cultural milieu beyond those of American Indian (Middlebrook, LeMaster, Beals, Novins, & Manson, 2001), Asian American (Range et al., 1999), Latino (Hovey & King, 1997), and African American (Walker & Bishop, 2005) youth and young adults in the U.S. Interestingly, European American, and African American college students did not differ significantly on religiosity associated with suicidal ideation (Walker & Bishop, 2005). To our knowledge, broad cultural studies of belief systems, behavioral acceptability, and sociocultural experiences have not been widely explored for European American college students. Nevertheless, such investigations contribute to a more comprehensive understanding of suicide risk.
+Suicide risk assessment for African Americans remains a complex task as emerging data reveals that African Americans’ pattern of suicide risk diverges significantly from previously identified patterns for delineating risk. Garlow, Purselle, and Heninger (2005) reported marked ethnic group differences in suicide mortality such that the mean age for Black suicide death in Fulton County, Georgia was 32 years compared with 44 years for White suicide deaths. This shift in age-risk may have implications for distinctive risk factors. Other studies suggest geographical (Willis et al., 2003) and familial differences (Roy, 2003) in suicide death among African Americans such that African Americans are more likely than Whites to have died in urban areas and less likely to have a family history of suicide death. Unexplained ethnic group differences in suicide behavior and mortality merit broader, cultural, and ethnic levels of analyses.
+African American college students are said to disclose suicidality less readily than their White counterparts (Morrison & Downey, 2000) even when suicide acts are imminent (Molock, Kimbrough, & Lacy, 1994). African American youth in transition to university settings may be faced with unique contextual experiences (e.g., increased perceived discrimination) that are predictive of suicide risk levels. Though suicidality was not explored in available studies of discrimination experiences, discrimination was implicated in 35% of stressful life experiences for Black college students. Swim (2003) found that students reported weekly experiences of racism on average. These environmental antagonists and other contextual experiences are rarely explored in suicide research.
+Acculturation
+Acculturation is a complex, psychosocial phenomenon that involves individual and group-level changes in cultural patterns for ethnic minorities as a consequence of contact with the ethnic majority (see Chun, Organista, & Marin, 2003). Acculturative stress is the stress that is associated with cultural adaptation, which may occur at the risk of certain psychological consequences. Acculturative stress has been linked to symptoms of suicide and depression in Latino populations (Hovey, 1998, 2000a, 2000b), depression in African, Asian, and Latin American international college students (Constantine, Okazaki, & Utsey, 2004) and bulimic symptoms in African American and Hispanic college students (Perez, Voelz, Pettit, & Joiner, 2002).
+Though studies have explored the significance of acculturative stress for African Americans both conceptually (Anderson, 1991) and empirically (Joiner & Walker, 2002), exploratory investigations of the psychological and emotional impact of acculturation and acculturative stress rarely include African Americans. Pope-
+Davis, Liu, Ledesma-Jones, and Nevitt (2000) linked acculturative stress to racial identity defined as “a measure of the importance that members of an ethnic group place on their cultural heritage” (p. 197). They remarked that conceptual ambiguities have hindered the development of studies that investigate racial and ethnic identity because theories typically fail to explain the process by which identification (with one’s cultural group) occurs. Nevertheless, Pope-Davis and colleagues asserted that, when studied together, acculturation and ethnic identity may create a more complete picture of African American psychosocial development.
+The Group for the Advancement of Psychiatry (GAP, 1989) and others (Davis, 1980; Gibbs, 1984, 1997; Gibbs & Hines, 1989; Walker, Utsey, Bolden, & Williams, 2005) posited that cultural changes may be related to African American suicide deaths. These changes have potentially occurred via acculturation that likely brings about an erosion of religious, spiritual, and social protective factors as well as cultural beliefs (e.g., suicide as unacceptable). Many studies in African American suicide have focused on the religiosity-spirituality spectrum as a protective factor in African American suicide deaths, citing religious well-being and spirituality as cultural buffers (Marion & Range, 2003), coping resources (Kaslow et al., 2002), or deterrents (Early & Akers, 1993). Other studies have emphasized the importance of social support as a protective factor in suicidal ideation (Compton, Thompson, & Kaslow, 2005; Nisbet, 1996). Though religiosity, spirituality, and social support have revealed important buffering conditions, the effects of other sociocultural variables have remained gravely understudied.
+Ethnic Identity
+The U.S. Public Health Service (2001) report cited ethnic identity and acculturation along with other factors in understanding the severity of mental health challenges for ethnically diverse groups. According to Phinney (1992), ethnic identity is a reliable construct for understanding adherence to values and beliefs that are reflected by a cultural group. In college student populations, identity resolution may be particularly salient as students separate from families of origin and venture independently into a new stage of life. Though both European American and African American youth experience group esteem, ethnic identification has been observed more saliently for African American adolescents (French, Seidman, Allen, & Aber, 2006).
+Studies have found that ethnic identity buffers potentially negative mental health outcomes. Ethnic identification has been linked to positive self-esteem in Black college students (Phelps, Taylor, & Gerard, 2001) and is suggested for incorporation in drug prevention programs for young African American adults (Brook, Balka, Brook, Win, & Gursen, 1998). Ethnic identity or other sociocultural factors may, at least in part, account for differences in depressed students who may or may not be suicidal.
+To our knowledge, cross-cultural investigations in suicide and identity are nonexistent though challenges with identity resolution are potentially suicidogenic across cultural groups. With one exception, investigations of African American adult suicide have largely ignored the potential relationship of identity and suicide risk. Kaslow et al. (2004) found that African American suicide attempters scored lower on the Multigroup Ethnic Identity Measure (MEIM; Phinney, 1992) than nonattempters. Thus, adult
+suicide attempters reported lower rates of belongingness and group orientation. Studies that include predominantly European American samples frequently cite sexual identity crises as precipitants to suicidal behavior (see Kulkin, Chauvin, & Percle, 2000, for review).
+Current Study
+The purpose of the present paper was to explore ethnic group differences in the relationship between suicide and depression, one of the most common risk factors for suicide ideation and attempts (Goldsmith, Pellmar, Kleinman, & Bunney, 2002). Importantly, we evaluated the depression-suicide relationship in the context of third variables, ethnic identity, and acculturative stress. Given that factors in African American suicide have differed unexpectedly from those of European Americans, we speculated that accultura-tive stress and ethnic identity, important sociocultural variables might distinguish certain subgroups of individuals who are at risk. Though Perez et al. reported evidence of acculturative stress in a sample of European American college students, we posited that acculturative stress might affect African American and European American depression and suicide differently. Additionally, ethnic identity is a cross-cultural construct in which a comparative study is advantageous in parceling out potentially unique factors across ethnic groups. The cross-cultural emphasis proposes a precise, model of the depression-suicide relationship that is expected to better predict suicide ideation for African Americans than European Americans and expand existing models of suicide, a complex phenomenon. Such precision broadens the spectrum of variables that are considered in suicide assessment and scientific inquiry (see Triandis & Brislin, 1984).
+The explicit hypotheses for the current study were: (a) depressive symptomatology is positively correlated with suicidal ideation for both African Americans and European Americans; (b) accul-turative stress moderates the relationship between depressive symptomatology and suicidal ideation for African Americans but not European Americans such that the relationship between selfreported depressive symptomatology on suicide ideation is increased for acculturatively stressed individuals and; (c) ethnic identity moderates the relationship between depression and suicidal ideation for African Americans but not European Americans such that the relation for suicide and depression is strengthened in the absence of a strong ethnic identity.
+Method
+Participants
+The participants were 459 university students who participated in this study to partially fulfill a requirement for an introductory psychology class or to gain some other academic credit. Mean age for the total sample was 20.88 years (SD = 3.08 years). The ethnic composition of the sample was 64.5% African American (n = 296) and 35.5% European American (n = 163). Female participants represented the majority of European American (60%; n = 168) and African American (70%; n = 114) participants. There were 248 (248; n = 54%) students enrolled in a predominantly White public university in the southeastern U.S. There were 163 (163; n = 36%) students were enrolled in a historically Black public
+university in the southeastern U.S. The institution-type was not reported for 10% (n = 48) of students.
+Measures
+Societal, Attitudinal, Familial, and Environmental (SAFE) Acculturative Stress Scale. Levels of acculturative stress were measured by a modified, short version of the original 60-item SAFE scale used in previous studies (Fuertes & Westbrook, 1996; Mena, Padilla, & Maldonado, 1987). The short version of the SAFE scale measured acculturative stress in social, attitudinal, familial, and environmental contexts, along with perceived discrimination toward immigrant populations (Mena, Padilla, & Maldonado, 1987). Example items include, “In looking for a job, I sometimes feel my ethnicity is a limitation,” and “It is hard to express to my friends how I really feel.” According to Mena and colleagues, scores on the SAFE scale correlated negatively with both “ethnic loyalty” (r = —.35, p < .001) and “loyalty to parents” (r = —.25, p < .001). Participants were required to rate each SAFE item that applied to them on a Likert Scale, ranging from 1-not stressful to 5-extremely stressful. In this study, items that were “not applicable” were skipped and scored “0.” Consequently, the individual total scores were prorated to reflect possible skipped items. The possible scores for the SAFE ranged from 0 to 120. Joiner and Walker (2002) previously detailed evidence for convergent and discriminant validity for African Americans. The SAFE has also been shown to be reliable for Asian Americans and international students (a = .89; Mena et al., 1987), a heterogeneous group of Hispanic Americans (a = .89; Fuertes & Westbrook, 1996), and Black college students (a = .87, Joiner & Walker, 2002; a = .87 (Perez, Voelz, Pettit & Joiner, 2002). Similar alpha was obtained in this sample (a = .89; n = 459).
+Multigroup Ethnic Identity Measure (MEIM). The MEIM (Phinney, 1992) is a measure of ethnic identification based on the elements of ethnic identity that are said to be common across ethnic groups (Phinney, 1992). Some evidence indicates that the MEIM is a useful global measure of ethnic identity (see Roberts et al., 1999). Phinney (1998) asserted that ethnic identity can be considered a component of acculturation that focuses on the individual’s attachment and relation to his or her own (sub)group of the larger society. Anderson (1991) further explained that racial pride equips Black people to cope with acculturative “threats.” The MEIM consists of 14 items that assess three aspects of ethnic identification (i.e., positive ethnic attitudes and sense of belonging; ethnic identity achievement; and ethnic behaviors/practices). In this study, participants were required to rate each item on a Likert Scale, ranging from 1-strongly disagree to 4-strongly agree. The items were summed for a total score; higher scores represented more positive ethnic group identity. The MEIM has been shown to be valid and reliable for Asian American, Black, Mexican American, and White students (see Phinney, 1992 for a review; see also Sellers et al., 1998). The MEIM was also found to be reliable in the current sample (a = .87; n = 449).
+Beck Suicide Scale (BSS). Suicidal ideation was measured by the BSS (Beck & Steer, 1993), a 21-item self-report inventory. Each item consists of groups of statements that represent increasing levels of severity on a scale ranging from 0 to 2. As an example, one “0” statement is “I have no wish to die.” The “2” statement in that group is “I have a moderate to strong wish to die”.
+Items 1 through 19 contributed to a possible total score that ranged from 0 to 38. Items 20 and 21 referred to past suicide attempts and were optional. The BSSs reliability and validity have been well supported (see Beck & Steer, 1993; see also Beck, Steer, & Ranieri, 1988). In the current study, a = .91; n = 431.
+Beck Depression Inventory (BDI). Levels of depressive symptoms were assessed by the BDI, a 21-item self-report inventory. Each item was rated on a scale ranging from 0 to 3. Thus, possible inventory scores ranged from 0 to 63 in which higher scores represented increased severity. Although the BDI is not indicative of the full clinical syndrome of depression, it is a reliable and well-validated measure of depressive symptomatology (see Beck, Steer, & Garbin, 1988 for a review; see also Kendall, Hollon, Beck, Hammen, & Ingram, 1987). In the current study, a = .84; n = 432.
+Procedure
+The present study was granted full institutional review board approval. Participants were solicited from undergraduate and graduate courses in two southeastern university psychology departments. Each participant was informed that she or he would be administered a questionnaire packet that included questions about their behavior, views, and feelings with regard to depression, cultural identity, and suicidal thoughts. Each participant was also given a consent form that stated that consent for participation in the study was assumed upon completion of the anonymous questionnaire packet. The primary investigator, a licensed clinical psychologist and suicidologist, was immediately available in the event that any study participant was at risk for imminent danger. Students were informed that participation in the study could cease at any time and referral to the university counseling center or psychology clinic for free services would be available if needed. None of the participants discontinued participation, requested a referral for psychological services, or demonstrated imminent risk for danger. Approximately 25 minutes were required to complete the questionnaires.
+Results
+Means, standard deviations, and intercorrelations for all measures are presented for African American and European American college students in Table 1. All values were within expected limits. As Table 1 shows, self-reported depressive symptoms were similarly correlated with suicidal ideation for both African American (r = .54, p < .01) and European American (r = .54, p < .01) college students such that the more depressive symptomatology that was reported, the more suicidal thoughts reported. As expected, both acculturative stress and ethnic identity were associated with suicidal thoughts for African American college students such that higher acculturative stress (r = .29, p < .01) and lower ethnic identity (r = —.23, p < .01) correlated with increased suicidal thoughts. Acculturative stress was also associated with suicidal thoughts in European American college students (r = .19, p < .05). This is consistent with Perez, Voelz, Pettit, and Joiner’s (2002) findings and may reflect culture-related stress as a function of being immersed in a novel setting (i.e., college setting).
+Hierarchical Multiple Regression
+Hierarchical multiple regression was used to identify the presence and nature of moderating effects (Aiken & West, 1991; Cohen & Cohen, 1983). As recommended, scale scores were centered to reduce multicollinearity between the main effect and interaction terms (Cohen & Cohen, 1983). Further, West, Aiken, and Krull (1996) noted that centering continuous variables ensures the interpretation of effects would occur at a meaningful value (i.e., the mean, which has a value of 0 with centered variables).
+Acculturative Stress as a Moderator for Depressive Symptoms and Suicide Ideation in African Americans
+To test a main hypothesis that acculturative stress moderates the relationship between depressive symptoms and suicide ideation for African Americans, a hierarchical regression equation was constructed with BSS scores as the dependent variable and BDI scores as the predictor variable. SAFE scores were added in the second step. In the third step, the interaction of BDI and SAFE scores was entered into the regression equation as a predictor. A significant overall effect was found for the regression equation (r = .61; F(3, 295) = 57.50, p < .001). Thus, together, depressive symptoms, acculturative stress, and the depressive symptoms x acculturative stress interaction accounted for 37.2% of the variance in predicting suicide ideation. The main effects for depressive symptoms and acculturative stress were significant in predicting BSS scores (see Table 2). The depressive symptoms x acculturative stress interaction was also significant (partial correlation = .30, t(294) = 5.45, p < .001).
+Holmbeck (1997) suggested evaluating high and low scores of the moderator variable to interpret the interaction. Accordingly, we examined the relation between BDI scores and BSS scores among two subgroups of participants: those who reported low and those who reported high levels of acculturative stress (i.e., those who scored one standard deviation above the SAFE mean, and those
+who scored one standard deviation below the SAFE mean). The regression equation predicted BSS scores for those high in acculturative stress (r = .62, p < .001), but not for those low in acculturative stress (r = .05, p = .75). This pattern of results indicates that the nature of the relationship between depression and suicide differed for individuals who reported high levels of accul-turative stress and those who reported lower levels of acculturative stress.
+Ethnic Identification as a Moderator for Depressive Symptoms and Suicide Ideation in African Americans
+To test a main hypothesis that ethnic identification moderates the relationship between depressive symptoms and suicide for African Americans, a hierarchical regression equation was constructed with BSS scores as the dependent variable and BDI scores as the predictor variable entered first into the regression equation. In the next step, MEIM scores were added. In step three, the interaction of BDI scores and MEIM scores was entered into the regression equation as a predictor. A significant overall effect was found for the regression equation (r = .62; F(3, 295) = 61.62, p < .001). Thus, together, depressive symptoms, ethnic identity, and the depressive symptoms x ethnic identity interaction accounted for 38.4% of the variance in predicting suicide ideation. The main effects for depressive symptoms and ethnic identity were significant in predicting BSS scores (see Table 2). The depressive symptoms x ethnic identity interaction was also significant ( partial correlation = —.32, t(294) = —5.86, p < .001), thereby demonstrating a moderating effect for ethnic identity.
+To interpret the interaction, we evaluated the relation between BDI scores and BSS scores among those who reported low and those who reported high levels of ethnic identity (i.e., those who scored one standard deviation below the MEIM mean, and those who scored one standard deviation above the MEIM mean). The
+regression equation predicted BSS scores for those high in ethnic identity (r = .55, p < .05), but more so for those low in ethnic identity (r = .75, p < .001). This pattern of results indicates that the strength of the depression-suicide relationship was greater for African American students who reported low levels of ethnic identity.
+Acculturative Stress as a Moderator for Depressive Symptoms and Suicide Ideation in European Americans
+To test a main hypothesis that acculturative stress moderates the relationship between depressive symptoms and suicide for European Americans, a hierarchical regression equation was constructed with BSS scores as the dependent variable and BDI scores entered as the predictor in Step 1 of the regression equation. In Step 2, SAFE scores were entered in the equation. In the third step, the interaction of BDI and SAFE scores was entered into the regression equation as a predictor. A significant overall effect was found for the regression equation (r = .55; F(3, 162) = 22.78, p < .001). Thus, together, depressive symptoms, acculturative stress, and the depressive symptoms x acculturative stress interaction accounted for 30.3% of the variance in predicting suicide ideation in European Americans. The main effect for depressive symptoms but not acculturative stress was significant in predicting BSS scores (see Table 2). The depressive symptoms x acculturative stress interaction was also not significant (partial correlation = .10, t(161) = 1.23, p = .221).
+Ethnic Identification as a Moderator for Depressive Symptoms and Suicide Ideation in European Americans
+To test a main hypothesis that ethnic identity moderates the relationship between depressive symptoms and suicide for European Americans, a hierarchical regression equation was constructed with BSS scores as the dependent variable and BDI scores entered as the predictor in the regression equation. MEIM scores were entered in Step 2. In Step 3, the interaction of MEIM and BDI scores was entered into the regression equation as a predictor. A significant overall effect was found for the regression equation (r = .55; F(3, 162) = 23.21, p < .001). Thus, together, depressive symptoms, ethnic identity, and the depressive symptoms x ethnic identity interaction accounted for 30.3% of the variance in predicting suicide ideation. The main effect for depressive symptoms but not ethnic identity was significant in predicting BSS scores (see Table 2). The depressive symptoms x acculturative stress interaction was also not significant (partial correlation = —.12, t(161) = -1.49, p = .138).
+Discussion
+The overall aim of the current paper was to investigate the relationship of acculturative stress and ethnic identity to selfreported depressive symptoms and suicidal ideation in a cross-cultural sample. As expected, we found that depressive symptomatology was correlated with suicidal ideation in both African American and European American college students. Our finding that the strength of the depression-suicide ideation correlation was similar for European American and African American college students is noteworthy as some studies have indicated that African
+Americans who die by suicide are less likely than European Americans to demonstrate symptoms of depression. It may be that African American college students are as likely to consider suicide when depressed, but this does not confer increased risk for a fatal suicide attempt. Because studies report high rates of suicide attempts for both African American males (Centers for Disease Control, 2004) and females (Nisbet, 1996) that mimic and/or exceed those of European Americans, additional studies of moderating and mediating effects of cultural phenomena in suicide fatalities are warranted.
+We found that acculturative stress was related to suicidal ideation in both African American and European American students. However, ethnic identity was only associated with suicide ideation in African Americans. Further, the depression-suicide relationship strengthened for a subgroup of African Americans. That is, we found that acculturative stress moderated the effect of depression on suicidal ideation for African Americans such that suicidal ideation was increased for African American college students who were depressed and also acculturatively stressed. Depression was not moderated for European American college students or for African American students who were not acculturatively stressed. This finding sheds light on subgroups of depressed African Americans who may consider suicide. That is, the experience of accul-turative stress, not low or high levels of acculturation per se, kindles suicide ideation. Though acculturation level was not included as a variable in this study, other studies have measured psychological effects of acculturation level with mixed results (see Neff & Hoppe, 1993; Rogler, Cortes, & Malgady, 1991, for review). Contradictory conclusions have emerged such that acculturation is said to be positively adaptive for some while others argue that adopting the mainstream culture is psychologically toxic. In a study of acculturation level and suicide attempts and ideation, Walker, Utsey, Bolden, and Williams (2005) found that selfreported suicidal thoughts and attempts decreased as a function of a higher acculturation status. Since this finding was contrary to prediction, Walker and colleagues speculated that “unacculturated persons [may] specifically experience more acculturative stress as a pressure to assimilate to mainstream society” (p. 213). Future studies should likely explore the relationships of both acculturation level and acculturative stress along with ethnic identity in predicting suicidal ideation.
+We hypothesized that ethnic identity would moderate the relationship between depression and suicide ideation such that the relation for suicide ideation and depression is strengthened in the absence of positive ethnic identity. Similar to the pattern of findings for acculturative stress, African American (but not European American) college students who were less attached to their ethnic group reported a stronger relationship of depression to suicidal ideation than those who endorsed a stronger attachment to their group. This is consistent with Kaslow et al.’s finding that African Americans who reported lower ethnic group identification were more likely to have attempted suicide than other Africans Americans who were seeking medical care (i.e., not in psychiatric crisis).
+To our knowledge, this is the first study to investigate the moderating effects of acculturative stress and ethnic identification in relation to depression and suicide ideation. We found convincing evidence that certain subgroups of African American college students who report symptoms of depression are more likely to consider suicide given poor group identity or high levels of accul-
+turative stress. European American college students, while stressed by the process of adjusting to a new environment, were not similarly at risk. Though the proposed model is not exhaustive toward discriminating cross-cultural determinants of suicide ideation, the findings offer important insight to how third variables might be informative in minimizing assessment errors (e.g., false positives).
+Overall, the current study highlights the relevance of cultural factors in the provision of mental health services, and therefore has implications for the evaluation, intervention, and treatment of African American college students in particular. As an example, the interactive risk of depressive symptomatology and accultura-tive stress (or ethnic identification) should be included in suicide risk assessment protocol. Though negative life events and stressful circumstances are known to trigger suicidal ideation and crises, stressors associated with the acculturative process amplify risk for African American college students. Future studies may also consider the compound effects of discrimination, perceived racism and other race-related stressors in addition to culturally relevant factors.
+Theoretical advances that embrace complex psychological, sociocultural, and biological models of suicide risk generate meaningful approaches to understanding and preventing suicide. In the current study, suicidal ideation increased in the presence of poor group identity and acculturative stress for African Americans. This conclusion factors into the multidimensional nature of suicide risk and the need for research that is more comprehensive, evaluation, and treatment.
+Some cautions and limitations should be noted. The first limitation of the current study is related to the selection of participants. The students’ suicide history was not known, and the overall variability in BSS scores was low. Though significant effects were observed despite the low variability in suicidal ideation, future studies may focus on clinical samples where suicide history is established. These studies might also represent more diverse age groups and levels of education. Given different rates of suicide across age groups (Garlow, Purselle, & Heninger, 2005), data that demonstrate suicide risk should be disaggregated such that risks for college age African Americans are not compared with those of elder African Americans. Older African Americans may respond differently to acculturative stress. Group identity may be even more resolved, given time and enduring effects of segregation. The range of education and perhaps, the range of socioeconomic status (SES) in the current study were restricted. Though the college sample used in the present study was consistent with those used in past suicide research (which presupposed that “advantaged” individuals, higher in SES and education, demonstrate higher levels of suicidality; see Selkin, 1983), the study’s generalizability beyond college samples is limited.
+The use of the single-informant, self-report, cross-sectional methodology added to the study’s limitations. Questionnaire items may have elicited minimization or exaggeration of psychological symptoms and cultural variables. Future studies would benefit from an outside, independent observation of the participants’ emotional and psychological status. In addition, the cross-sectional nature of the study only provided a snap-shot in time, and therefore, the data were not sufficient for causal assumptions. Future studies incorporating longitudinal analyses could potentially provide evidence that increases or decreases in acculturative stress,
+ethnic identification, and depressive symptomatology affect changes in level of suicidal ideation.
+Overall, this study makes a timely contribution to the suicide literature. As the U.S. population continues to increase in cultural diversity, a more “inclusive” understanding of suicide risk is needed. The outcome of this study provided empirical evidence for the negative impact of acculturative phenomena and low ethnic identification differentially for African American and European American college students. Moreover, the data indicated that vulnerability toward suicidal ideation was associated with acculturation-related distress and insufficient group identity. Both quantitative and qualitative investigations should fully explore culturally relevant phenomena in suicide risk. Because African Americans’ patterns of suicide defy conventional models of suicide risk, investigations of culturally relevant factors are fundamental to studies of African American suicidal behavior.

An Exploration of the Relationship Between Spirituality, Religion and Mental Health Among Youth Who Identify.txt

@@ -0,0 +1,56 @@
+Background
+There is a growing interest in addressing spirituality and religion in health care, with evidence emerging that personal spiritual and religious practices, and support of these by practitioners, can influence mental health in a positive way. Spirituality is understood, in this context, as a search for connectedness and meaning, transcendence and belonging. Being religious is conceptualised as the outward practice
+of spiritual beliefs situated in a particular organised religion (Moreira-Almeida et al. 2016). For youth who identify as lesbian, gay, bisexual, transgender plus [plus other minority sexual groups] (LGBT+), there are distinct challenges to spiritual and religious expression (Liboro 2015). Negative thoughts or experiences arising from LGBT+ youths’ personal religious beliefs (Hamblin and Gross 2013), attitudes from others or cultural experiences of historical religious beliefs, can affect the youth’s self-perception in a negative way leading to mental health issues (Liboro 2015). Conversely, there is the potential for religious or spiritual beliefs to provide both personal and community support in a positive manner (Tenner 2015). Given a young LGBT+ person’s vulnerability at a significant time of their life, developing an understanding of the implications of spirituality and religion for this group is essential. The success of health-care agencies towards this population can therefore depend on their capacity to subscribe to a spiritual approach, routinely assessing spiritual needs or having a religious understanding (Kralovec et al. 2014).
+There is evolving evidence indicating that the expression of spirituality and religion can have a positive impact upon people’s lives, with some studies supporting the assertion that having a faith can lead to better mental health outcomes (Ream and Savin-Williams 2005; Rodriguez 2009). In a review of the literature addressing the relationship between the incidence of mental disorders and religion or spirituality in the general population, a significant number of studies (72.1%) reported positive outcomes between spiritual or religious involvement and mental illness including depression, substance use and suicide; far less (4.7%) showed negative results (Bonelli and Koenig 2013). However, the potential positive influence of religion and spirituality on mental health has been the subject of debate. King (2014) concluded that those who described themselves as spiritual (but not religious) appeared to be more vulnerable to psychological issues, suggesting that ‘those with a spiritual view of life appeared to be vulnerable to mental and substance misuse disorders’ (King 2014:109).
+There has been a major about turn in attitudes towards people with same-sex attractions and a greater acknowledgement of the religious and spiritual lives of people who identify as LGBT+ (Gibbs and Goldbach 2015). Still, LGBT+ youth can face significant challenges in establishing a sense of identity in a predominately heterosexual and transphobic world (Matthews and Salazar 2012). Many LGBT+ youth have to face many challenges alone without the support of family and peers and often in hostile and unsupportive environments (Levy and Edmiston 2014; McCann et al. 2019).
+Support from religious organisations may be helpful in challenging and stressful times. However, non-affirming societal beliefs around same-sex intimacy or gender identity can exacerbate minority stress and internalised homophobia (Meyer 2003; Rostosky and Riggle 2017). The minority stress model demonstrates the potential psychosocial stressors related to being LGBT+ and the damaging effects on health and well-being. Negative societal responses can lead to feelings of guilt, shame, demoralisation, low self-esteem and social exclusion (Meyer 2003; Lease et al. 2005; Rosario et al. 2006). This phenomenon has been associated with a significant increase in depression, anxiety and suicidal thoughts and behaviours. There are also strong links with substance use and eating disorders (Barnes and Meyer 2012).
+Despite the identified mental health challenges, affirming religious beliefs have been shown to be a protective factor in counteracting harmful stressors among sexual minority youth (Wilkinson and Pearson 2009; Barnes and Meyer 2012; Foster et al. 2011). Coming out, a significant time for LGBT+ youth, may lead to rejection from their spiritual or religious community. The resultant existential conflict can lead to increased anxiety and depression (Gibbs and Goldbach 2015). Young people may become more distant from their family of origin through moving away from their spiritual faith thus limiting access to emotional support during times of need (Barnes and Meyer 2012). By re-examining spirituality, youth may develop coping strategies and resilience through a renewed sense of faith and finding affirmative people and spiritual communities that are open, supportive and can validate expressions of sexuality and gender identity and encourage good mental health (Koenig 2009). However, the picture remains incomplete; hence, the current systematic review focuses on the subjective experiences of LGBT+ youth regarding spirituality and religiosity that may guide and inform future mental health practice and service developments.
+Methods
+The aim of this review was to synthesise current evidence regarding the experiences and perceptions LGBT+ youth regarding the expression of their spirituality and their mental health needs. Therefore, the questions of this review are:
+1.	What are the experiences and mental health needs of youth who identify as LGBT+ regarding religion or spirituality?
+2.	What are the implications for mental health services in relation to the religious/ spirituality needs of youth who identify as LGBT+ ?
+Search and Selection Strategy
+A subject librarian assisted with the literature search strategy. The databases used in the search were CINHAL, MEDLINE, PsychINFO and Sociological Abstracts. The search terms used were: spiritual* OR relig* OR sacred OR transcendent AND GBLT OR gay OR lesbian OR bisex* OR trans* OR intersex OR queer AND mental health OR psychosocial OR well-being OR self-esteem OR homonegativity AND youth OR adolecen*. The inclusive dates were 31 May 2008 to 1 June 2018 to best capture contemporary mental health practices and individual experiences in the changing sociopolitical climate for youth who identify as LGBT+. The search strategy utilised in one of the electronic databases is contained in Table 1.
+The search yielded 314 hits in total. Following the removal of duplicates and a check for relevance, 44 papers remained. Full texts of papers were screened leaving 10 papers suitable for the review. To be included in this review studies had to be empirical peer-reviewed research in English and focus on mental health and spirituality or religious experiences of youth up to the age of 25 years who identified
+as LGBT+. Studies not meeting the criteria were excluded. Reasons for exclusion included wrong population, wrong subject or failed to address the research questions (Fig. 1).
+Quality Assessment
+A quality assessment tool was used to review the papers (Critical Appraisal Skills Programme 2018). Specific questions were applied to each of the relevant studies (Table 2). Each question was scored zero, one or two out of a score of 20. A score of zero was given if the paper had no information, one if there was a moderate sum, and a score of two if the question was fully addressed (Rushbrooke et al. 2014). A score of 17 and above, demonstrating a high-quality study, was achieved by three of the studies (Gattis et al. 2014; Page et al. 2013; Quinn et al. 2016). A total of five studies scored between 14 and 16, indicating deficits in the clarity of aims, data collection methods, research relationships considered and ethical considerations (Eick et al. 2016; Gold and Stewart 2011; Jeffries et al. 2014; Kubicek et al. 2009; Lauri-cella et al. 2017). The remaining two studies scored below 14, due to limited information that impacted on the overall quality and were related to the aims, ethics, and clarity and detail of findings (Hatzenbuehler et al. 2012; Nielson 2017). All of the studies were included in the review as they met the study inclusion criteria.
+Characteristics of the Selected Studies
+The ten studies that addressed the review questions are presented in Table 3. The majority of studies (n = 9) were conducted in the USA, with the remaining study carried out
+in Israel. The studies had sample sizes ranging from 1 to 1413 participants. The age of youth participants ranged from 12 to 25 years (n=7). Five of the studies used quantitative methods, two studies used qualitative methods, and two were mixed methods studies.
+Data Extraction and Analysis
+The process of data analysis and synthesis involved the extrapolation of themes that addressed the aims of the research. These were coded from the results of the included studies, organised according to concepts and verified and agreed by the research team (Caldwell et al. 2011).
+Findings
+The aim of this systematic review was to consider empirical studies regarding the spirituality and religious experiences of LGBT+ youth regarding the expression of their sexuality and their mental health needs. Following data analysis, three main themes emerged. These were (1) attitudes and beliefs; (2) individual sexuality experiences; and (3) spirituality as coping and support.
+Attitudes and Beliefs [Discrimination, Gender Differences and Shame]
+Adolescence is a crucial time in the formation and development of a person’s sexual and religious/spiritual identity. It is often a period of experimentation and of testing one’s own beliefs and ideas and engaging in critical reflection on life’s possibilities and future directions. The situations where these experiences may be carried out can present challenges, particularly in perceived heterosexist environments. Some of the studies included in the review identified schools as potentially discriminatory and stressful environments where homophobia, biphobia and transphobia often exist (Eick et al. 2016; Gattis et al. 2014). Victimisation experiences, including bullying, shaming and violence, can lead to poor academic performance, motivation and attendance. The challenges faced by LGBT+ youth can also lead to higher rates of anxiety, depression, suicidality, substance use and prostitution than in the heterosexual population (Gattis et al. 2014).
+In one study, addressing prejudice and stereotyping towards homosexual students in Israeli schools using contact interventions (Allport 1954), there were improvements in student emotional, cognitive and behaviourial attitudes (Eick et al. 2016). This study by Eick was a mixed student population of straight and LGB youth and the improvements concerned this whole sample. Studies that examined the relationship between religion, mental health and internalised homophobia in LGBT+ youth found that belonging to a religious denomination that was gay affirming can act as a protective factor for discrimination and depression (Page et al. 2013). Conversely, where homonegativity prevails, in the form of discrimination, stigma and persecution, there can be a disintegration/dissonance between religiosity and sexuality. Tensions can often exist creating feelings of incompatibility, imbalance and doubt. Individuals may feel alienated, isolated and marginalised through the discrimination displayed by some religious organisations (Gattis et al. 2014; Page et al. 2013; Quinn et al. 2016). Further conflict can exist between religion, spirituality and sexual identity (e.g. ‘reparative therapy’). As a result, LGBT+ youths can be wary of ‘organised’ or established religious groups (Gattis et al. 2014). In the Black Church, where the dominant position was non-LGBT+ affirming, LGBT+ people tended to be ‘closeted’ and sexually secretive to cope with the challenges of homonegativity. However, due to social, political and family centrality, Black LGBT+ members often remained active in the church. Religion and spirituality remained prominent in young Black men’s lives despite heteronormativity. Some study respondents thought that challenging the negative views of clergy towards LGBT+ congregation
+members was futile. Although some commentators agree that stigma reduction strategies can reduce internalised homophobia, increase self-esteem and reduce depression and isolation in LGBT+ youth, many felt let down and had to eventually leave faith communities (Gattis et al. 2014; Quinn et al. 2016).
+Individual Spirituality Experiences [Conflict, Oppression, Identity Formation]
+A conflict was found to exist between sexual and spiritual identity and religious teachings about LGBT+ concerns. This tension appeared to lead to the LGBT+ community becoming increasingly marginalised from many faith-based communities. Approximately 90% of more than a dozen nationally representative survey respondents described present-day Christianity as anti-homosexual (Barnes and Meyer 2012). Perhaps as a reaction to this, or as a means of coping, many youth who identify as LGBT+ have dissociated from non-affirming religious institutions. The conflict between religion and sexuality is strongly associated with internalised homonegativity and poor mental health (Lauricella et al. 2017; Page et al. 2013). Early on in a person’s sexual development, LGBT+ youth are often not able to clarify their sexual orientation, may have little or no contact with the LGBT+ community and often get involved in religious activities as a way of suppressing their own desires (Lauricella et al. 2017). Later in their development, some of these individuals may still hold on to feelings of shame concerning their sexual identity and try to eliminate their urges through prayer and other means. Some people, however, may go on to find a more accepting spiritual community or find other ways of reconciling their sexuality with their childhood religion (Lauricella et al. 2017).
+Oppression is a social construct that creates the closet in which LGBT+ people reside either partially or fully (Rhodes 1994). It is recognised as the place between self-identifying as gay and disclosing one’s sexual orientation to others. In a webbased survey of 47 respondents, Gold and Stewart (2011) explored how LGB undergraduate students negotiated and defined their spiritual identities during this coming out process. When attempting to navigate their burgeoning sexual identity with that of their spiritual identity, students spoke of experiences of irreconciliation, progressive development, arrested development, completed development and reconciliation. The authors considered that it was through these processes that the individuals were able to begin to negotiate and in turn construct their own new internal identities (Gold and Stewart 2011).
+Spirituality as Coping and Support
+Spirituality has been described as acceptance and ‘loving kindness’. It can involve personal relationships with a powerful essence, a strong connection to nature and a respect for all forms of life. It has to do with love, understanding and compassion. There may be a closeness to a higher being or ‘god’ (Gold and Stewart 2011). Although there is the beginning of a sea change in the psychology of religion, with an increased acknowledgement of the religious and spiritual lives of people who identify as LGBT+, there still persists a need for more evidence-based research
+1 Springer
+for young people coming to terms with their sexuality and exploring their religious beliefs (Page et al. 2013; Ream and Rodriguez 2014). The anti-homosexual stance previously held by organised religious groups, however, may be changing, since a 2011 survey found that 58% of respondents believed that society should accept homosexuality (Pew Research Center 2011).
+Hatzenbuehler et al. (2012), exploring religion and health risk behaviours, identified that religious climate among youth who identify as LGB was a predictor in excessive alcohol use and risky sexual behaviour. The study demonstrated that LGB youths living in countries with more supportive religious climates showed fewer health risk behaviours, meaning religion can also be protective factor for LGB youths. The authors highlighted the need to develop prevention intervention programmes for LGB youth living in high-risk environments, in particular youth living in unsupportive religious climates. Similarly, Jeffries et al. (2014) advocated the need to consider factors involving religion and spirituality in young HIV-infected men as a way to help tailor appropriate interventions for this population and help enhance faith-based practice developments.
+In another study investigating individual resilience experiences, Kubicek et al. (2009) explored the role of religion and spirituality in the lives of a sample of young gay men and looked specifically at how homophobic messages taken from religious contexts were internalised by this group. This unique mixed methods study looked at how these young men attempted to resolve the conflict between these messages and their sexual identity and discovered how they had made a number of important conscious choices about their lives, including religious and spiritual beliefs in an effort to solidify their identity. The study describes their experiences and processes in identifying the positive and nurturing aspects of religion such as feeling a sense of acceptance and support from a higher power. The group at times had to reframe or simply reject the negative messages they had heard whilst growing up which had the effect of incorporating a stronger sense of spirituality into their lives. It is important to note that for the participants of this study, they relied on the belief that sexual orientation as an innate and unchangeable aspect of their selves.
+Discussion
+The development of a LGBT+ sexual identity is a complex and often difficult process. This review has demonstrated both the positive and negative experiences of LGBT+ youth in relation to faith-based or spiritual upbringing. Important issues have been raised and will now be discussed further through the implications for practice, education and future research.
+Implications for Practice
+The World Psychiatric Association proposes that full consideration should be given to spirituality in holistic assessments, that is, the biopsychosocial, cultural and spiritual elements (Moreira-Almeida et al. 2016). Despite this, there is no evidence of
+formal training about spiritual elements in the education and training of mental health practitioners (Castaldelli-Maia and Bhugra 2014; Schuck and Liddle 2001), or any formalised, recognised way of going about this. The underlying principles should be person-centred approaches to care, supports and treatment including respect, sensitivity and curiosity for spirituality experiences. Practitioners should be able to demonstrate awareness, respect and sensitivity to peoples’ spiritual experiences and beliefs. Furthermore, clinicians should be aware of the potential benefits and the harm of religious, spiritual and secular world views. The importance of maintaining a strong sexual and gender identity for LGBT+ youth and the development of resilience is indicated in the current review (Page et al. 2013). Parents, teachers and mental health practitioners can focus on relevant stressors that are evident in the lives of LGBT+ youth and provide the necessary supports and psychosocial interventions that may promote greater resilience and coping strategies for LGBT+ youth. Also, there needs to be more collaborative work with faith leaders to support LGBT+ youth and their families (Moreira-Almeida et al. 2016; Page et al. Rodriguez 2009). Given that spirituality and religion can be sensitive issues, there needs to be training and education to underpin any such practice (Lease et al. 2005).
+Implications for Education
+Schools are important places to address discrimination, prejudice and victimisation. The review has revealed that through positive and supportive environments, where negative attitudes and beliefs were challenged, knowledge, beliefs and attitudes improved (Eick et al. 2016). Non-LGBT+ affirming religions have been associated with greater internalised homonegativity (Barnes and Meyer 2012), emotional distress (Wilkinson and Pearson 2009) and poorer self-esteem (Ream and Savin-Williams 2005). There is also a strong association between religion, mental health and minority stress (Newcomb and Mustanski 2010). There is a need for more evidencebased research into young people coming to terms with their sexuality and exploring their religious beliefs (Ream and Rodriguez 2014), and greater exploration of potential for tolerance and acceptance among religious communities (Park et al. 2016). There needs to be a recognition and development of multicultural competencies. Educational and training initiatives should contain sexuality and spirituality components for holistic practitioners (McCann and Brown 2018). Raising awareness and increasing knowledge of the ways that social identities can influence students who are searching for meaning and purpose in their lives is important. Reflection and increased dialogue around acceptance and tolerance should be supported and encouraged (Park et al. 2016). There should be appropriate spaces in campus for exercising spiritual activities such as meditation, prayer and reflection.
+Implications for Future Research
+The review has identified several areas where more research is needed to better support youth who identify as LGBT+. Whilst empirical research has identified significant links between spirituality, religion and health, there needs to be more
+1 Springer
+LGBT+-specific research establishing needs and evaluating potential interventions (Hill and Pargament 2003; Castaldelli-Maia and Bhugra 2014; Yip 2008), and also the potential pitfalls related to youths’ experiences of religion and spirituality and how this might have a negative influence. Some of the emerging issues that require further investigation are the prevalence of depression and substance use among LGBT+, and the relationship of spirituality and religion to the manifestation of these issues. Spiritual understandings and experiences also need to be taken into account during diagnosis, so that accurate account of spiritual and religious views or effects from these are clearly articulated and understood, rather than being categorised as a constituent element of mental disorder. Spiritual distress, for example, which manifests in a feeling of a lack of meaning in life, could be mistaken for depression. Spiritual distress has been classified as a nursing diagnosis in NANDA International (NANDA-I) since 1978. It is defined as a ‘state of suffering related to the impaired ability to experience meaning in life through connectedness with self, others, world or a Superior Being ’ (Herdman and Kamitsuru 2014: 372).
+More research needs to be carried out around suggested interventions and treatments that may have a spiritual element such as self-help groups; religious communities; talking therapies; mindfulness; and tai chi, for example. Overall, more needs to know about the potential of addressing and supporting LGBT+ youths’ spiritual needs and its effect on their overall outcome (recovery and staying well if they are diagnosed with mental health issues) and also prevention of mental health issues. If there is potential for spiritual interventions to improve quality of life and well-being, then more needs to be done to explore this possibility in this cohort. What has become increasingly apparent from this systematic review is the distinct lack of empirical research that specifically addresses the psychosocial experiences of LGBT+ youth with regard to spirituality and religiosity. There are opportunities to conduct multi-centre, international and longitudinal research studies that utilise a range of methodologies and designs that will contribute significantly to the evidence base and increased understanding of the relevant issues and concerns for LGBT+ youth.
+Strengths and Limitations
+There is an increasing interest in the spirituality and religious experiences of youth who identify as LGBT+ and its importance in relation to mental health and psychosocial well-being. This systematic review has revealed valuable sources of information that may guide practitioners, service providers, educators and researchers. There are limitations in the studies included in this review primarily due to the relatively small sample sizes, the robustness of some of the study designs and the absence of intervention and evaluation studies. The authors have attempted to exercise rigour in their selection of studies and have utilised relevant frameworks and methodological strategies throughout to address these issues.
+Conclusion
+Practitioners need to be aware of and sensitive to individual religious and spirituality issues. Negative experiences of religious institutions may affect self-perceptions and a willingness to engage in healthy behaviours. Religious and spiritual activities may help with negative coping behaviours such as drug use, risky sex and prostitution. Spiritual coping may promote better mental health and increase self-esteem. It may support healthy living and help motivate people to make positive changes in their lives. Support for marginalised groups should be a pivotal point for all churches and religious institutions that are open, non-judgemental and accepting of all, and given the potential (positive or negative) influence of spirituality on the LGBT+ youth, particularly in relation to their mental health (Kralovec et al. 2014), it is important that health researchers lead the way in promoting this support and providing a distinct evidence base to support it.
+Funding No funding was received for this project.
+Compliance with Ethical Standards
+Conflict of interest The authors declare that they have no conflict of interest.

Annual Research Review A meta-analytic review of worldwide suicide rates in adolescents.txt

@@ -0,0 +1,84 @@
+Introduction
+Suicide is a leading cause of death worldwide. Current estimates indicate that an individual will die by suicide somewhere in the world every 40 s (World Health Organization (WHO), 2014). This public health concern is perhaps even more alarming and puzzling when it comes to suicide death among youth -estimated to be the second leading cause of death among young people 10-24 years old (Centers for Disease Control and Prevention (CDC), 2017b; Patton et al., 2009; WHO, 2014). The purpose of the current review is to provide a recent estimate of worldwide suicide mortality rates in adolescents and to examine cross-national trends in these rates. Extending prior research, this study explores suicide mortality data in detail, including patterns in suicide methods, how access to lethal means relates to suicide rates, and how suicide rates vary cross-nationally based on indices of economic quality and inequality.
+Although the specific causes of suicide among young people are complex and remain somewhat elusive (Bridge, Goldstein, & Brent, 2006; Cha et al.,
+2018; Hawton, Saunders, & O’Connor, 2012; Tur-ecki & Brent, 2016), it is clear that suicide is a major public health concern among adolescents. Suicidal thoughts and behaviors are relatively rare during childhood but increase significantly during the transition to adolescence (Dervic, Brent, & Oquendo, 2008; Hepp, Stulz, Unger-Koppel, & Ajdacic-Gross, 2012; Nock, Borges, Bromet, Cha, et al., 2008; Nock et al., 2013). In addition to the increased prevalence during adolescence, there is also significant escalation from suicidal thoughts to suicidal behaviors during this developmental period. Most youth who transition from suicidal thoughts to suicidal behaviors will do so within 1-2 years after the onset of suicide ideation (Glenn et al., 2017; Nock et al., 2013). Moreover, available country-level estimates suggest that the suicide rate among adolescents has increased in recent years (OECD, 2017b). For example, in the United States of America (USA), the age-adjusted suicide rate increased by 24% from 1999 to 2014; the increase in rates for females was greatest among those aged 10-14 years, while males aged 10-14 years experienced the second largest percent increase among males during this time (Curtin, Warner, & Hedegaard, 2016). Among 15- to 19-
+year-olds, suicide rates increased for both sexes from 2007 to 2015; among females, the rate in 2015 was higher than any time in the prior 40 years (Curtin, Hedegaard, Minino, Warner, & Simon, 2017). Taken together, adolescence is a key developmental period for effective suicide intervention and prevention (Gould, Greenberg, Velting, & Shaffer, 2003; NAASP, 2014; WHO, 2014; Wyman, 2014).
+A number of prior studies have estimated crossnational trends in suicide mortality rates among youth. Most of this previous research has used the World Health Organization’s (WHO) Mortality Database (WHO, 2018b), which provides one of the best sources of information about worldwide mortality rates. Using this database, Wasserman, Cheng, and Jiang (2005) estimated a worldwide suicide rate for 15- to 19-year-olds of 7.4/100,000 people based on suicide mortality data collected in 1995 from 90 countries. The highest suicide rates in youth have been observed in New Zealand (Bridge et al., 2006; Chaet al., 2018; Kõlves & De Leo, 2016; McLoughlin, Gould, & Malone, 2015; Roh, Jung, & Hong, 2018), Finland (Bridge et al., 2006; Cha et al., 2018; McLoughlin et al., 2015; Roh et al., 2018), Ireland (Bridge et al., 2006; McLoughlin et al., 2015), Guyana (Koõlves & De Leo, 2016), Sri Lanka (Wasserman et al., 2005), and a range of former Soviet Union states (Bridge et al., 2006; Cha et al., 2018; Koõlves & De Leo, 2014, 2016; McLoughlin et al., 2015; Roh et al., 2018; Wasserman et al., 2005). In addition, suicide rates are found to be higher in older versus younger youth (Bridge et al., 2006; Cha et al., 2018; Roh et al., 2018). Finally, adolescent suicide deaths are much more common (2-4x higher) in males than females (Bridge et al., 2006; Cha et al., 2018; Kõlves & De Leo, 2014, 2016; McLoughlin et al., 2015; Roh et al., 2018; Varnik et al., 2009; Wasserman et al., 2005), consistent with sex differences in suicide rates observed among adults (Bachmann, 2018; Canetto & Sakinofksy, 1998; Chang, Yip, & Chen, 2019; Nock, Borges, Bromet, Cha, et al., 2008; Schrijvers, Bollen, & Sabbe, 2012; World Health Organization (WHO), 2016b). The major exceptions to this sex difference in youth have been observed in China (Bridge et al., 2006; McLoughlin et al., 2015; Wasserman et al., 2005), India (McLoughlin et al., 2015), Cuba(Wasser-man et al., 2005), Ecuador (Wasserman et al., 2005), El Salvador (Wasserman et al., 2005), and Sri Lanka (Wasserman et al., 2005), all of which have reported higher suicide rates among females than males in at least one study.
+The current review provides an updated estimate of worldwide suicide mortality rates among youth, aged 10-19 years, and examines cross-national trends in suicide rates. Like most prior studies examining worldwide suicide rates, this review uses the WHO Mortality Database. The present review builds on prior studies in three important ways. First, this review examines the time period from early to late adolescence (10- to 19-year-olds) and compares rates
+among younger (10- to 14-year-old) and older (15- to 19-year-old) adolescents. The current focus on this age range is critical, as most prior studies have examined either narrow age ranges (e.g., 1519 years) that leave out key periods of adolescence, or wider age ranges extending into early adulthood (e.g., 5-29 years). Prior research reveals marked differences in the incidence of, and risk factors for, suicide-related outcomes across adolescent and adult developmental periods (Koõlves & De Leo, 2015; Lewinsohn, Rohde, Seeley, & Baldwin, 2001; Nkansah-Amankra, 2013), highlighting the need to more precisely examine the period of adolescence. Second, this review provides a more recent estimate of suicide rates by focusing on data since 2010. The majority of prior reviews have examined suicide rates over the past 15-20 years. Given the significant changes in suicide death rates over time (Curtin et al., 2016; OECD, 2017b), an updated review is needed. Third, this review considers a number of cross-national trends in suicide mortality rates. In addition to examining cross-national suicide rates as a function of age and sex, this review examines crossnational trends in specific suicide methods (Hepp et al., 2012; Kõlves & De Leo, 2017; Varnik et al., 2008, 2009), how rates vary based on access to lethal means (e.g., firearms and railways), and how rates vary as a function of economic quality and inequality (Bachmann, 2018; Shah, 2012).
+Method
+Search strategy for suicide mortality data
+In line with prior studies (Bachmann, 2018; McLoughlin et al., 2015; Nock, Borges, Bromet, Cha, et al., 2008), we used two main strategies to obtain cross-national data on suicide deaths in 10- to 19-year-olds: (a) We accessed publicly available data sources of either cross-national or country-specific mortality data, and (b) we conducted a systematic review of empirical studies reporting national or cross-national suicide death data.1 The resulting source for suicide mortality data was the WHO’s Mortality Database (WHO, 2018b). The WHO Mortality Database (last updated May 2018) collects mortality and vitality statistics directly from nations’ civil registration systems across the world and presents standardized mortality data by age, sex, year, and cause of death (coded according to the International Classification of Diseases, 10th revision [ICD-10; WHO, 2016a]). Available national data are categorized by data quality. Although the WHO recognizes 194 member states as of 2016, the completeness of data coverage varies by country, and several countries do not submit mortality statistics to the WHO. Developed countries more consistently report annual and complete data than developing nations, which often submit partial data covering subnational regions.
+We restricted our use of the WHO database in three ways. First, we only included countries that had data available since 2010 in order to provide the best estimate of recent suicide rates. Data for the most recent year available for each country were included in this review. Second, since the present review focused on examining suicide death data in detail, we only included data for countries that were evaluated as ‘high’ quality (e.g., identifying ICD codes for the vast majority of suicide deaths), and excluded countries with ‘medium’ and ‘low’ quality data. Death registration data quality
+classifications are based on three indices: (a) whether mortality data is submitted by ICD code, (b) whether mortality data has been submitted for multiple years, and (c) average usability of data submitted since 2007 (WHO, 2018a). Usability scores account for the proportion of reported deaths that are assigned to a poorly defined ICD death code. As of 2016, a nation’s data are considered ‘high’ quality if that nation has supplied at least 5 years of data since 2007 that have achieved an average usability score of 80% or higher. A classification of ‘medium’ quality denotes that a nation’s mortality data have an average usability score between 60% and 80%. ‘Low’ and ‘very low’ quality indicate usability scores below 60% and 40%, respectively.
+Based on these criteria, 45 countries were included in this review2 : Africa (n = 1), Asia (n = 6), Europe (n = 28), North America (n = 6), Oceania (n = 2), and South America (n = 2; see Table 2). For each country, we extracted suicide mortality data for 10-to to 19-year-olds using the WHO Cause of Death Query Online (CoDQL) tool. When available, we also extracted the following suicide mortality data: (a) age group: 10- to 14-year-olds and 15- to 19-year-olds, (b) sex: male and female, and (c) suicide method. WHO data provide method of suicide death based on the International Classification of Diseases, 10th revision (ICD-10; WHO, 2016a) codes X60-X84 indicating selfinflicted death3 (see Data Analysis section).
+Population data
+Population estimates were extracted from the United Nations (UN) Population Division’s World Population Prospects 2017 database (UNPD, 2017), which provides population data for all countries included in this review. The World Population Prospects database compiles national census data and data from specialized population surveys to provide population estimates by country, age, and sex (UNPD, 2017). Given that several countries do not report population estimates directly to the WHO, the WHO Mortality Database only provided total population data for approximately half of the countries included in this review. To calculate age-standardized death rates for nations that do not regularly report population data together with vital registration data, the WHO collaborates with the UN Population Division to collect global health statistics and population totals (WHO, 2018a). When possible, WHO population estimates and UN estimates were compared by year and age and were found to be nearly identical.
+Access to lethal means
+Data on access to lethal means were obtained from publicly available datasets. For train- and firearm-related suicide death, access to means was operationalized as density of means (i.e., railways and firearms), either per geographic area or persons.
+Railway density data (km of lines per 1,000 km2) per country were obtained from the International Union of Railways (International Union of Railways (UIC), 2016). Countries were categorized into one of seven density ranges (lowest density = 0-5 km of lines per 1,000 km2; highest density > 75 km of lines per 1,000 km2). Although railway density information was not obtained for the same year as the mortality data, these estimates have increased only marginally (3.6%) over the past decade (UIC, 2016) and thus remained relatively stable during the period of data collection for this study.
+Firearm accessibility was measured as the estimated number of civilian firearms per 100 persons, with data for each country obtained from the Small Arms Survey (Small Arms Survey, 2018). It is important to note that although the proportion of suicides deaths via firearm is sometimes used as aproxy for gun ownership (Alvazzi del Frate & Pavesi, 2014), the Small Arms Survey did not use suicide by firearm in
+calculating rates of civilian firearm ownership per country (Karp, 2018). Therefore, the firearm access and suicide death by firearm variables are independent, allowing their association to be examined. However, unlike railway estimates, firearm density has increased significantly over the past decade (estimated increase of 32% from 2006 to 2017 due to enhanced research methods and increased civilian holdings; Karp, 2018) and therefore was not a stable estimate over the data collection period for this study.
+Urban population data (percentage of a country’s total population living in urban areas in 2018) were obtained online from the United Nations Population Division (UNPD, 2018). These data were used as a proxy for access to tall structures, or heights, for jumping.
+Economic quality and inequality
+Included countries were classified by economic level according to the World Bank Income Groups (2019). These groupings are determined by the gross national income (GNI) per capita, reflecting the average income of a country’s citizens. Groups are defined as high-income (>$12,506 in US Dollars), uppermiddle-income ($3,896-$12,055), lower-middle-income ($996-$3,985), and low-income (<$995) (World Bank Group, 2019). Previous studies have used the World Bank Income Groups to examine how a country’s economic quality relates to mental health outcomes (Ayuso-Mateos, Nuevo, Verdes, Naidoo, & Chatterji, 2010; Bromet et al., 2011; Nock, Borges, Bromet, Alonso, etal., 2008; Stein et al., 2010). The World Bank Income Group ratings were obtained for the same year as the most recent available WHO mortality data for each country.
+Economic inequality was measured with the Gini index, or Gini coefficient, which measures income distribution in a country and is the most commonly used measure of economic inequality. The Gini index is assessed on a scale of 0 to 1, with 0 representing the least possible amount of inequality and 1 representing the greatest possible inequality in a country (Subramanian & Kawachi, 2004). Prior research has used Gini coefficients to compare how economic inequality relates to a range of psychiatric disorders (Burns, Tomita, & Kapadia, 2014; Cifuentes et al., 2008; Johnson, Wibbels, & Wilkinson, 2015; Yu, 2018). For the current study, Gini coefficients were obtained from the World Bank (2019) for the most recent available year. Although this may not align with the same year as the WHO mortality data for a given nation, Gini coefficients have been relatively stable over time (Li, Squire, & Zou, 1998).
+Data analysis: estimates of suicide mortality
+Pooled estimates. To estimate the pooled suicide mortality rates across all available countries (n = 45), we used the ‘metafor’ R package (Viechtbauer, 2010) to conduct a series of random-effects meta-analyses. Because suicide is a relatively rare event, there were several instances, especially pertaining to subgroups (e.g., females 10-14 years old), for which there were no suicide deaths. To account for the existence of these cases, we used the Freeman-Tukey transformation (Freeman & Tukey, 1950), which allows for proportions that equal 0. We calculated pooled estimates for suicide death by all methods, cross -tabulated by age group (10- to 19-year-olds, 10- to 14-year-olds, 15- to 19-year-olds) and sex (males and females combined, males only, females only). Each analysis produced an estimate of prevalence, which we standardized to prevalence per 100,000 people, as well as a 95% confidence interval for the estimate. The meta-analysis also produced two metrics of heterogeneity: the I2 statistic, which quantifies the percent of variability across cases that is not due to chance, and a Q statistic, which, when significant, reflects a
+high level of heterogeneity between cases (Higgins & Thompson, 2002).
+Estimates by country. For country-level data, we calculated the mortality rate for each country, standardized to suicide deaths per 100,000 people. Given that we were interested in country-to-country differences, we did not use meta-analysis. In line with recommendations for reporting mortality rates (United States Department of Health, 2018) and consistent with prior reviews (Kolves & De Leo, 2017), we excluded from analyses any cell with fewer than 10 events (i.e., suicide deaths).4 Therefore, of the 45 countries with data available for suicide by any method, analyses included anywhere from 10 to 37 countries (M = 21.78 countries, SD = 11.30). When examining estimates by country and by method, there were ultimately fewer countries included due to the possibility that there were 0 suicides by any given method. We calculated statistics cross-tabulated by age group (10- to 19-year-olds, 10- to 14-year-olds, 15- to 19-year-olds) and sex (males and females combined, males only, females only). We also calculated the ratio of suicide mortality rates by males and females.
+Suicide methods. We created higher-level groupings of suicide methods based on ICD-10 codes (WHO, 2016a), leading to a total of nine groups of methods: (a) self-poisoning, including drugs, medications, solvents, gases, and pesticides (codes X60-X69); (b) hanging/suffocation (code X70); (c) drowning (code X71); (d) firearms (codes X72-X74); (e) explosion, fire, steam, or hot objects (codes X75-X77); (f) sharp or blunt objects (codes X78-X79); (g) jumping from a height or jumping/lying in front of a moving object (codes X80-X81), which were combined because counts were too small for each code to be examined separately (referred to collectively as ‘jumping/lying’ from this point forward); (h) motor vehicle (code X82); and (i) other/unspecified methods (codes X83X84). Using meta-regression, we calculated deaths per 100,000 people for males and females together, as well as males and females separately.
+Data analysis: moderators of suicide mortality
+Economic quality. To explore whether suicide rates differed by income group across all countries, we conducted a moderated meta-analysis (i.e., meta-regression with dummy variables) based on the recommendations provided by Viecht-bauer (2010). As with the other meta-analyses performed, because meta-analysis is robust to very infrequent event counts, we included in the analysis any country with available data, even if they did not have more than 10 suicide deaths. Of the 45 countries with income group data, 34 were high, nine were upper-middle, and two were lower-middle. Given that we did not want to have two countries drive the moderated metaanalysis, we combined the lower- and upper-middle-income countries into one ‘middle-income’ group. We were also interested in whether the ratio of male:female suicides differed by income group. To explore this, we conducted a t-test using the male:female suicide death ratio as the outcome and income group as the predictor, in all adolescents (10- to 19-year-olds, among countries with >10 suicides) and 15- to 19-year-olds (also among countries with >10 suicides). We did not examine this relationship for the 10- to 14-year-old group because there were too few countries with more than 10 suicides (n = 8) to make a meaningful inference about the data.
+Economic inequality. To explore whether suicide rates differed by economic inequality, we conducted a set of analyses similar to those for economic quality, but used the Gini coefficient instead of the economic quality group. Because the Gini coefficient, and therefore economic inequality, is a
+continuous variable, these analyses differed from those for economic quality in two ways: (a) The moderated meta-regression did not use dummy codes and (b) we conducted a correlation between Gini coefficients and male:female ratios instead of t-tests.
+Access to lethal means. We examined how suicide methods varied as a function of lethal means access -specifically firearms (number of firearms per 100 people), railways (rail density per 1,000 km2), and access to tall structures (% of individuals residing in urban areas). For each of the three moderators, we calculated a series of moderated meta-regressions for each method (i.e., separate models for each method) across all ages and sexes, rather than separately by age groups and sex in order to avoid potential type I errors as a result of multiple comparisons. In these cases, a significant Q statistic indicated the presence of a moderation effect.
+Results
+Suicide mortality
+Pooled estimates. Table 1 displays the pooled suicide rates for adolescents by age group and sex. The pooled suicide rate across all sexes, 10- to 19-year-olds, was 3.77 per 100,000 (95% CI = 3.15-3.45, I2 = 96.87%, Q = 1,587.92, p < .001). There was considerable heterogeneity across analyses, reflecting the variability in cross-national suicide rates and
+pointing to a need for caution in interpreting these findings.
+Estimates by country. Table 2 displays the country-level descriptives by age group and sex. The rates for all ages (10- to 19-year-olds) inclusive of males and females ranged from 1.31/100,000 people (Israel) to 9.72/100,000 people (Estonia). The rates for 10- to 14-year-olds, both sexes (20/35 countries had <10 cases and were excluded), ranged from 0.28/100,000 people (United Kingdom) to 4.71/ 100,000 people (Kyrgyzstan). The rates for 15- to 19-year-olds, both sexes (0 countries had <10 cases), ranged from 2.30/100,000 people (Israel) to 17.6/ 100,000 people (New Zealand).
+Pertaining to the ratio of males to females who died by suicide, males were more likely than females to die by suicide in all countries except Uzbekistan (where the male:female ratio was 0.95). In all other
+countries across both age groups, the male:female suicide ratio ranged from 1.14 (Sweden) to 2.73 (Italy). A similar pattern was found across most countries when examining 15- to 19-year-olds only; the male:female ratio ranged from 1.21 (South Korea) to 3.13 (Italy), with the exception of Uzbekistan (ratio: 0.87). We do not report ratios for 10- to 14-year-olds because data were only available for both sexes for 7 out of 35 countries. However, for almost all countries where data were available for both sexes of 10- to 14-year-olds, rates were higher among males than females (with the exception of Canada).
+Suicide methods. Figure 1 presents stacked bar charts showing the distribution of suicide methods across countries, for both males and females combined and separately. Several findings were consistent across countries. Hanging/suffocation was the
+most common method of suicide across all countries and for both sexes. With few exceptions (e.g., Estonia, New Zealand, Uzbekistan, Kyrgyzstan, Mexico, and Ireland), jumping/lying was the second most common method of suicide across both sexes. When examining sex differences in suicide methods across all countries with >10 cases, a significant difference was found only for self-poisoning (Q = 21.25, p < .001) such that males were more likely to selfpoison than females. The average self-poisoning mortality rate per 100,000 people was 0.42/ 100,000 for males and 0.21/100,000 for females. (Of note, this study was underpowered to examine self-poisoning by specific substance, such as drugs vs. other substances.) For all other methods, sex differences were nonsignificant (Q = <.0001-2.15 p = .142-.995).
+Moderators of suicide mortality rates
+Economic quality. Results of the moderated metaanalysis showed that pooled estimates did not differ by World Bank Income Group across any of the demographic groups (i.e., overall and by age or sex; Q = 0.02-3.25, p = .071-.895). There was no difference in the male:female ratio for the 10- to 19-year-olds (t = 0.29, p = .786) or the 15- to 19-year-olds (t = 0.75, p = .495).
+Economic inequality. Results of the moderated meta-analysis showed that pooled estimates did not differ by Gini coefficient across any of the demographic groups (i.e., overall and by age or sex; Q = 0.14-1.03, p = .310-.710). There was also no correlation between the male:female suicide ratio for 10- to 19-year-olds and Gini coefficient (r = .29, p = .226). However, there was a significant correlation between the male:female suicide ratio for 15- to 19-year-olds and Gini coefficient such that in countries with more inequality (higher Gini), there was a larger ratio of male:female suicides (r = .55, p = .023).
+Associations between access to lethal means and suicide mortality rates
+Access to firearms. When examining number of firearms per 100 people as a moderator of suicide mortality rates, we found that greater firearm access moderated the rate of suicide due to firearms (Q = 32.40, p < .001) but was unrelated to all other methods of suicide (p = .272-.979).
+Access to railways. When examining access to railways as a predictor of suicide mortality rates, we found significant omnibus tests for hanging/ suffocation (Q = 16.88, p = .009) and jumping/lying (Q = 16.51, p = .011). Omnibus tests for other suicide methods were nonsignificant (Q = 1.98-6.95, p = .326-.921). When examining pairwise post hoc
+comparisons for hanging/suffocation, there were, in general, differences between the lowest and highest densities, with higher rates of suicide due to hang-ing/suffocation in the areas with lower railway density. Specifically, rail density of 0-5 km per 1,000 km2 of lines differed significantly from all higher densities (all p < .001), > 100 km per 1,000 km2 of lines differed significantly from all lower densities (p = <.001-.002), and 5-10 km per 1,000 km2 and 50-75 km per 1,000 km2 differed significantly (p = .037). When examining pairwise post hoc comparisons for jumping/lying, there were significant differences among all pairs (Q = 6.91122.67, all p < .001), where areas of lower rail density had lower rates of suicide due to jumping/ lying than areas of higher rail density.
+Urban population. When examining whether percent of the population in an urban area moderated the suicide mortality rate, we found no moderation for any of the suicide methods (Q = 0.003-2.72, p = .099-.951).
+Discussion
+Our findings replicate and extend prior research in six important ways. First, our review provides an estimated suicide rate for the 10- to 19-year-old period (using WHO Mortality Data from 2010 to 2016) of 3.77/100,000 people. There are two important considerations when interpreting this rate: (a) Considerable heterogeneity was found in suicide mortality rates cross-nationally (which we discuss in the next section), and (b) our analyses included all available high-quality WHO mortality data, but only represent a subset of primarily Western countries worldwide (an issue we discuss in the Limitations section). However, this overall suicide rate among 10- to 19-year-olds is consistent with a prior study (Roh et al., 2018) that found a suicide rate of 3.94/ 100,000 people among 10- to 19-year-olds from 29 Organisation for Economic Co-operation and Development (OECD) countries during the period from 1995 to 2012. In addition, our 15- to 19-year-old rate of 6.04/100,000 people is similar, although lower, than the rate of 7.4/100,000 people found by Wasserman et al. (2005) among older (15-19 years old) adolescents from 90 WHO countries in 1995. Therefore, although these studies included data from different countries over different time periods, rates were relatively consistent for this population.
+Second, at the country level, this review replicates higher suicide rates for adolescents from New Zealand, as well as Estonia and Uzbekistan (both former Soviet Union States) (Bridge et al., 2006; Cha et al., 2018; Kõlves & De Leo, 2016; McLoughlin et al., 2015; Roh et al., 2018; Wasserman et al., 2005). Given that high rates in these regions have been documented for decades, a variety of explanations have been provided, although many have not
+been examined empirically or compared cross-na-tionally.
+High rates of suicide mortality among youth in New Zealand have been recognized as a major public health concern for decades (Associate Minister of Health, 2006). Disproportionately high suicide death rates have been found among youth from indigenous Maori populations, especially young Maori males. This disparity may be partially explained by the socially and economically disadvantaged status of Maori populations in New Zealand, evidenced by the disproportionate number of Maori youth receiving welfare services (Beautrais & Fergusson, 2006).
+Additionally, the elevated suicide rate among Maori youth may reflect the unique effects of colonization experienced by indigenous youth, including cultural alienation and identity confusion (Beautrais & Fer-gusson, 2006). Moreover, New Zealand consistently reports high rates of child abuse and neglect and bullying in school. A longitudinal study of 55,000 New Zealand children (under the age of 18) found that 23.5% had a report about their welfare made to Child Protective Services (CPS) by the age of 17 (Rouland & Vaithianathan, 2018). Adolescents involved with CPS and other social welfare systems were found to be at elevated risk of suicide death
+(Beautrais, 2001). In a 2015 cross-national study of OECD countries, New Zealand reported the second highest adolescent bullying rate of the 51 countries examined, with over 25% of adolescents experiencing some form of bullying multiple times a month, and 18.1% of adolescents met the criteria for ‘frequent bullying’ - more than twice the rate of the 50 other countries surveyed (OECD, 2017a). Government initiatives in New Zealand have aimed to address this elevated suicide risk and improve mental health care (Associate Minister of Health, 2006). However, prevention efforts, including means restrictions, are challenging, as most suicide deaths occur by hanging in private dwellings (Taylor, 2010). Of note, hanging is used more commonly among Maori than nonindigenous populations (Taylor, 2010). A recent review indicated that, although research is growing, few intervention and prevention programs in New Zealand have been evaluated (Coppersmith, Nada-Raja, & Beautrais, 2018).
+Elevated suicide rates were also reported among youth in Estonia and Uzbekistan, consistent with prior studies finding higher rates among youth living in former Soviet Union states (Bridge et al., 2006; Cha et al., 2018; Kõlves & De Leo, 2014, 2016; McLoughlin et al., 2015; Roh et al., 2018; Wasserman et al., 2005). However, unlike New Zealand, little research has explored specific risk factors that contribute to high rates in these regions. Under the Soviet Union, suicide was a classified subject and suicide statistics were kept secret or discarded, delaying and discouraging the emergence of suicide as an acknowledged public health concern in this region (Wasserman & Varnik, 1998). The collapse of the Soviet Union in 1991 led to numerous social, political, and economic difficulties as former Soviet states rebuilt as independent nations. Estonia, in particular, experienced a sharp increase in suicide deaths among its Russian immigrant minority following independence from the Soviet Union as Russian immigrants lost their previously privileged status (Varnik, Kõlves, & Wasserman, 2005). Suicide rates in this region may also be related to the transition from a strict Soviet campaign against alcohol to more lax policies regulating alcohol (Koõlves & De Leo, 2016). Increased accessibility of alcohol led to a higher incidence of alcohol-related suicide deaths among adults (Koõlves, Milner, & Varnik, 2013; Varnik, Kõlves, Vali, Tooding, & Wasserman, 2007), and studies conducted in some former Soviet countries found that changes in national alcohol consumption were linked with fluctuations in mortality rates (Koõlves & De Leo, 2016; Koõlves et al., 2013). More research is needed, however, to clarify how alcohol consumption among adolescents in these countries may contribute to higher rates of suicide among youth.
+A third finding from this review replicates a well-established trend that suicide rates are higher among older adolescents (Koõlves & De Leo, 2017;
+Roh et al., 2018) and young adults (Bridge et al., 2006; Cha et al., 2018) compared with younger adolescents. The age finding is also consistent with research describing the trajectories of suicidal thoughts and behaviors during adolescence, specifically that the onset of suicide ideation typically occurs during early adolescence (around ages 1113) and, for a subgroup of youth, transitions to suicidal behavior during later adolescence (around age 15 or 16; Glenn et al., 2017; Nock et al., 2013). This previous research may suggest the existence of a developmental process (or set of processes) by which adolescents become more capable of engaging in suicidal behavior as they transition from early to later adolescence. Although the nature of these developmental processes remains unclear, there are several key differences between younger and older adolescents that may be relevant to changes in suicide risk (Dervic et al., 2008). First, the rise in suicidal thoughts and behaviors across adolescence coincides with increases in rates of other forms of psychopathology that confer risk for suicide, such as nonsuicidal self-injury, depression, substance use disorders, and certain anxiety disorders (Costello, Copeland, & Angold, 2011; Glenn et al., 2017; Nock et al., 2009). Second, compared to younger adolescents, older adolescents engage in more risk-taking behaviors (Braams, van Duijvenvoorde, Peper, & Crone, 2015), another known risk factor for suicidal thoughts and behaviors (Ammerman, Steinberg, & McCloskey, 2018). Third, older adolescents have more fully developed cognitive facilities than younger adolescents, which may intensify the complexity and severity of maladaptive thinking. For instance, increases in metacognition and abstract reasoning may enhance the ability to ruminate (Papageorgiou & Wells, 2003), and increases in future thinking abilities may facilitate hopelessness (Kosnes, Whelan, O’Donavan, & McHugh, 2013). Thus, types of negative cognition commonly associated with suicidal thoughts and behaviors (Cha, Wilson, Tezanos, DiVasto, & Tolchin, 2019) may become more advanced during older adolescence. Moreover, normative developmental changes in adolescent social networks that make peers more influential may also contribute to increased potential for imitation of risky behaviors, including modeling of suicidal behavior (Pickering et al., 2018). Finally, individuals with histories of multiple suicide attempts are at especially high risk of later suicide death (Kochanski et al., 2018). Beginning at age 17, most suicide attempts are repeat attempts (Goldston et al., 2015); thus, older adolescents may be at increased risk of suicide death due to increased experience engaging in suicidal behavior. Taken together, these observations suggest that higher rates of suicide during older adolescence may be due, at least in part, to other developmental changes during this period.
+Fourth, higher suicide rates were reported among males (4.83/100,000 individuals) compared with females (1.95/100,000 individuals) in this age range. This sex effect is consistent with many prior studies in youth (Bridge et al., 2006; Cha et al., 2018; Kõlves & De Leo, 2014, 2016; McLoughlin et al., 2015; Miranda-Mendizabal et al., 2019; Roh et al., 2018; Varnik et al., 2009; Wasserman et al., 2005) and is also found among adults (Bachmann, 2018; Canetto & Sakinofksy, 1998; Chang et al., 2019; Nock, Borges, Bromet, Cha, et al., 2008; Schrijvers et al., 2012). Higher suicide death rates among males have been attributed to a range of factors, including greater use of lethal means (e.g., hanging and use of firearms; Callanan & Davis, 2012) and higher incidence of risk factors related to suicide death, such as substance use and aggressive and risk-taking behaviors (Bozzay, Liu, & Kleiman, 2014). Although the sex effect (with higher rates among males) was observed in most countries, the main exception to this trend was Uzbekistan (ratio male:female 0.95), where suicide rates between sexes were relatively comparable. In addition, 10 other countries (across North America, Europe, and Asia) had a male:female ratio of <2, which is surprising given prior findings that the suicide rate among males is at least 2-4 times higher than among females (Bridge et al., 2006; Cha et al., 2018; Kõlves & De Leo, 2016; McLoughlin et al., 2015; Roh et al., 2018; Varnik et al., 2009; Wasserman et al., 2005). Taken together with prior studies, these findings suggest that although overall suicide rates are higher among males than females, this sex difference is not uniform cross-nationally nor stable over time. Variation in suicide deaths by sex underscores an important role for cultural factors in suicidal behavior among youth.
+Fifth, this review extends prior work by examining cross-national differences in suicide methods among adolescents. Hanging/suffocation was the most common method of suicide death worldwide among 10- to 19-year-olds, followed by jumping/ lying in front of a moving object or jumping from a height, consistent with prior studies in youth (Hepp et al., 2012; Kõlves & De Leo, 2017; Varnik et al., 2008, 2009). Although hanging/suffocation is also a common method of suicide death among adults, previous research has found that jumping from a height and railway suicide deaths are much more common in youth than adults, and intoxication is less common among youth than adults (Hepp et al., 2012).
+Although certain methods were more common overall, there were also differences between sexes. Males were more likely to die by self-poisoning than females, contrary to prior findings of higher rates of self-poisoning in females compared with males (Hepp et al., 2012; Koõlves & De Leo, 2017). However, a number of studies have found that selfpoisoning by drugs is higher for females, while
+self-poisoning by other substances is higher for males (Rajapakse, Griffiths, Christensen, & Cotton, 2014; Varnik et al., 2008, 2009). Although the current study was unable to examine differences in self-poisoning by specific source, this remains an important direction for future research. Interestingly, there were no sex differences in suicide by firearm, as has been found in previous research (Hepp et al., 2012; Kõlves & De Leo, 2017; Varnik et al., 2009). However, given that suicide death counts were low, the nonsignificant findings should be interpreted with caution.
+Notably, we found that use of particular suicide methods varied based on cross-national differences in access to these methods. Specifically, increased access to firearms within a country was strongly related to suicide death by firearm in that country, but not to suicide death by other methods. However, as discussed in the Methods section, firearm access was not stable over the period of data collection for this study - a notable limitation and opportunity for future research. Nevertheless, these findings highlight the importance of means restriction of firearms in countries with greater firearm access. Improved firearm legislation in New Zealand (Beautrais, Fergusson, & Horwood, 2006) and firearm storage in the United States (Brent et al., 1991; Grossman et al., 2005) have been related to reduced suicide rates. In addition, greater access to railways was associated with jumping/lying in front of a moving object (e.g., train) or jumping from a height. Some successful prevention strategies for railway deaths include use of sliding doors to limit access to rail track and creating deep channels between rails (Krysinska & De Leo, 2008; Pirkis et al., 2015). Moreover, a recent meta-analysis of prevention strategies for ‘suicide hotspots’ (where most deaths were due to jumping from a height) found evidence for reduced suicide rates with means restriction by building fences or rails to limit access (Pirkis et al., 2015).
+Although we were able to measure access to certain suicide methods (i.e., firearms and railways), we were unable to estimate access to methods used for hanging/suffocation - the most commonly reported method among youth. Just as it is difficult to measure access, it is also difficult to restrict access to means used for hanging/suffocation - an issue that has made suicide prevention by hanging/suffo-cation in youth extremely challenging (Sarchiapone, Mandelli, Iosue, Andrisano, & Roy, 2011). This is particularly alarming in light of recent findings from the United States, indicating that suicide deaths by hanging/suffocation are on the rise among youth (Bridge et al., 2015; Sullivan, Annest, Simon, Feijun, & Dahlberg, 2015).
+Sixth, the current review found that suicide rates overall were not significantly associated with the included indices of economic quality and inequality. Although ours is not the only study to find that
+economic indices did not significantly relate to suicide rates (Bremberg, 2017; Vijayakumar, Nagaraj, Pirkis, & Whiteford, 2005), this finding is somewhat surprising in light of converging evidence, suggesting that economic factors have a significant impact on suicidal behavior (Bachmann, 2018). For instance, economic crises, and high unemployment rates in particular (Nordt, Warnke, Seifritz, & Kawohl, 2015), have been broadly linked to suicide deaths (WHO, 2014). In addition, lower socioeconomic indices are associated with suicide attempts across the world (Andres, Collings, & Qin, 2009; Burrows & Laflamme, 2010; Fang, 2018; Ki, Sohn, An, & Lim, 2017). However, less research has examined how economic indices relate to suicide deaths cross-nationally. It is important to note that in the current study, the majority of included countries (93%) were from high- or upper-middleincome groups. Therefore, the null findings may be due to restricted worldwide coverage - an issue we discuss in the Limitations section. Moreover, the included measure of national economic inequality may fail to capture important heterogeneity of inequality between cities within a country (Glaeser, Resseger, & Tobio, 2009).
+Although economic indices were not related to suicide rates in the overall sample, there was a moderate to strong correlation between economic inequality (Gini coefficient) and the male:female suicide death rate ratio among 15- to 19-year-olds; that is, greater income inequality was associated with a higher suicide rate for males compared with females within a country. This finding is consistent with some research, suggesting that economic hardships may be related to poorer mental health outcomes among males compared with females. For instance, in a large cross-national study, Gini coefficients were significantly related to depressive symptoms in males but not females (Yu, 2018). In addition, in the United States, lower-income school contexts have been related to increased suicide ideation and attempts among males but not females (Fang, 2018). Moreover, a study in Denmark found that lower income and unemployment were related to suicide deaths for all adults, but effects were greater among males (Andres et al., 2009). Greater risk among males in poorer economic circumstances may be related to their role as primary income earners for their families in many countries (Mann & Metts, 2017). However, further research is needed to understand the mechanism of risk, particularly among youth, and to suggest potential targets for prevention.
+Limitations and future directions
+Although this review significantly extends knowledge of cross-national suicide trends in youth, limitations of this research warrant discussion. First, this review was limited in its worldwide
+coverage of only 45 (mostly high- and middleincome) countries out of the 194 WHO member countries. There are surprisingly little cross-national data on suicide mortality rates that are publicly available beyond the WHO Mortality Database. Of the countries included in the WHO database, data that are determined to be of ‘high’ quality are predominantly from countries in Europe, North America, Asia, and two high-income countries in Oceania (New Zealand and Australia); coverage of South America and Africa is limited. As a result, the current findings may not accurately estimate suicide rates in youth worldwide, particularly among these underrepresented regions.
+The dearth of good-quality suicide mortality data worldwide may be due to significant underreporting (e.g., stigma) and misclassification of suicidal behaviors (e.g., lack of knowledgeable medical professionals), particularly in countries where suicidal behavior is illegal (Bachmann, 2018; De Leo, 2015; WHO, 2014). Additionally, many countries - including India, China, and the majority of nations in Africa - have not yet developed national death registration systems. As of 2010, <30% of the global population resided in countries with established death registration systems, resulting in lower quality mortality data for these regions (Bhalla, Harrison, Shahraz, & Fingerhut, 2010). The WHO estimates that suicides in countries without goodquality data account for approximately 71% of global suicide deaths annually. Good vitality registration data are disproportionately available for wealthier countries, with high-quality coverage for 95% of suicides in high-income countries but only 8% of all estimated suicide deaths in low- or middle-income countries (WHO, 2014). Greater cross-national coverage is greatly needed to more accurately estimate worldwide suicide mortality rates. Notably, in 2014, the World Bank and the WHO published a global investment plan to increase the number and quality of national civil registration and vital statistics systems (CRVS) in low- and middle-income countries (World Bank & World Health Organization (WHO), 2014). The initiative aims to both strengthen existing national CRVS systems and to catalyze the implementation of new systems by developing model CRVS legislation and expanding training for physicians and other medical staff responsible for registering vital statistics (World Bank & World Health Organization (WHO), 2014)).
+Second, due to low counts, we were limited in our ability to examine all suicide methods. For instance, we had to combine jumping from a height and jumping/lying in front of a moving object. Although both are violent methods, prevention efforts, such as restricting access to means, may be distinct. In addition, we were unable to examine differences in the specific substances used for self
+poisoning. As already noted, there may be important sex differences in overdosing via drugs (higher in females) versus other substances (higher in males) (Rajapakse et al., 2014; Varnik et al., 2008, 2009).
+Third, we examined cross-national trends based on binary female and male sex rather than gender or gender identity. At present, there is a substantial dearth of information regarding suicide death rates among nonbinary, transgender, and gender-expansive youth. Only recently have large-scale, nationally representative studies, such as the USA’s Youth Risk Behavior Surveillance System (CDC, 2019) and the National Violent Death Reporting System (NVDRS; CDC, 2016), begun to consider including gender-expansive demographic characteristics in their measures (CDC, 2017a). Recent reporting data suggest important sex-based differences in suicide rates and forms of suicidal behavior. For example, a report of data from the NVDRS between 2013 and 2015 examining suicide decedents, 12- to 29-year-olds, in 18 states within the United States (Ream, 2019), found that 13% of transgender males and 8% of transgender females' suicide deaths were due to firearms as compared to 55% and 34% of cisgender, heterosexual males and females, respectively (Ream, 2019). However, these data were limited in two key ways: They were not nationally representative, and transgender identity was coded based on information included in reports from law enforcement and medical examiners (Ream, 2019). The latter is a significant limitation given that gender identity is often not listed on death certificates or coroners’ reports (Haas & Lane, 2015; Haas et al., 2010; Ream, 2019). However, data from individual studies conducted in the United States suggest rates of suicide ideation (i.e., thoughts of killing oneself) and suicide attempts (i.e., selfdirected injury with at least some intent to die) may be as high as 31% and 17%, respectively, among transgender youth, compared with 11% and 6% in matched cisgender peers (Reisner et al., 2015). There is little information available for transgender youth in Asian, African, and South American countries (Adams, Hitomi, & Moody, 2017; McNeil, Ellis, & Eccles, 2017). Future crossnational research would benefit from incorporating comprehensive demographic questions surrounding gender and gender identity to estimate the suicide death rate among nonbinary, transgender, and gender-expansive youth.
+Fourth, although this review examined several important cross-national trends, many potential moderators of interest could not be examined (see Cha et al., 2018; for a review of other important sociodemographic factors). For instance, we were not
+able to examine racial and ethnic differences within countries given the limited demographic information provided in the WHO Mortality Database. Examining suicide rates only at the aggregated country level may mask important differences based on race and ethnicity within countries. Illustrative of major differences are high suicide rates among indigenous, or native, youth in countries such as New Zealand (Beautrais & Fergusson, 2006), Australia (Cantor & Neulinger, 2000), Brazil (Coloma, Hoffman, & Crosby, 2006), and the United States (Leavitt et al., 2018). Moreover, high suicide rates have been reported among racial minority youth (e.g., Black children, 5-11 years old, in the United States, Bridge et al., 2015). Differences in youth suicide risk also have been reported as a function of immigrant generation status (Pena et al., 2008). These examples highlight the importance of examining crossnational and intranational trends based on racial, ethnic, and generational status factors in future research.
+Finally, this review provides primarily descriptive and correlational information about worldwide suicide rates in adolescents. Although useful for understanding current cross-national trends, inferences should not be made about causation. Future research is needed to understand how factors such as access to lethal means and economic inequality may directly influence suicide rates.
+In summary, the current review provides an updated estimate of worldwide suicide rates in adolescents, 10-19 years old, using the WHO Mortality Database from 2010 to 2016. Replicating prior research, suicide deaths were overall more common among male and older (15-19 years old) adolescents, hanging/suffocation was the most common method, and highest rates were found in Estonia, New Zealand, and Uzbekistan. This review contributes new information through findings that access to firearms and railways were related to suicide deaths by firearms and jumping/lying in front of a moving object or jumping from a height, respectively. Similar to prior reviews, this study was limited in its worldwide coverage of suicide rates and trends. Important future research directions include expanding the worldwide coverage to more low- and middle-income countries, examining suicide trends among nonbinary gender groups and by race/ethnicity within countries, and clarifying factors that account for cross-national differences in suicide rates.
+Acknowledgements
+The research was partially supported by a grant from the National Institute of Mental Health (L30 MH101616;

Annual Review of Clinical Psychology Transforming the Treatment of Schizophrenia in the United.txt

@@ -0,0 +1,72 @@
+1.	INTRODUCTION
+The schizophrenia spectrum disorders (henceforth referred to as schizophrenia) are neurode-velopmental illnesses with a lifetime prevalence near 1%; they can cause extensive functional impairment and have for too long carried low expectations for recovery (Lieberman et al. 2013). Only 10-15% of people with schizophrenia are employed, and many remain on disability (Harvey et al. 2012). In 2013, excess total costs of schizophrenia in the United States were estimated at $155.7 billion, including significant direct health care costs but mostly indirect costs related to losses to the labor market (Cloutier et al. 2013). In 2009 the National Institute of Mental Health (NIMH) funded a set of research studies called Recovery After an Initial Schizophrenia Episode (RAISE) in order to build on national and international studies to change this gloomy state of affairs (Heinssen et al. 2017). The RAISE studies contributed to the creation of a new way to organize treatment, called coordinated specialty care (CSC), which has the promise of improving the course of schizophrenia (Dixon et al. 2015, Kane et al. 2016). The creation and dissemination of CSC programs across the United States and the contribution of the RAISE projects can be understood as the intersection of trends in both science and policy that converged to create the foundation for changes in care and care delivery (Dixon 2017a,b). This article discusses the key dimensions of these dramatic changes anchored in the RAISE projects.
+This review is divided into four sections. Section 2 considers the pre-RAISE era, with a focus on the scientific and policy context of the project in the United States: What led to RAISE? Section 3 focuses on the findings of the RAISE studies, including both scientific and policy/service delivery dimensions. We emphasize the RAISE early treatment program (RAISE-ETP) project, which is the large randomized trial of a CSC model (Kane et al. 2016). Section 4 discusses key unanswered questions and challenges in the aftermath of the RAISE studies. Section 5 concludes.
+2.	UNDERSTANDING THE SCIENTIFIC AND POLICY CONTEXT FOR RAISE AND OTHER FIRST-EPISODE PSYCHOSIS SERVICES IN THE UNITED STATES
+2.1.	Policy and Service System Issues
+Early intervention services for psychotic disorders have been implemented in Australia and Northern Europe for over two decades, survived experimental tests for efficacy in Denmark and the United Kingdom, and have since 2000 been part of a national implementation plan in England (Dep. Health 2000, Srihari et al. 2012). The prospect of rapidly providing care after the onset of psychosis is consistent with approaches to other medical disorders, and it presented itself as a “best bet” for many national health care systems that invested in this opportunity even as research and testing of specific models were still underway (McGorry 2012). Why was the United States so late in developing a national strategy for early psychosis?
+The rise of the community mental health movement in the middle of the twentieth century reflected the belief that early intervention would reduce chronic disability for many mental illnesses (Grob & Goldman 2006). This had been a central promise of the moral treatment era in the midnineteenth century, embodied in the rise of asylums, and also a central promise of the mental hygiene movement of the early-twentieth-century progressive era, embodied in the development of psychopathic hospitals and youth guidance centers. Unfortunately, the interventions at each turn of these reform cycles failed to deliver on their promises. By the mid-1970s the community mental health centers were criticized for their failure to sufficiently prevent the severe disability associated with chronic mental illnesses (Gen. Account. Off. 1977, Tessler & Goldman 1982).
+Thus, the disappointment following the initial optimism led to a series of policies that turned away from early intervention and neglected the potential for such treatments and services. It did not help that the clinical and neuroscience evidence at the time did not support moving further. A weak technology does not advance service delivery.
+The mid-1970s critique of the NIMH community mental health center program for its failure to focus on chronic mental illness led the NIMH to support the development of community support programs and system reforms, redirecting public sector attention to improving services for individuals already disabled by mental illness. These individuals, such as people with mostly chronic schizophrenia, became the target population for public mental health systems now in the throes of a community support reform cycle. Individuals with less disabling or early-stage mental illnesses were not targeted, or even eligible, for services (Grob 1994, Grob & Goldman 2006, Tessler & Goldman 1982).
+The lack of public sector priority for individuals in the early stages of psychosis was exacerbated by the fact that the system increasingly relied on Medicaid for funding (Frank et al. 2003). Access to Medicaid for young adults at the peak age of onset for psychosis was dependent on eligibility for the Supplemental Security Income (SSI) disability program. To be eligible one had to be disabled already. Single individuals in the early stages of psychosis typically did not qualify for SSI, and thus they were also ineligible for Medicaid unless they had dependent children and were impoverished. In some states, access to public sector services was difficult for individuals who were not on Medicaid (Goldman et al. 2013).
+Individuals in the private sector were also disadvantaged by insurance rules. They were removed from parental health insurance unless they paid very high COBRA premiums or were full-time students. They lost insurance if they left the workplace due to their illness. The classification of their mental illness as a preexisting condition allowed commercial payers to place them into a new high-risk insurance pool and thereby inflate their premiums or exclude them from coverage
+altogether. In any case, they fell out of the private sector and, as we learned above, also had trouble qualifying for public sector services (Goldman et al. 2013).
+It was not until the Affordable Care Act (ACA) was passed in 2009 that some of those exclusionary rules were weakened, allowing more individuals with a first episode of psychosis to retain insurance coverage (Goldman 2010). Patients gained access to parental and other private sector insurance through new underwriting rules for health insurance exchanges, and they could qualify for Medicaid in states that accepted federal support to extend this entitlement to low-income individuals. There were other reforms in the ACA that provided better services for individuals experiencing the early stages of a psychotic illness. Furthermore, supplements to the federal block grant specifically earmarked for early intervention made more services available (Goldman & Karakus 2014).
+2.2.	The Development and Identification of Evidence-Based Practices for Schizophrenia in the United States: The PORT Initiative
+An important antecedent to the RAISE studies and the dissemination of CSC programs was the identification of evidence-based practices in general. The components of CSC tested in RAISE were based almost entirely on evidence-based interventions for established schizophrenia, applied to the early stages of psychosis. The Agency for Health Care Policy and Research (now called the Agency for Healthcare Research and Quality) began to fund so-called Patient Outcomes Research Teams (PORTs) in the late 1980s and early 1990s in recognition of the fact that many (if not most) treatment decisions in medicine were made without any systematic input from scientific data about efficacy, effectiveness, and cost. The earliest PORTs focused on management of back pain, acute myocardial infarction, and cataracts (Goldberg & Cummings 1994). The first PORT that addressed a mental illness was awarded to investigators at the University of Maryland and Johns Hopkins, and it focused on schizophrenia. The PORT studies attempted to systematically review evidence from relevant clinical studies to make treatment recommendations to clinicians for specific patient populations.
+Subsequently, three sets of PORT recommendations for schizophrenia were published, all of which largely identified recommended treatments based on empirical support rather than expert opinion (Buchanan et al. 2010, Dixon et al. 2010, Kreyenbuhl et al. 2010, Lehman & Steinwachs 1998a, Lehman et al. 2003). In the first set of PORT recommendations (Lehman & Steinwachs 1998a), 18 of the 30 recommendations focused on the use of antipsychotic medications for acute and maintenance treatment. The recommendations identified appropriate dosage ranges and also highlighted the utility of clozapine. Relevant to early psychosis treatment, one of the original recommendations specified that patients experiencing a first acute episode should be treated with dosages in the lower end of the overall recommended range for people with more long-standing conditions.
+Regarding psychologic and psychosocial treatments, the team recommended vocational rehabilitation (for individuals having characteristics associated with good employment outcomes), family support, individual and group therapies consisting of education and cognitive and behavioral skills training, and assertive community treatment (ACT). The first Schizophrenia PORT was also funded to assess the extent to which routine care conformed to evidence-based treatment recommendations; Lehman & Steinwachs (1998b) found that overall conformance to the recommendations was modest (generally below 50%) and higher for pharmacological than for psychosocial treatments. The findings of the initial Schizophrenia PORT underscored the gap between science and practice.
+By the time the second PORT recommendations for schizophrenia (Lehman et al. 2003) were issued, the scientific literature had been able to address the impact of second-generation
+antipsychotic agents, and the specificity and availability of psychosocial interventions had improved. This updated report continued to recommend implementation of ACT and family interventions lasting at least nine months and including illness education, crisis intervention, emotional support, and training in how to cope with illness symptoms and related problems. The update also elaborated on group and individual therapy to include cognitive behavioral therapy (CBT) as the therapy of choice for residual psychotic symptoms. Social skills training was newly recommended. The team also identified supported employment as the service of choice in place of the broader concept of vocational rehabilitation for anyone interested in obtaining employment. It is notable that the second PORT continued to clarify the empirical foundation for future CSC programs.
+The last set of PORT recommendations, published in 2010, continued to highlight the need to use lower doses of antipsychotic medications and to avoid the use of clozapine and olanzapine as a first-line treatment in early psychosis (Buchanan et al. 2010, Kreyenbuhl et al. 2010). The emerging evidence for psychosocial treatments provided more precise information on relevant populations and expected outcomes (Dixon et al. 2010). In addition, these recommendations supported alcohol and substance use services and weight management, given the high co-occurrence of these comorbidities with schizophrenia and the availability of effective treatments. The key elements of treatment for alcohol or drug use disorders for persons with schizophrenia include motivational enhancement and behavioral strategies that focus on engagement in treatment, coping skills training, relapse prevention training, and its delivery in a service model that is integrated with mental health care. Regarding weight loss, a psychosocial intervention that is at least 3 months long that includes psychoeducation focused on nutritional counseling, caloric expenditure, and portion control; behavioral self-management including motivational enhancement; goal setting; regular weigh-ins; self-monitoring of daily food and activity levels; and dietary and physical activity modifications was recommended. This 2010 review did not find sufficient evidence for an overall recommendation of a specific single or multicomponent treatment for early psychosis. However, as discussed below, several preliminary studies found results favoring family interventions, CBT, and supported employment, all in antipsychotic-treated populations. Furthermore, the review reported on evidence from international randomized controlled trials (RCTs) supporting multielement interventions for early psychosis that provide comprehensive packages of psychosocial and medication supports.
+The most recent Schizophrenia PORT identified five published papers on CBT for early psychosis, of which three included actual CBT trials. Three of the five papers came from a UK longitudinal study named the Study of Cognitive Reality Alignment Therapy in Early Schizophrenia (SoCRATES) (Lewis et al. 2002, Tarrier & Wykes 2004, Tarrier et al. 2006). In the study, the SoCRATES intervention group received a stage-based, manualized CBT intervention, whereas the control groups received supportive counseling or treatment as usual. The SoCRATES intervention group showed greater improvements on delusions and auditory hallucinations indexes compared to treatment as usual (TAU) and supportive counseling groups, but SoCRATES was only better than TAU on the Positive and Negative Syndrome Scale (PANSS) positive symptom subscale. Moreover, whereas individuals in the CBT and supportive counseling groups appeared to get better faster than those in TAU, medical records indicated that the three groups did not differ significantly on rehospitalization or relapse rates. The fourth CBTpaper reported on a quasi-experimental study conducted in Australia and found no differences between the CBT intervention group and standard care (Jackson et al. 2005). However, the lack of an adequate control sample, a weak CBT intervention, and enriched standard care (offered in the pioneer EPPIC program) limited inferences. The final CBT study compared active cognitive therapy for early psychosis to befriending as part of an early intervention program in the United States. Again, this study found no significant differences between the CBT and comparison groups (Jackson et al. 2008).
+The Schizophrenia PORT identified four published papers on family interventions, of which two were controlled studies. Of the controlled studies, one study that took place in China found greater symptom improvement and lower rates of hospitalization among those who received the family intervention (Zhang et al. 1994). The second controlled study, which took place in the Netherlands, found no differences between individuals receiving the family intervention and individuals not receiving it (Linszen et al. 1996). In the follow-up study, five years later, individuals who had received the family intervention had spent less time living in institutional settings (Lenior et al. 2001, 2002).
+With regard to supported employment, only one study had been identified at the time of the 2010 Schizophrenia PORT. This study evaluated an occupational intervention for early psychosis (Killackey et al. 2008). Specifically, the EPPIC program in Australia randomly assigned patients either to receive individual placement of support (IPS) along with EPPIC’s services for early psychosis or to receive EPPIC services alone. Both groups were followed for six months, and the results indicate that those individuals who received IPS had better employment outcomes.
+In addition to the monotherapies mentioned above, PORT took into account three RCTs across eight publications of comprehensive, multielement, psychosocial treatment programs in Europe. Each of these programs used ACT or an equivalent approach as the treatment structure and enhanced it by including various evidence-based treatments, including CBT, skills training, and psychoeducation. Five of the studies came from the OPUS project in Denmark (Bertelsen et al. 2008; Jeppesen et al. 2005; Kassow et al. 2002; Petersen et al. 2005a,b), two came from the Lambeth Early Onset (LEO) project in the United Kingdom (Craig et al. 2004, Garety et al. 2006), and one came from a small RCT in Norway (Grawe et al. 2006). Notably, these studies were pragmatic in design: They tested ecologically relevant interventions in real-world samples and measured a range of salient outcomes across clinical (relapse, remission, rehospitalization), functional (social, education, and employment) and economic (cost) domains. Notably, these second-generation studies (Srihari et al. 2012) went beyond establishing efficacy for single-component interventions (e.g., CBT) to test comprehensive models of care that responded to the diverse needs of patients and families presenting for care. Positive outcomes compared to usual care were reported over follow-up periods of up to two years across these domains. Notably, as discussed below, these improvements were not sustained when individuals were assessed three years after being discharged from these specialized services to usual care (Bertelsen et al. 2008, Gafoor et al. 2010).
+In summary, research conducted in other countries had made great strides in demonstrating the effectiveness of comprehensive early intervention services (Srihari et al. 2012). These countries did not have in place policies, such as those in the United States, which impeded the development and widespread dissemination of these services. Notable public sector pioneers in the United States included Oregon’s EASA program (established in 2001; http://www.easacommunity.org/), Massachusetts’s PREP (2003; Caplan et al. 2013), North Carolina’s OASIS (2005; Uzenoff et al. 2012), San Francisco’s PREP (2006; http://felton.org/ social-services/early-psychosisschizophrenia-prep/), and Connecticut’s Specialized Treatment Early in Psychosis (STEP) program, which launched the first US pragmatic RCT of teambased care for early psychosis (2006; Srihari et al. 2009). All these programs were serving early psychosis patients, beginning to move beyond delivering standard psychopharmacology trials and toward delivering different types of comprehensive treatment models. Thus, by the first decade of the millennium, there was an emerging foundation for an evidence-based approach for first-episode psychosis. The United States had clearly lagged behind many other countries in developing the clinical and policy context necessary to launch such programs. The time was ripe for the RAISE studies to tackle this situation and build on the evidence base that international colleagues had developed.
+3.	THE CREATION AND RESULTS OF RAISE PROJECTS
+The NIMH RAISE initiative aimed to develop and test an intervention that would engage individuals with early psychosis, improve recovery trajectories, and prevent or limit long-term disability, while reducing the costs associated with psychotic disorders. RAISE supported the development, testing, refinement, and implementation of CSC in real-world, community-based behavioral health centers in the United States. The focus was on the feasibility, effectiveness, and acceptability of the program in real-world settings. Clinicians who were already members of the behavioral health workforce, rather than specially trained research staff, delivered the program after modest amounts of training, within already existing treatment centers, and using preexisting billing/reimbursement structures to pay for the services whenever possible. The RAISE initiative also intended to foster the rapid expansion of CSC services in the community once the studies ended (Azrin et al. 2015). The initiative funded two studies: the RAISE Early Treatment Program (RAISE-ETP) and the RAISE Implementation and Evaluation Study (RAISE-IES). The initial contracts were funded with economic stimulus dollars made available as a response to the Great Recession of 2008.
+3.1.	Key Findings from RAISE-ETP
+The RAISE-ETP was built around the CSC intervention labeled NAVIGATE (Mueser et al. 2015). The four manual-based key interventions were psychopharmacology, for which a computerized prescriber decision support system called COMPASS was developed (Robinson et al. 2018); individual resilience therapy; family therapy/psychoeducation; and supportive employment/education (see http://raiseetp.org for manuals). A cluster-randomized design was employed and involved 34 nonacademic, community mental health centers in 21 states across the United States. Seventeen clinics were randomized to deliver NAVIGATE and 17 clinics were randomized to provide usual care. The staff at the NAVIGATE-assigned clinics were then trained in all four modalities utilizing a variety of tools and techniques. Research diagnostic interviews and major outcome assessments were conducted by blinded, remote, centralized raters using live twoway video. The primary outcome measure was the Heinrichs-Carpenter Quality of Life Scale. A total of 404 first-episode psychosis patients with a mean age of 23 were enrolled (Kane et al. 2015).
+At the two-year follow-up, the NAVIGATE-treated patients did significantly better on the Quality of Life Scale, the Positive and Negative Syndrome Scale, the Calgary Depression Scale for Schizophrenia, the length of time staying in treatment, and the degree of improvement in work/school engagement. There was no significant difference in the rate of hospitalization between the two groups, though rates overall were relatively low (Kane et al. 2016). The median duration of untreated psychosis (DUP) in this sample was 74 weeks (Addington et al. 2015). When the influence of DUP on quality of life outcomes was examined, it proved to have a highly significant moderating effect, with individuals having a DUP shorter than 74 weeks deriving significantly more benefit from the CSC than those with longer DUP (Addington et al. 2015, Kane et al. 2016). These findings further underscore the potential value of reducing DUP.
+RAISE-ETP utilized a computerized prescriber decision support system, which also helped to facilitate evidence-based care (Robinson et al. 2018). Over the two years, the 223 NAVIGATE participants compared to the 181 clinician-choice participants had more medication visits, were more likely to be prescribed an antipsychotic (and also an antipsychotic conforming to NAVIGATE prescribing principles), and were less likely to be prescribed an antidepressant. (As noted previously, at the same time they also had significantly lower scores on the Calgary Depression Scale for Schizophrenia.) NAVIGATE participants experienced fewer side effects and also gained less weight; other vital signs and cardiometabolic laboratory findings did not differ between treatments.
+www.annualreviews.org • The RAISE Initiative 243
+Adherence estimator scores (McHorney 2009) decreased (fewer beliefs associated with nonadherence) with NAVIGATE but not clinician-choice care.
+The recruitment of 404 individuals receiving treatment at community mental health centers across the United States after the onset of a first episode of psychosis also provided a window into the medication histories and medical status of these individuals at the time of referral (Robinson et al. 2015b). A total of 159 patients (39.4% of the sample) were identified as potentially benefiting from changes in their psychotropic prescriptions. Of these, 8.8% received prescriptions for recommended antipsychotics at higher-than-recommended dosages; 32.1% for olanzapine (often at high dosages); 23.3% for more than one antipsychotic; 36.5% for an antipsychotic and also an antidepressant without a clear indication; 10.1% for psychotropic medications without an antipsychotic; and 1.2% for stimulants.
+With regard to medical status (Correll et al. 2014), in 394 of404 patients with cardiometabolic data [mean (SD) age = 23.6 (5.0) years; mean (SD) lifetime antipsychotic treatment = 47.3 (46.1) days], 48.3% were obese or overweight, 50.8% smoked, 56.5% had dyslipidemia, 39.9% had prehypertension, 10.0% had hypertension, and 13.2% had metabolic syndrome. Prediabetes (glucose based = 4.0%; hemoglobin Ajc based = 15.4%) and diabetes (glucose based = 3.0%; hemoglobin A1c based = 2.9%) were less frequent. Total psychiatric illness duration correlated significantly with higher body mass index, fat mass, fat percentage, and waist circumference (all P < 0.01) but not elevated metabolic parameters [except triglycerides to HDL-C ratio (P = 0.04)]. Conversely, antipsychotic treatment duration correlated significantly with higher non-HDL-C, triglycerides, and triglycerides to HDL-C ratio and with lower HDL-C and systolic blood pressure (all P < 0.01). Olanzapine was significantly associated with higher triglycerides, insulin, and insulin resistance, whereas quetiapine fumarate was associated with significantly higher triglycerides to HDL-C ratio (all P < 0.02).
+In patients with first-episode schizophrenia syndrome, cardiometabolic risk factors and abnormalities are present early in the illness and are likely related to the underlying illness, unhealthy lifestyle, and antipsychotic medications, which interact with each other. Given that these risk factors become even more pronounced in chronic psychosis populations, CSC providers are presented with an opportunity to engage in prevention of cardiovascular morbidity and mortality (Srihari et al. 2013). Specific approaches include smoking prevention and cessation, counseling and lifestyle modification to prevent or limit weight gain, preferred use of lower-risk antipsychotics (Tek et al. 2015), routine monitoring, and referral to and coordination of access to appropriate medical care.
+In terms of cost effectiveness, the Net Health Benefits Approach was used to evaluate the probability that the value of NAVIGATE benefits would exceed the program’s costs relative to community care from the perspective of the health care system (Rosenheck et al. 2016). The NAVIGATE group improved significantly more on the Quality of Life Scale (QLS) and had higher outpatient mental health and antipsychotic medication costs. Effectiveness was measured as a one standard deviation change on the Quality of Life Scale (QLS-SD). The incremental costeffectiveness ratio was $12,081/QLS-SD, with a 0.94 probability that NAVIGATE was more cost effective than community care at $40,000/QLS-SD. When converted to monetized quality-adjusted life years (QALY), NAVIGATE benefits exceeded costs, especially at future generic drug prices. Notably, low-DUP and high-DUP patients had a somewhat different pattern of cost effectiveness. Among low-DUP patients, the total costs of NAVIGATE averaged $1,368 per patient per six months less than community care (14.8%; P = 0.72); among high-DUP patients, NAVIGATE showed increased costs of $3,839 (64%; P = 0.05) per patient per six months. The incremental cost-effectiveness ratio (ICER) was calculated as the difference in average annualized total costs divided by the difference in effectiveness (improvement in the QLS from baseline).
+Bootstrap analyses produced an ICER of $1,035/QLS-SD among low-DUP patients, compared to an ICER of $41,307/QLS-SD among high-DUP patients, with wide 95% confidence intervals (CIs).
+RAISE-ETP investigators performed a number of secondary analyses that shed light on some of the core processes and relationships among symptoms in early psychosis. NAVIGATE-treated patients experienced increased perceived autonomy support, which was related to improved quality of life (Browne et al. 2017). NAVIGATE treatment was also associated with a greater increase in participation at work or in school; this difference appeared to be mediated by the use of supported employment and education services. No group differences were observed in earnings or public support payments (Rosenheck et al. 2017a). Interestingly, obtaining benefits was predicted by more severe psychotic symptoms and greater dysfunction and was followed by increased total income, but it was also associated with fewer days of employment and reduced motivation (e.g., sense of purpose, greater anhedonia) (Rosenheck et al. 2017b). At the same time, during the first year of NAVIGATE treatment, tests of the bidirectional associations between motivation and social and occupational functioning suggest that motivation contributes to better occupational functioning but not better social functioning. Higher social functioning, on the other hand, predicted increased motivation. This suggests that improving occupational functioning in this population may benefit from targeting patient motivation directly (e.g., through motivational interviewing) or indirectly (e.g., by improving relationships and support networks) (Fulford et al. 2017).
+Overall, the RAISE-ETP project demonstrated that CSC could be delivered at a range of community mental health centers, and that such care was associated with significantly better outcomes in a number of different domains. Health economic analysis also indicated that overall the intervention was cost effective (Rosenheck et al. 2016). These results provided further encouragement to national efforts to make CSC more broadly accessible to patients (and their families) experiencing a first episode of schizophrenia.
+3.2.	Key Findings and Products of RAISE-IES
+The RAISE-IES study was initiated as an RCT comparing the RAISE connection model—what we would now call a CSC—to case management plus usual care. However, NIMH redirected the project in 2010 to other tasks, as described below. First, the program was implemented in two sites, recruiting a total of 65 individuals and following them for up to two years. Participants had reduced symptoms and improved social and occupational functioning over time (Dixon et al. 2015). Processing speed was identified as a significant moderator of improvement in occupational global assessment of functioning; treatment fidelity, engagement, and family involvement were found to be mediators of improvement in occupational and social functioning; and processing speed was identified as a significant moderator of improvement in occupational functioning (Marino et al. 2015). A closer examination of work and school participation revealed that individuals who engaged in vocational activity typically did so within months: 28 participants (43%) engaged in work or school at baseline, rising to 44 participants (68%) reporting vocational activity at some time in the first 6 months and 51 (78%) reporting activity in the first 12 months; only two additional participants began vocational activity after their first year of participation. Almost all participants (N = 59) met with the supported employment and education specialist at least three times (Humensky et al. 2017).
+RAISE-IES also conducted two qualitative sub-studies focusing on engagement of clients and family members (Lucksted et al. 2015, 2017). Four factors were associated with engagement of clients, including tailored care, engagement of family members, attributes of the program, and personal factors. A main factor contributing to engagement was the program’s ability to focus
+on the patients’ goals and to demonstrate that the team cared about helping individuals achieve these goals. Participants found nonclinical services such as those focused on employment and education to be a key facilitator of engagement. Other important components included shared decision making, individualized care, flexibility, and warm and respectful communication from staff (Lucksted et al. 2015). The authors concluded by recommending that teams provide recovery-oriented, flexible services that show compassion and warmth while focusing on patients’ life goals (Lucksted et al. 2015).
+The study of engagement among family members underscored that critical family member experiences of engagement included outreach, communication and support from teams, flexibility within the program model, and individualized treatment. Family members also shared their own challenges to engagement, which included personal responsibilities, lack of time and resources, and balancing the autonomy of their loved one with providing care and support (Lucksted et al. 2017). The authors concluded by recommending that teams provide families with individualized support while also helping them manage the stress related to their members’ experiences (Lucksted et al. 2017).
+The RAISE-IES project developed resources and tools to help administrators and individuals start their own CSC programs, including treatment manuals and program guides (see https://www.nimh.nih.gov/health/topics/schizophrenia/raise/coordinated-specialty-care-for-first-episode-psychosis-manual-i-outreach-and-recruitment.shtml and https://www. nimh.nih.gov/health/topics/schizophrenia/raise/csc-for-fep-manual-ii-implementation-manual_147093.pdf). RAISE-IES devised practical strategies to monitor treatment fidelity (Essock et al. 2015b), created an online interactive tool to estimate costs and resources for early psychosis care across a population (Humensky et al. 2013), and outlined approaches to financing the CSC program (Frank et al. 2015). RAISE-IES also showed it was possible to sustain a long-term program by collaborating with state mental health authorities to fund CSC services (Essock et al. 2015a). As a result, the New York Office of Mental Health (OMH) implemented the OnTrackNY initiative, a statewide first-episode psychosis treatment program which builds on the successful RAISE initiatives in New York State (Bello et al. 2017). This study demonstrated the feasibility of starting and maintaining a CSC program within the US health care system.
+As the RAISE studies were being completed and reports published, the US Congress recognized the value of CSC programs by adding 5% to the community mental health block grant program. This amounted to an additional $25 million for states and federal territories to share. Notably, the legislation required that the monies be used to develop and support evidence-based programs for individuals experiencing early psychosis. The 5% set-aside for CSC programs continued in 2015, and the allocation was doubled in 2016, providing an additional $50 million for states to share to develop CSC programs (Dixon 2017a). In 2008, only a few states had such programs. By 2016, 36 states had begun implementing one or more CSC programs. By 2018, that number will grow to 48 states (R. Heinssen, personal communication).
+4.	THE US LANDSCAPE POST-RAISE: WHAT NEXT?
+The RAISE studies are best contextualized within a two-decade-long international literature that began with observational studies of increasingly mature service interventions and resulted in a growing consensus on the principles that should inform the care of early schizophrenia (Edwards & McGorry 2002). This set the stage for a progression of experimental studies, which were necessary for translating knowledge from research into public health benefit. This project was advanced by a series of pragmatic randomized trials that retained the experimental benefit of minimizing selection bias (via randomization) while also allowing for more realistic samples, interventions,
+and patient-oriented outcome measures (Hotopf et al. 1999). The pioneering OPUS and LEO trials both tested ACT-style services with the ability to provide community outreach as well as high-intensity and well-resourced care (clinician:patient ratios of 1:10 to 1:12), and established the efficacy of comprehensive specialty care services for early psychosis (Srihari et al. 2012). The STEP RCT extended these results with a model of care designed for the constraints of a US public mental health center, with office-based care, limited outreach and clinician:patient ratios of 1:50. This trial demonstrated the effectiveness of CSC in a real-world US setting (Srihari et al. 2015), a finding that was further elaborated by the RAISE-ETP (Kane et al. 2016). Subsequent reports from RAISE have supported cost effectiveness in the United States (Rosenhecket al. 2016), again adding to the similar conclusions of the international literature on societal economic benefit (Alison et al. 2012) and strengthening consensus on the need for policy commitments to support further implementation and refinement of models of care (Fleischhacker et al. 2014, Lieberman et al. 2013). EASA in Oregon and OnTrackNY in New York State provide two examples of the increasing number of states that are attempting to disseminate CSC statewide.
+In this context, the status quo of current care systems in the United States is indefensible. Individuals with new-onset psychosis and their families face unnecessary suffering: delays to care are inordinately long (Addington et al. 2015, Compton et al. 2011), and best-practice services are not routinely available (Dixon 2017a). The stakes are high, with mounting morbidity, premature mortality (Pompili et al. 2011, Schoenbaum et al. 2017), and economic costs (Alison et al. 2012) that only partially measure the true human costs of delayed and inadequate care. This is an important and optimistic moment in US healthcare policy for vulnerable early schizophrenia patients. The community mental health block grant set-asides of 2014 and 2016 seeded the growth of CSC programs. The inclusion of this funding in the recently passed 21st Century Cures Act (http:// docs.house.gov/billsthisweek/20161128/CPRT-114-HPRT-RU00-SAHR34.pdf) has established this modest but important financial incentive within US health care policy, and it offers a backbone upon which the US implementation gap can be closed. Several influential national agencies, including the NIMH, the Robert Wood Johnson Foundation, the National Association of State Mental Health Program Directors (NASMHPD), the Centers for Medicare and Medicaid (CMS), the National Alliance on Mental Illness (NAMI), and Mental Health America (MHA), have supported wider dissemination of specialized models of care for early psychosis. The United States thus appears poised to catch up with implementations of early intervention services in other developed economies.
+Several challenges and questions delimit the potential impact of CSCs on the disease course and overall health of individuals diagnosed with schizophrenia; they will be the focus of this final section. These include both failures to implement what we know and significant knowledge gaps that require research. One way to structure these challenges is to consider the gaps and what patients need before, during, and after CSC is delivered.
+4.1.	Before Coordinated Specialty Care
+The time from onset of diagnosable illness to effective treatment is measured in months to years across mental illnesses in the United States (Kessler et al. 2005); psychotic disorders are no exception, with an average DUP ofover ayear (Addington et al. 2015). DUP has been robustly associated with poor outcomes across health care systems (Marshall et al. 2005, Perkins et al. 2005). These unacceptable delays to care occur during periods of highest risk for self-harm and aggression (Nielssen & Large 2010, Pompili et al. 2011), but they more commonly cause avoidable suffering for the affected youth and their families as they traverse chaotic and disorganized pathways to care. Therefore, maximizing the benefits of CSC requires optimized efforts to identify, refer, and
+promote engagement with CSC treatment as soon as possible after onset. [Notably, ongoing and interleaved research efforts have focused on identifying those at risk and testing approaches to prevent the onset of psychosis (Fusar-Poli et al. 2012, 2014), but consideration of that important task is beyond the scope of this review.] Early psychosis populations in any area can be divided into two groups for outreach purposes: those who are yet to seek help and those who have already come into contact with the health care system but are yet to receive CSC or best-practice care. Each group requires separate attention.
+Multiple attempts across the world to reduce DUP provide a wealth of lessons and some notable successful examples (Lloyd-Evans et al. 2011). The seminal TIPS program demonstrated that multipronged efforts that address lack of awareness (via a public information campaign) and at the same time provide clear direction on how to access responsive services can halve DUP in a large geographic sector (Friis et al. 2005). Several ongoing early detection efforts in the United States funded by a recent NIMH Request For Applications, including a quasi-experimental replication of TIPS (Srihari et al. 2014), will deliver more rigorous information on how to effect early detection and referral. Another project will address identification delays by using standard targeted provider education plus novel technology-enhanced screening and at the same time address engagement delays by using a mobile community-based, telepsychiatry-enhanced engagement team (Carter 2016). Other research studies will test methods to increase community literacy in the Latino population (Lopez 2017) and develop Internet-based strategies to reach young people through social media (Kane 2015). New York City has taken a public health approach and now requires all individuals hospitalized with first-episode psychosis to be identified and reported (https://www1. nyc.gov/site/doh/providers/reporting-and-services/notifiable-diseases-and-conditions-reporting-central.page); the city also offers a critical time intervention model staffed by a peer and a professional, called NYCStart, aimed at enhancing optimal follow-up care (https://www1. nyc.gov/site/doh/health/health-topics/crisis-emergency-services-nyc-start.page). Simon et al. (2017) have developed an algorithm to identify individuals experiencing the first presentation of psychosis using chart reviews and claims. Other strategies will need to reach into jails and prisons as well as schools and other community structures to identify and engage youth who are experiencing the onset of psychosis (Ford 2015).
+4.2.	During Coordinated Specialty Care
+This category subsumes all of the questions we have about how to implement what we know and how to expand our knowledge of what works and for whom. Although CSC has been defined as including specific care components—including medication and primary care coordination, family support and education, case management, psychotherapy, and supported employment and education—there is to date no standard CSC program and no well-validated measure of fidelity, though Addington et al. (2016) have begun this process. Although the RAISE-ETP study and the block grant’s facilitation of the national rollout of CSC have created a vast array of experiences across the chaotic US health care system, systematic knowledge regarding how to deliver CSC in different settings to different populations is lacking. Training of the workforce to deliver CSC and the development of strategies to sustainably finance it are two foundational challenges yet to be met (Dixon 2017a).
+Several approaches are available to help organize the task of spreading evidence-based care models (Aarons et al. 2011). One approach from the Institute of Medicine offers a compelling way to address the challenges of delivering care in the US system, with its myriad regulatory demands, inefficient medical record systems, and limited reimbursement for psychosocial services. Learning health networks have been proposed as a means to engender a collaborative model
+wherein “science, informatics, incentives, and culture are aligned for continuous improvement and innovation ... and new knowledge is captured as an integral by-product of the care experience” (Inst. Med. 2013, p. ix). These, or related approaches, help support CSC implementations, allow knowledge sharing, and maintain quality. This approach to creating a learning community may be employed in the government evaluation of the use of the SAMHSA block grant supplement to support implementation of CSC. Participating sites are convening to evaluate technical assistance, with the added benefit of creating contacts among the various CSC sites across the United States.
+The challenge of refining and improving CSC is no less daunting than the challenge of delivering the current best practices. The knowledge gaps are vast. To name a few, problems with cognition (Revell et al. 2015), substance use (Seddon et al. 2016), and suicidality (Coentre et al. 2017) require further attention. Cognitive remediation is not considered a required component of CSC at this point, but models are being tested and there is some evidence of effectiveness. A recent systematic review of RCTs investigating cognitive remediation after a first episode of psychosis found that one of seven neurocognitive domains showed a significant positive effect (verbal learning and memory), and five others showed borderline significant benefits. There was a significant effect on functioning (0.18; CI = 0.01, 0.36; p < 0.05) and symptoms (0.19; CI = 0.02, 0.36; p < 0.05). The effect of cognitive remediation on functioning and symptoms was larger in trials with adjunctive psychiatric rehabilitation and small group interventions (Revell et al. 2015).
+Although some studies have demonstrated reductions in hospitalization with CSC, hospitalization and psychotic relapse persist, stimulating efforts to improve the utilization of clozapine and long-acting injectable medications. Clinical trials have demonstrated the effectiveness of comparatively low doses of antipsychotic medication in the early stages of schizophrenia, with the majority of patients achieving substantial improvement in psychotic signs and symptoms (Robinson et al. 2015a). At the same time, first-episode patients are potentially more vulnerable to side effects, even with lower doses. In many cases, they are highly ambivalent about taking medication in the first place, so that tolerability and early identification and management of adverse effects become a high priority. The fact that no specific antipsychotic medication has shown to be superior in reducing positive symptoms in first-episode patients underscores the importance of selecting treatments based on tolerability. However, clozapine has shown to be effective when patients have failed two or more adequate trials of other medications, even during the first episode of treatment (Agid et al. 2011).
+The recommendations for longer-term maintenance pharmacologic treatment have come a long way since the earliest controlled trials (Kane et al. 1982) indicating that patients who had recently recovered from a first episode of schizophrenia would benefit from continued antipsychotic medication to reduce the risk of subsequent psychotic relapse. Additional studies confirmed the efficacy of antipsychotic medications in reducing the risk of relapse following a first episode of psychosis (Robinson et al. 2005). However, not allpatients will experience an exacerbation of symptoms following medication discontinuation, though the majority will (Robinson et al. 2005). At the same time, we remain hard pressed to identify the subgroup who might not require such treatment, at least during the early phase of illness. Alvarez-Jimenez et al. (2016) recently reviewed studies of treatment discontinuation in first-episode psychosis, including affective psychosis. They suggest that individuals who do not have a diagnosis of schizophrenia, achieve clinical remission for at least three months, and attain early functional recovery with strong support may be possible candidates for discontinuation of antipsychotic medication accompanied by effective psychosocial interventions. Further, there is a clear need to learn more about the adverse cardiometabolic effects of antipsychotic medications, even as they remain essential tools to manage psychotic symptoms and associated aggression and they can reduce suicidality (in the case of clozapine). Studies of CSC have also demonstrated that there is still a group of nonresponders whose care demands further research.
+Overall, it is important not to permit awareness of the benefits of CSC to prevent consideration of the well-known heterogeneity in early psychosis samples in terms of prognosis (without treatment) and of responsiveness to available treatments and to efforts at early detection. A one-size-fits-all approach based on average effects from even rigorously conducted clinical studies risks over- and undertreating different subgroups and delaying identification of those who are refractory to current best practice. Careful ascertainment of sociodemographic and clinical characteristics can be leveraged in predictive models to allow us to determine what works for whom. Also, emerging knowledge of distinct etiologies and pathophysiologies currently categorized within the schizophrenia spectrum may yield more personalized treatments. Moreover, we need better-validated measures of functional outcome or community adaptation to help define the value of CSC for affected youth and their families, but also society at large. Whereas composite measures such as QALY allow comparisons across medical conditions, these may not be adequately sensitive to meaningful changes in the state of individuals with psychotic illnesses (McCrone 2011). A panel of measures that assess distress, impairment, and disability will likely be necessary to evaluate the societal value of early intervention services and to calibrate the level of policy support for wider dissemination of such services.
+An additional question that inevitably arises when considering the implementation of CSC is to whom it should be offered. As discussed above, the RAISE-ETP study showed much greater benefits for individuals with DUP of less than 74 weeks (Kane et al. 2016). Here, DUP was defined as the time between onset of psychosis and exposure to antipsychotics in a sample in which participants did not have more than six months of total exposure to antipsychotic medication (Addington et al. 2015). DUP varies widely across the many early psychosis studies, which differ in inclusion criteria as well as definitions and assessment strategies for DUP, making cross-study comparisons difficult (Golay et al. 2016). Some CSC programs offer services only to individuals within a specified time of illness onset (e.g., two years) regardless of previous treatment, which will by definition cap the DUP of the individuals served (Bello et al. 2017). There is no evidence to support a DUP after which the CSC model has minimal value over usual care. Three months is a commonly accepted DUP target (Cotter et al. 2017) but the fact that CSC was compiled from treatment known to be effective in chronic schizophrenia suggests such team based, specialty care models may benefit patients later in the illness course.
+Another common question facing policy makers is whether to offer CSC to individuals with affective psychotic disorders such as major depression and bipolar disorder. Arguments against this decision are that the CSC research has largely focused on schizophrenia-type disorders, the benefits of CSC in other psychotic illnesses are less well tested, and the impact of DUP is less clearly delineated. At the same time, other scholars argue that it is very difficult to differentiate these illnesses in youth, and there is not likely any specificity to the benefits of this comprehensive teambased model for all young people with psychosis. A policy framework that focuses on providing evidence-based treatment to youth with behavioral health care disorders at their earliest phases rather than focusing on specific disorders may be the most coherent population-based approach.
+4.3.	After Coordinated Specialty Care
+The issue of patients’ needs after CSC includes a consideration of how long CSC should last and what young people experiencing psychosis and CSC care need in an ongoing way. Followup studies of several CSC RCTs to date, including OPUS and LEO, suggest that the benefits observed at the time of program completion are not sustained 5 and 10 years later (Bertelsen et al. 2008, Gafoor et al. 2010, Secher et al. 2015, Sigrunarson et al. 2013). Interestingly, the TIPS study that focused on early detection did observe greater rates of recovery in the early-detection versus
+usual-detection group after 10 years (Hegelstad et al. 2012). Would extending the duration of CSC programs mitigate the erosion of benefits? The Prevention and Early Intervention Program for Psychoses in Ontario provided extended continuity of lower-intensity care for three additional years after the two-year standard CSC program (Norman et al. 2011). Scholars examining the program found that the improvements observed at two-year follow-up were maintained at five years, with ongoing improvement in global functioning. Chang et al. (2015, 2017) performed an RCT in Hong Kong that compared individuals who had a one-year extension of the two-year CSC program called EASY to individuals who got stepped-down care. Individuals with extended EASY had improved outcomes in numerous domains immediately after the one-year extension (Chang et al. 2015), but there were no group differences one year later (Chang et al. 2017). Another RCT compared individuals who had received two years of OPUS followed by usual treatment with individuals who had received five years of OPUS. Group differences were limited to increased likelihood of remaining in contact with specialized mental health services, higher client satisfaction, and stronger working alliance (Albert et al. 2017). Overall, the treatment extension and follow-up studies do not reveal uniform findings. There are signs indicating that ongoing treatment produces persistent benefits, whereas evidence of the persistence of such benefits after CSC is lacking. There are many possible explanations for these findings, including sampling and attrition issues, variability in the quality of treatments compared, limited implementation of CSC treatment, and variability in DUP, to name a few. More research is clearly needed on the overall optimal length of CSC and what should come next.
+The overall failure to produce sustained benefits in the aftermath of CSC treatment presents the field with an enormous challenge. Alvarez-Jimenez et al. (2013) have developed an online approach called HORYZONS that uses expert moderation and “super-users” (peer moderators) to provide follow-up care to young people as they are completing their course of CSC treatment at the Orygen program. This is being tested in a randomized trial. CSC programs are focusing on developing approaches to step down and follow up, including ongoing vocational and educational supports, family education, and alumni groups (personal communication, T. Sale, EASA). The heterogeneity of responses to CSC demands tailored solutions. Within the US health care system, the separation of CSC programs from the overall delivery system likely hinders the seamless integration of CSC into optimal longitudinal care. The transformation and integration of CSC programs into learning health networks may produce self-correction of CSC practices as systems learn what works and what does not, and it may perhaps contribute to an overall improvement of usual care.
+5.	CONCLUSION
+In the end, evidence is mounting for the positive impact of CSC on a range of outcomes for individuals in the early stages of psychosis. The studies hint at the benefits of early intervention to reduce DUP and to improve outcomes. To some extent, implementing CSC within a system of mental health services increases its capacity to provide evidence-based care for individuals at any stage of a psychotic illness. Earlier treatment means earlier benefits in terms of immediate outcomes but may not improve longer-term outcomes and prevent disability. The ultimate promise of prevention of long-term disability, which has motivated so many of the cycles of mental health service reform in the past, remains elusive. CSC programs have established their value in improving early outcomes; they should be available as standard care for new-onset psychosis and can provide a humane and rigorous platform upon which to build further studies, develop new treatments, and refine the delivery of services. Doing what we know works can thus support ongoing research to answer lingering questions and to avoid paralysis in the face of important uncertainties.

Annual Review of Clinical Psychology.txt

@@ -0,0 +1,102 @@
+INTRODUCTION
+Depression is one of the most frequent causes of disability and lost workdays worldwide, and women are more likely than men to suffer from this common mental disorder (Hasin et al. 2018, Kessler 2003, Marcus et al. 2012). In the United States, women are twice as likely to suffer from depression than men, and approximately 25% of women will be diagnosed with depression in their lifetime (Hasin et al. 2018, Kessler et al. 2003). Depression profoundly affects social functioning and results in lower rates of labor force participation, reduced work hours, and lower earnings (Bland et al. 1988, Jayakody & Stauffer 2000, Lerner & Henke 2008). Such reduction in one’s capacity to be part of the labor force results in the economic deprivation and financial hardship that begin to define poverty, which also includes social, political, and cultural factors (UNESCO 2019), such as age, class, and race/ethnicity.
+In the United States, women are more likely than men to live in poverty and are more likely than men to have food insecurity, inadequate nutritional intake, unstable housing, partner conflict, and other difficulties that affect mental health (Chant 2006, Edin & Kissane 2010). Managing sporadic income and making difficult decisions on purchasing and basic needs also increase cognitive load (often referred to as the amount of attention and working memory required for a task) (Mani et al. 2013). As these concerns preoccupy daily life, cognitive resources available to guide choice, behavior, and emotional regulation are reduced. Such stressors also increase allostatic load or “wear and tear on the body,” as characterized by McEwen & Stellar (1993, p. 2094; see also McEwen 2003), and the neurobiology of regions such as the hippocampus, amygdala, and prefrontal cortex undergoes structural remodeling, altering behavioral and physiological responses and making it even more difficult to function (McEwen et al. 2015, Nasca et al. 2017).
+Another main factor associated with poverty and depression in women is parenting status (England & Sim 2009). Although the intersectionality of depression, parenting, and poverty in women has been acknowledged in the literature and reported across diverse geographical regions, societies, populations, and social contexts, there is limited literature that explores the links among mental health, parenting, and economic stability for women. Moreover, data are limited on how these life-altering factors relate to a broader intervention and policy agenda.
+This review summarizes the mental health and economic literature regarding how maternal depression intersects with intergenerational poverty. We provide a conceptual model asserting that treatment of depression and integration of its treatment into social services systems with employment opportunities can improve work productivity and enhance the capacity to care for one’s family. Finally, this review discusses challenges and recommendations regarding interventions and policies to treat maternal depression in large-scale social services systems.
+The review utilizes three theories from social epidemiology that highlight the relationship between depression and economic status: social causation, social selection, and interactionist. According to the social causation theory, environmental and societal conditions lead to increased risk of depression. By contrast, the social selection hypothesis suggests that individual differences in the likelihood of depression influence the likelihood of employment and impact potential for earnings, making it more likely that a person with depression will be poor. The interactionist hypothesis combines the two theories of social causation and social selection and posits that individual differences influence economic outcomes, which in turn have impacts on depressive symptoms (Conger & Donnellan 2007, Wadsworth & Achenbach 2005).
+We recognize that the causes of depression are multifactorial and include a combination of psychosocial, environmental, genetic, cognitive, and neurobiological factors and that an understanding of the complex and multidimensional nature of poverty and depression is necessary. We also acknowledge the need for public health, population-based approaches that address the fact that on average, women are twice as likely to be diagnosed with major depression compared with men over all ages and nations (Hyde & Mezulis 2020). Addressing genetic and neurocognitive and developmental factors at a population level remains difficult. As such, this review focuses on the research addressing relationships between maternal depression and economic status while considering the social causation, social selection, and interactionist theoretical frames. Included in this review are examples of programs intended to improve economic status and research on the effects of treating depression in low-income pregnant and parenting women, with resultant economic benefits (potential evidence for the social selection theory). The interactionist hypothesis and the policy implications of the reported relationship between depression and economic status specific to low-income mothers also are reviewed and discussed with a call for additional research that can help to establish the optimal sequencing and combination of depression treatment and poverty alleviation interventions and policies.
+Current evidence indicates that all major racial and ethnic groups have reductions in employment associated with poor mental health (Demirhan & Demirhan 2019). Additionally, most of the research cited in this review is based on interventions in and policies of the United States. Although a few studies from other countries are included, addressing global policies is beyond the scope of this article. Compared with other countries, the United States has distinct policies, including employment and economic policies (Cambron et al. 2015). For example, even though parents and caregivers are working longer hours, America’s child poverty rate is twice that of most wealthy countries (Hardy et al. 2018). For these reasons, the conclusions in this review are generalizable within the context of US policy affecting low-income pregnant and parenting women.
+POVERTY AND WOMEN IN THE UNITED STATES
+Women may be particularly influenced by the causes and effects of poverty, and women’s experience of poverty differs from that of men (Aydiner-Avsar & Piovani 2019, Chant 2006). Chant (2006) described three factors that contribute to women’s poverty relative to men. First, women have fewer possibilities to translate work into income because of (a) their extensive responsibility for reproductive, caregiving, and domestic roles, including cleaning, cooking, and child care; (b) the conceptualization of their productive activities as “helping” men; and (c) their concentration within sectors that are either an extension of their reproductive roles (and thus lower paid) and/or within the informal economy (Edin 2000, Edin & Kissane 2010). Second, even when women earn wages, family structures and social norms often interfere and influence women’s decision-making capacity and decisions on how income is used. When women do make economic decisions, they are less likely to make decisions that improve their personal well-being (Edin 2000, Edin & Kissane 2010, Oliker 1995). Third, economic resources that enter the household via women are more likely to be spent on household and children’s needs. In addition to these gender differences, there is evidence that the presence of major depression is more strongly associated with job loss in women than in men (Andreeva et al. 2015, Martínez et al. 2020).
+POVERTY AND PARENTING
+Longitudinal studies with large samples support the conclusion that alterations in the quality of caregiving are one pathway by which poverty adversely impacts child development. Support from friends and family can improve the parent-child relationship in the context of poverty (Elder et al. 1985, London et al. 2004, Lundberg & Pollak 2007, Moore et al. 2017, Perry et al. 2019, Zaslow et al. 2005). A large literature demonstrates that parenting quality in stressful circumstances, such as those of scarcity and ill health, influences children’s biology and behavior. The detrimental effect of poverty in childhood on health and well-being has been widely documented (Aber et al. 1997, Caughy et al. 2003, Wood 2003), and researchers have argued that economic disadvantage increases the chances that children will fail to thrive (Shaefer et al. 2018). However, recent research demonstrates that although poorer households have poorer health, the impact of income is relatively small compared with the impact of the mother’s own health and parenting quality, which plays a much larger role in determining child outcomes (Ciciolla et al. 2017, Luthar & Ciciolla 2015, Perry et al. 2019, Tirumalaraju et al. 2020, Washbrook et al. 2014).
+MECHANISMS LINKING DEPRESSION AND ECONOMIC MOBILITY IN LOW-INCOME PREGNANT AND PARENTING WOMEN
+At the population level, depression has been associated with work absenteeism, impaired work performance, and increased health care costs for employers (Fournier et al. 2015, Mojtabai et al. 2015, Moussavi et al. 2007). Among mothers, depression is the mental health problem most likely to be associated with poverty. Lower-income mothers are more likely to be depressed than higher-income mothers (28% versus 17%, respectively) (Golin et al. 2017), and depressive symptoms are four times more common among lower-income women who are parents than among middleincome mothers (Green et al. 2016).
+Particular to women who are pregnant or parenting, depression and depressive symptoms have also been shown to have a negative impact on the transition from welfare to work (Bailey & Danziger 2013, Danziger et al. 2001) and subsequent lack of employment (Mojtabai et al. 2015, Whooley et al. 2002). In a longitudinal study of 2,235 nationally representative mothers, those who reported a poverty-level income were more likely to have high depressive symptoms than
+the women who were never below the poverty level (Pascoe et al. 2006), and in several samples of parents who applied for social services to aid the poor, close to half of the parents had clinically significant depressive symptoms (Fuller & Kagan 2000, Gupta & Huston 2009, Pavetti et al. 1996, Quint 1994).
+Policymakers are focused on enhancing women’s economic status through increasing employment in the paid workforce as lack of employment has also been found to contribute to an increased risk of major depressive disorder (Dooley et al. 1994, Kessler et al. 1989, Pieters & Klasen 2020). Traditionally, the focus of increasing women’s employment has been on social factors that may impact employment (e.g., child care) and other structural barriers like transportation and flexible scheduling. Factors that impact an individual woman’s ability to participate in the workforce, such as level of education and training, are considered by many policy makers, yet policies and many studies on women’s economic advancement make little mention of psychological difficulties (Cambron et al. 2015). Depression is an important barrier to economic advancement and to willingness to enter the labor force (Mossakowski 2009), and the existence, duration, and age of onset of depressive symptoms may prevent some pregnant and parenting women from leaving welfare for work in a timely manner (López-López et al. 2020). Despite the lack of focus on the mental health problems of women receiving social services for the poor, recent research indicates that women receiving welfare assistance may experience higher levels of depressive symptoms and general psychiatric distress than the general population and that this distress can affect economic self-sufficiency. Psychological factors thus play a critical role in the success of economic and social policy efforts and are often overlooked in the economic opportunity landscape for pregnant and parenting women (Coley et al. 2007, Danziger et al. 2001, Dooley & Prause 2002, Gibson et al. 2018).
+Poverty and depression are likely to be bidirectional in terms of causation and are hypothesized to operate in a cycle that perpetuates poor economic and psychiatric outcomes (Lund et al. 2010). The onset of mental illness may increase the risk of poverty (social selection or drift), and conversely, the experience of poverty may increase the risk of depression (social causation). However, it could be that the cycle of poverty and depression is linked to a third set of factors related to the intersection of poverty, gender, and mental illness, such as exposure to violence, access to treatment and health care, and chronic medical conditions (Ridley et al. 2019). Another hypothesis is that poverty leads to stress and negative affect (social causation) and that, in turn, stress and negative affect increase risk aversion, which could make it more difficult to take the steps needed to escape poverty (Haushofer & Fehr 2014).
+Research on the epidemiology of depression finds a consistent and robust relationship between depression and socioeconomic status as measured by income, education, and employment status (Gariépy et al. 2016). Researchers have argued that this relationship is the result of lower-socioeconomic-status individuals experiencing a greater number of stressful life events and having fewer financial resources to buffer the impact of the stressors (Lorant et al. 2003). In one of the most comprehensive reviews on the topic, Lund et al. (2010) surveyed 115 studies and found that although the direction and strength of the poverty-mental health relationship vary across studies, taken in its entirety, the evidence suggests that some aspects of poverty (e.g., lower education, food insecurity, financial stress, lower socioeconomic status) are consistently related to depression, and the association between depression and other measures of poverty, such as income and employment, is less clear. Specific to depression, those with more assets may be less likely to experience depressive symptoms, as assets provide financial resources that can be used to buffer the impact of stressful life events. Assets also may have a positive effect on depressive symptoms because they reduce economic pressure on individuals and offer more opportunities (Enns et al. 2016, 2019; Rohe et al. 2017).
+www.annualreviews.org • Mental Health and Wealth 185
+Specific to low-income pregnant and parenting women, it is possible that having a job reduces the probability of having depressive symptoms, while the lack of employment results in an increased likelihood of depressive symptoms (Richard & Lee 2019). An alternative interpretation suggests that depressive symptoms may prevent low-income pregnant and parenting women from undertaking the tasks necessary to find employment or that parenting women with depressive symptoms may lack the agency and sense of efficacy needed to take on new challenges. Even after someone has obtained employment, depressive symptoms can play an important role in selfsufficiency outcomes. Some pregnant and parenting women may succeed in obtaining employment but have difficulty keeping their jobs or performing them effectively because of depressive symptoms that interfere with daily functioning (Jayakody & Stauffer 2000) and work-child care balance. Depressive symptoms and disorders affect a woman’s productivity and social functioning: The degree of impairment is statistically comparable to the impairment associated with chronic medical conditions (Aydiner-Avsar & Piovani 2019, Demirhan & Demirhan 2019, Raver 2003). Depression can also play a role in the success of education and job training programs because those suffering from depression are more vulnerable to interpersonal problems and irritability and may experience diminished social functioning (Schless et al. 1974, Seedat et al. 2009, Weissman et al. 1971). In the most severe forms, depression can make job search and work participation impossible. Furthermore, the experience of poverty among low-income pregnant and parenting women means that their children face the related dimensions of disadvantage and the environmental stressors associated with living in poverty.
+Social Causation Hypothesis
+The associations between economic circumstances and depressive symptoms in mothers are well documented, but important questions remain regarding fundamental causal processes. We focus on three hypotheses to frame how depressive symptoms are associated with economic outcomes in women and resultant interventions and policy approaches (Gupta & Huston 2009, Marcus et al. 2012): (a) low economic status causes depression (social causation) (Aydiner-Avsar & Piovani 2019); (b) depression causes low economic status (social selection) (Blane et al. 1993); or (c) there is an ongoing bidirectional bridging relationship between economic circumstances and depression, with each affecting the other (Bruce et al. 1991, Conger & Donnellan 2007, Dohrenwend & Dohrenwend 1969, Schofield et al. 2011).
+According to the social causation model, environmental, sociopolitical, and job loss and income declines precipitate depression. From this perspective, taken at its extreme, if all women were exposed to the same social environments from birth, they would achieve a similar level of economic success. If this hypothesis held true, low-income mothers who faced significant adversity, discrimination, and other stressful circumstances would be likely to develop depressive symptoms. Although the cross-sectional nature of many studies prevents us from disentangling the causal direction of the high rates of depressive symptoms experienced by lower-income women compared with the general population (Morris 2008, Ribeiro et al. 2017, Silva et al. 2016), data on the links between early stressful life experiences and job loss and income declines lend support to the social causation framework. Natural experiments have demonstrated that loss of employment or income reduces mental health (Pierce & Schott 2020) and that large income increases improve mental health (Apouey & Clark 2015, Lindqvist et al. 2020, Wolfe et al. 2012). One randomized experiment in Oregon found that receiving health insurance reduced rates of depression by about a quarter among low-income individuals (Finkelstein et al. 2012). Longitudinal data with controls for individual characteristics and repeated measures allow for an examination of the relationship of depressive symptoms and economic outcomes across time.
+Support for the social causation hypothesis is evident in one of the few longitudinal studies that simultaneously tracked family income, parenting style, and child outcomes using US cohort data as analyzed by Dearing et al. (2004). These authors found that reductions in income were significantly associated with maternal depression in the first 3 years of children’s lives. Furthermore, they observed that it was the stress of poverty that caused depression (rather than the other way around) and that depression was likely to result in harsher and/or more inconsistent parenting.
+Additionally, nationally representative samples found that earning a low income or being unemployed when in a low-income bracket appears to increase risk for depressive symptoms (Dooley et al. 1994). In a sample of low-income women from the Project on Devolution and Urban Change, those who were not employed had a higher risk of depressive symptoms, regardless of whether they received welfare assistance, compared with those who were employed (Polit et al. 2001). In the population-based US National Household Survey on Drug Abuse, low-income mothers had significantly higher rates of poor mental health compared with higher-income mothers, and the percentage of women with major depression in the not-working group was higher than in the working-at-all group (Jayakody & Stauffer 2000).
+Social Selection Hypothesis
+The social selection hypothesis posits that characteristics of individuals, including genetic composition and cognitive and behavioral attributes, predispose some individuals to poor mental health that leads them to reduced earnings and employment over the life course (Mojtabai et al. 2015, Whooley et al. 2002). According to the social selection hypothesis, a mother with depressive symptoms would be unable to obtain stable employment because of her psychological distress, including the motivation or ability to seek a new job, and her resultant lack of earnings would prevent her from escaping poverty (Dooley et al. 2000, Mossakowski 2009).
+Specific to depressive symptoms, the social selection hypothesis posits that clinically significant levels of depressive symptoms may lead to lower earnings and/or increase the likelihood of unemployment. Unemployment and lower earnings could then result in an increase in use of social services programs that provide aid to the poor (Lerner & Henke 2008). For example, Noonan et al. (2016, p. 201) found that the presence of maternal depressive symptoms during the first year of a child’s life “increases the likelihood that children and households experience food insecurity” from 50% to 80% by the time the child is 2 years old. Additionally, Noonan et al. (2016) found that elevated levels of maternal depressive symptoms increased the likelihood of enrollment in social services programs that aided the poor, including the Supplemental Nutrition Assistance Program (SNAP), Medicaid, and Temporary Assistance for Needy Families (TANF). In a representative population-based survey, the National Longitudinal Survey of Youth, mothers rated as at risk of depression on the Center for Epidemiologic Studies Depression Scale (CES-D) (a depressive symptom screening tool) were significantly more likely to enroll in cash assistance for the poor at a 2-year follow-up point compared with mothers with lower CES-D scores at baseline (Dooley & Prause 2002).
+It is also possible that high levels of depressive symptoms in mothers might conversely lead to reduction in the receipt of welfare benefits as depressive symptoms may interfere with a mother’s ability to adhere to requirements, such as employment and training requirements, of social services programs that aid lower-income families. Support of this theory was found in a cross-sectional study where mothers with positive depression screens were more likely to have been sanctioned for not meeting the participation requirements of a welfare program in the past 12 months compared with mothers with depressive symptoms (Casey et al. 2004, Lindhorst & Mancoske 2006).
+www.annualreviews.org • Mental Health and Wealth 187
+Studies of young mothers with children enrolled in the federally funded Head Start program have found that mothers with higher depressive symptoms at baseline reported lower future earnings compared with mothers with lower depressive symptoms at baseline (Raver 2003). In another study of mothers with depressive symptoms participating in 20 federally funded welfare-to-work experimental programs, welfare-to-work programs increased earnings less for the most depressed mothers than for the moderately depressed and the least depressed over a 3-year period (Bloom & Michalopoulos 2001).
+Although not specific to mothers, it is worth noting that recent data from the Avon Longitudinal Study of Parents and Children (ALSPAC) identified the chronicity, recency, and level of depressive symptoms in early childhood and adolescence as predictors of poor educational attainment and low income in early adulthood (López-López et al. 2020). Similarly, in another longitudinal sample of low-income employed mothers, those with high levels of depressive symptoms at the onset of the study had increased odds of unemployment during the subsequent 5 years compared with those not at risk for depression (Whooley et al. 2002).
+Regarding TANF programs that maintain the strictest restrictions on work participation and sanctions for violations of TANF policy, a recent study found that low-income single mothers in receipt of TANF in states with the most stringent work requirements were much more likely to have worse mental health than their counterparts living in states with flexible work requirement and sanction policies (Davis 2019). A 2007 study that focused on low-income mothers with depressive symptoms when they first received welfare found that depressed mothers were less likely to report that they engaged in job search activities compared with those who did not have depressive symptoms at baseline (Zabkiewicz & Schmidt 2007).
+Possible pathways by which mental distress leads to reduced income include an inability to procure skills, training, and social benefits due to diminished energy and higher levels of discouragement (Krueger & Mueller 2011), poor physical health resulting from increased psychological distress (Scott et al. 2016), and a change in family structure or environment (housing instability) due to depression and, subsequently, a reduction in household resources (Cambron et al. 2015). In a study by Whooley et al. (2002), people at risk of depression at baseline were almost twice as likely to have low income (<$25,000 in 1995-1996) 5 years later compared with those without risk of depression. Unfortunately, these results were not disaggregated by gender.
+This research demonstrates the association of depression and low employment. Although depression may overlap with other personal traits and social factors, it is reasonable to expect that effective treatment of depressive symptoms could help women seek and maintain employment and increase earnings (Ridley et al. 2019).
+Interactionist Hypothesis
+The interactionist model conceptualizes reciprocal influences of mental health, wealth, and wellbeing by incorporating both social causation and social selection (Conger & Donnellan 2007). In this context, the term interactionist means bidirectional. Figure 1 provides an illustration of our proposed model to guide research characterizing the links among health (particularly mental health), wealth, and social and economic well-being for pregnant and parenting women. In our model, both (a) characteristics of depressed mothers and (b) social opportunities and threats affect each other in an interactive way.
+Conger & Donnellan (2007) used the term interactionist to describe reciprocal or bidirectional processes, although they did not necessarily examine interactions or statistical techniques of moderation. The authors noted that both individual differences and conditions of the social and economic environment affect economic well-being. They proposed that individual cognitive
+and personality characteristics affect the likelihood of attaining high or low socioeconomic status as an adult. In turn, adults’ socioeconomic status most likely contributes to depressive and other psychiatric symptoms. For the purposes of this review, the model can be used to understand the consequences of an integrated, aligned intervention that improves women’s depressive symptoms and economic status, increases social and economic mobility (e.g., education, employment, social capital), and improves outcomes for their children.
+A bidirectional negative relationship between major depressive disorder and employment has been found in several studies (Andreeva et al. 2015, Dooley et al. 2000, Olesen et al. 2013). In a longitudinal analysis of low-income women participating in an employment-based antipoverty program from 1994 through 1998, Gupta (2006) examined depressive symptoms and earnings for women at two points in time across 3 years. Women who worked more hours and had higher incomes reported a larger decline in depressive symptoms from time 1 to time 2 compared with women who worked fewer hours and had lower incomes (Gupta 2006). In support of the social causation theory, there was a trend in Gupta’s study suggesting that women with lower welfare receipt and higher earnings had lowered depressive symptoms from time 1 to time 2. The social selection hypothesis was also supported because women with lower levels of depressive symptoms at time 1 were more likely than those with higher levels of depressive symptoms to have reductions in welfare receipt and increased incomes over the subsequent 3 years (Gupta 2006).
+Although not focused solely on depression, several studies have examined the impact of interventions for trauma and interpersonal violence in a TANF context. A study by Mascaro et al. (2007) detailed the complicated interaction between mental health and employment through an examination of depressive symptoms in women who had reported interpersonal violence and suicidality and were involved in a trauma intervention. At 6 months and 1 year after completion of the intervention, women who had gained employment were less likely to be depressed on a depression symptom screener compared with women who had remained unemployed or lost their employment. When the authors controlled for baseline employment status, this initial finding was attenuated: Women with high depressive symptoms at baseline were more likely to lose employment and less likely to gain employment over the course of the yearlong study compared with those women with low levels of depressive symptoms at baseline. The importance of examining subgroups of women has been noted in additional studies of women with high levels of trauma symptoms. Exposure to early trauma and adversity was associated with long-term unemployment
+in a sample of low-income women, with the mechanisms that helped explain these associations being depressive symptoms (Cambron et al. 2015). Research on the social stigma associated with receipt of welfare also has highlighted the interactions that occur and can be statistically assessed in the relationships between receipt of welfare, depressive symptoms, and employment. A theoretical body of work on social status and psychological distress has identified a perception or “signal” of low social rank associated with low income as the primary mechanism to increase depression among low-income populations. Recently, Pak (2020) used data from the 2008-2014 Health and Retirement Study to examine depressive symptoms associated with food stamp participation and noted that the stress and stigma of receiving benefits were mechanisms identified in increasing the risk of major depressive disorder for men and not women.
+Overall, the longitudinal research presented supports the interactionist hypothesis, which suggests a cascading effect of economic circumstances affecting depressive status, which in turn affects future economic mobility. Research on the relationship between mental health and economic outcomes could be amplified to test the interactionist hypothesis in understanding how depressive symptoms and economics interact with one another over time.
+INTERVENTIONS BASED ON THE SOCIAL CAUSATION HYPOTHESIS
+In this section, we examine evidence testing the social causation hypothesis to uncover how earnings, history of welfare receipt, employment, and income affect or predict depression and depressive symptoms in pregnant and parenting women.
+Temporary Assistance for Needy Families
+The Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996 created a shift from a welfare system based mainly on the provision of cash assistance without time limits to one requiring employment and imposing other participation criteria as well as time limits on cash assistance. One major goal of this legislation was to move single parents into the workforce. The changes in federal welfare laws in the United States resulted in a dramatic decrease in the welfare rolls and an increase in single mothers entering the workforce (Danziger et al. 2001, Mueser & Troske 2003, Slack et al. 2007). The shift in welfare laws also catalyzed an examination of structural barriers to employment faced by low-income single mothers. PRWORA was based on the key assumption that most welfare recipients could gain employment, but some researchers noted that depressive symptoms and other psychiatric problems would pose significant barriers to gaining and maintaining employment (Danziger et al. 2001, Hall et al. 2017, Jagannathan et al. 2010, Moore et al. 2017).
+Although welfare receipt provides cash assistance, there is little evidence that the intervention improves mental health. Many studies suggest that women who are current or former recipients of cash assistance have higher levels of depressive symptoms compared with their counterparts who have never received cash assistance (Dooley & Prause 2002). Two larger studies have documented that the effects on depressive symptoms for women when leaving welfare for work appear to be mixed. In the Three-City Study (Coley et al. 2007), which followed close to 2,000 low-income single mothers in Boston, Chicago, and San Antonio across two waves (1999 and 2001), mothers who became employed or remained employed across both waves showed reduced depressive symptoms compared with mothers who left work or never became employed.
+In the second study, the Minnesota Family Investment Program (MFIP), different components of welfare and their effects on family outcomes were examined by researchers prior to the 1996 federal welfare reform legislation. In a cohort of 879 mothers, Gennetian & Miller (2002) found
+that MFIP increased employment rates, decreased poverty, and decreased maternal depressive symptoms. Through an experimental design, the authors found that incremental increases in income for mothers reduced depressive symptoms when compared with a control group that did not receive these increases.
+Employment
+The most rigorously designed experimental studies of policies to increase employment in low-income parenting women have demonstrated little impact on reducing depressive symptoms. Zaslow et al. (2001) reviewed 18 sites and a total of seven experiments across sites prior to the changes made in federal welfare regulations in 1996. Morris (2008) analyzed the same 18 sites several years later and noted that any impact on maternal depressive symptoms depended on both the age of the children in the family and the type of program tested. Women with school-age children demonstrated reduced depressive symptoms depending upon the program, whereas for women with preschool-age children, the programs increased depressive symptoms, but these effects depended partly on the policy tested. Of three programs that offered financial incentives for work (but did not mandate employment), two reduced depressive symptoms and a third had no impact according to follow-up measurements. Among parents of preschool children, programs that emphasized rapid employment were the most likely to increase maternal depressive symptoms (Morris 2008). The latter finding could be partially explained by the fact that the stressors associated with increased employment, such as the need for transportation and child care and maternal concern for the well-being of children, outweigh any beneficial effects of increased earnings on maternal mental health (Chase-Lansdale et al. 2003, Edin & Kissane 2010).
+Increasing Earnings and Wages
+Increased wages from employment appear to lessen depressive symptoms in women. For example, the work trajectories of women participating in an employment-based antipoverty program were categorized across 2 years (Yoshikawa et al. 2006). While controlling for family demographics and work experience, the study showed that those with full-time employment and wages that increased in value reported lower levels of depressive symptoms compared with women in either the parttime low-wage employment or the rapid-cycle (in and out of jobs) groups. Specifically, CES-D scores were on average five points lower in the full-time employed group with wage growth compared with the two other groups of women. Working more hours was also associated with lowered depressive symptoms. Raver (2003) found that mothers’ increased work hours over a period of several months predicted lowered depressive symptoms.
+Strong associations between income and mental health are reported in cross-sectional and longitudinal analyses, but the evidence concerning causal direction is less consistent (Gugushvili et al. 2019, Platt et al. 2016, Zimmerman & Bell 2006). The varied findings are likely due to different samples, measurement of income and depressive symptoms, and consideration of gender.
+Individual Development Accounts
+Individual development accounts (IDAs) have been posited to help the poor (and, specifically, poor women) develop assets, which, in turn, would have a number of economic, social, and psychological benefits for families. IDA programs provide participants with incentives to save for the purchase of specific assets, such as a home, an education, or the development or expansion of a business. If the participants’ savings are used to purchase a program-approved asset, those savings are matched
+with program funds. IDA programs typically require program participants to take both general financial literacy training and asset-specific financial education courses, such as home ownership education or small business management. Federal funding was allocated to support IDA programs with the enactment of the Assets for Independence Act (AFIA) in 1998. The Assets for Independence Program (Mani et al. 2013) is now the largest funding source of IDAs in the United States, with sponsored IDA programs in 49 states and the District of Columbia. Yet few studies have directly examined the impacts of IDAs on depression. One of the largest studies used longitudinal data collected as part of the American Dream Demonstration experiment, in which applicants to a large IDA program were randomly assigned to either an IDA program or a control group (Rohe et al. 2017). Assignment to the IDA program was not associated with reduced depressive symptoms; rather, the value of assets and perceived financial stress were inversely associated with depressive symptoms at 10-year follow-up. Results by gender are difficult to disentangle because the experiment did not specifically focus on women.
+Another type of IDA, which has typically been more targeted at pregnant and parenting women, is a child development account (CDA)—a type of asset-building account created for children at birth. In Oklahoma, primary caregivers of children born during 2007 were randomly offered a CDA (n = 1,358) or no CDA (control group; n = 1,346). Baseline and follow-up surveys measured the participants’ depressive symptoms with a shortened version of the CES-D and found that CES-D scores for the CDA group were significantly lower than for the control group when controlling for baseline CES-D score (Huang et al. 2014). Although often framed as an economic intervention for children, CDAs may improve mothers’ psychological well-being, an effect that may be partially mediated through changes in children’s social-emotional development.
+Earned Income Tax Credit
+The Earned Income Tax Credit (EITC) is a refundable tax credit that has lifted millions of families out of poverty (Simon et al. 2018). The credit provides a subsidy as a percentage of income and thus effectively increases the wages of the working poor. A broad base of research suggests that the EITC improves health outcomes and that its most robust results are seen among single mothers and children (Gangopadhyaya et al. 2019). The specific mechanism for these improvements is a reduction in maternal stress (Simon et al. 2018); other examined pathways include improvements in health insurance coverage and employment for mothers (Gangopadhyaya et al. 2019). Specifically, one study examining the impact of the EITC among lower-income mothers found increases in happiness and feelings of self-worth as a function of EITC receipt (Boyd-Swan et al. 2016). In this study, a reduction in self-reported symptoms of depression was found in married mothers but not in single mothers. Given the overall positive effects on health from the EITC, there have been recent calls to expand its use specific to low-income pregnant and parenting women. Such efforts have called for lower-income pregnant women to become automatically eligible for the EITC (Simon et al. 2018). Extant evidence suggests that expanding the EITC with a focus on mothers may be more likely to improve health than expansions focused on fathers or single men, but this conclusion may reflect a need for more research to uncover whether male health remains unchanged following receipt of EITC (Evans & Garthwaite 2014).
+It also is worth noting that in Canada, there is a Canada Child Benefit (CCB) paid to parents of children aged 0-17 years. Unlike in the United States, benefits do not depend on earned income specifically, so families with no income still qualify for the benefit, and there is a National Child Benefit (NCB) that is province-specific in implementation. Milligan & Stabile (2011) used data on child benefits across province, time, and family type to study outcomes spanning test scores and maternal and child mental and physical health. Their findings suggested that child benefit
+programs in Canada had significant positive effects on both children’s test scores and maternal mental health.
+The research on tax benefits and child benefits indicates broad benefits to maternal mental health and child outcomes in most developed countries. Child benefit programs, as well as social assistance programs that target groups such as single mothers with young children, expand the budgets of qualifying families. Economists note two potential mechanisms through which this increase of the family budget may improve outcomes for low-income mothers and their children. First, families may use the income to purchase more goods and services, including those goods that are valuable in maintaining child well-being and enhancing child development, such as food, clothing, educational resources, andbooks (Mayer 1997, Yeung et al. 2002). Second, indirect effects can occur, such as reduced stress and improved marital and family relations and support, increasing opportunities for employment, which may in turn benefit women and children (Currie et al. 2010, Dooley & Prause 2002, Mascaro et al. 2007).
+Conditional Cash Transfers
+Conditional cash transfer programs aim to reduce poverty by making welfare programs conditional upon a receiver’s actions. Money is transferred only to persons who meet certain criteria. In 2007, the Center for Economic Opportunity of the New York City Mayor’s Office initiated the first conditional cash transfer program in the United States, Opportunity NYC-Family Rewards (hereafter, Family Rewards), which provided assistance to 2,377 New York City families. The program was explicitly modeled after Mexico’s Oportunidades (Aber 2009). The New York City program was privately funded and operated for 3 years (2007-2010) to provide cash rewards in the areas of children’s education, preventive health care, and employment (Riccio et al. 2010). There were two main mechanisms through which it was hypothesized that Family Rewards could improve the health of low-income families. First, through health-related incentives, the program might encourage participating families to increase their use of preventive care services. Second, the increase in family income brought about by the cash transfer could increase the ability of families to invest in healthy lifestyles and reduce financial stress. The main study of Family Rewards compared the outcomes of the Family Rewards participants with those of a control group of 2,372 families. Ninety-four percent of Family Rewards participants were low-income mothers. The experiment led to improvements in health insurance coverage and in mothers’ perceptions of their own health and hope for the future, mainly through improvements in reported financial well-being. Specifically, improved financial well-being explained 32% of the gap in “hope” scores between the intervention and control groups at 42 months, while preventive care use explained 21% of the difference (Courtin et al. 2018).
+Outside of the United States, there is a larger evidence base developing on the indirect effect of cash transfers on poverty alleviation and mental health in women. Overall, this literature has demonstrated that positive “economic shocks” delivered to individuals through cash transfer programs (Haushofer & Shapiro 2016) or through economic transfers and poverty alleviation programs (Banerjee et al. 2015) yield mental health benefits in terms of reduced depressive and anxiety symptoms. In South Africa, Green et al. (2016) andFernald et al. (2008) found no effects of entrepreneurship assistance on depressive symptoms in women. In Fernald and colleagues’ (2008) study, a subgroup analysis suggested that credit access decreased depressive symptoms only among men. Green et al. (2016) hypothesized that the gains women derived from increased economic security were offset by stressors associated with planning, launching, and maintaining a new business.
+Haushofer & Shapiro (2016) randomized Kenyan villages and households to receive unconditional cash transfers (programs that aim to reduce poverty by providing cash without any
+conditions upon the receivers’ actions) of $400 or $1,500 compared with a group that received no money; results showed that the cash transfers had a positive impact on self-reported distress and depressive symptoms among adults. Recipients of the largest transfers also exhibited reduced levels of the stress hormone cortisol. Similarly, Ozer et al. (2011) compared Mexican women who participated in Oportunidades, a government-sponsored conditional cash transfer program, with a matched sample of women not exposed to the program and found that women in the treatment group had lower self-reported depressive symptom scores. The authors also presented evidence that this quasi-experimental effect was mediated by reductions in perceived stress and increases in perceived control.
+Summary of Findings on Interventions Based on Social Causation Hypothesis
+In summary, there is some support for the social causation hypothesis (Gibson et al. 2018, Moore et al. 2017). Longitudinal studies that have used repeated measures designs have found that increases in income from earnings and wages or other mechanisms have led to reductions in depressive symptoms. Yet, results from studies incorporating additional methods are mixed, especially because employment and increased hours and employment-related demands can also increase role strain for pregnant and postpartum women.
+Cash welfare receipt appears to have neither positive nor negative effects on mental well-being for mothers. Other features of employment and social services policy appear to be important. For example, welfare policies that require quick job entry for mothers with young children increase depressive symptoms, perhaps because of child care and other barriers and low-quality jobs. However, at the causal level, the evidence supporting the power of economic factors to change depressive symptoms is still sparse.
+Earnings, employment, and tax and cash transfer policies need further examination to understand their effects on maternal mental health and which subgroups of mothers they are most likely to affect. In other words, what are the mechanisms (mediators and/or moderators) by which economic factors benefit maternal mental health? Some research examining the social causation hypothesis has been conducted globally. In this work, the impact of economic interventions on mental health symptoms and subsequent improvements in economic and social mobility has been examined, and cash transfer programs have been found to reduce depressive symptoms in mothers compared with controls (Samuels & Stavropoulou 2016).
+INTERVENTIONS BASED ON THE SOCIAL SELECTION HYPOTHESIS
+Tests of the social selection hypothesis can be found in studies that examine the impacts of the provision of depression treatment to low-income pregnant and parenting women and examine employment outcomes. While fewer studies exist that have examined economic outcomes of depression treatment for women, a few studies (Booshehri et al. 2018, Brenninkmeijer et al. 2019, Lagerveld et al. 2012) have indicated economic improvement after mental health treatment and provided support for the social selection hypothesis. Yet most of the studies have found that the treatment of depression alone does not substantially improve labor force participation (Bee et al. 2010, Brenninkmeijer et al. 2019, Hollinghurst et al. 2010, Nieuwenhuijsen et al. 2014).
+In one longitudinal study (Simon et al. 2001), persons with major depression were randomly assigned to receive one of three different antidepressants, and improvements in depressive symptoms were significantly related to increased employment after 1 year of treatment. Specifically, those in remission and without depressive symptoms had a higher probability of paid employment
+and missed 10 fewer days of work compared with those with persistent depressive symptoms that met criteria for major depressive disorder. Unfortunately, the data are not available by gender.
+Outside of the United States, Bass et al. (2016) conducted a randomized controlled trial of a group-based economic intervention in the South Kivu province of eastern Democratic Republic of Congo. Bass et al. (2016) investigated the impact ofvillage savings and loans associations on economic, social, and psychological outcomes among female sexual violence survivors (all mothers) with elevated mental health symptoms and impaired functioning. While female sexual violence survivors with mental health symptoms were successfully integrated into a community-based economic program, the immediate program impact was seen only for increased food consumption and reduced experience of stigma. Impacts on depression severity were not realized. Bass and colleagues (2016) suggested that targeted mental health interventions may be needed to improve psychological well-being among women.
+POLICY IMPLICATIONS
+How we conceptualize the association between poverty and depression in pregnant and parenting women has important implications for policy. In reviewing the evidence, we have focused on studies in the United States. Most public policies for low-income mothers have included job and skill training or work requirements. Sometimes policies will also include economic incentives and work supports such as child care and transportation, but few have focused centrally on treatment for depression as a way of helping women to escape poverty.
+One policy response to the social selection hypothesis is to increase access to and coverage of depression treatment (Green et al. 2016) or paid family leave (Ybarra & Noyes 2019), and the available evidence, although still quite limited, suggests that such interventions can improve economic outcomes (Lund et al. 2010). From this perspective, high levels of depressive symptoms among low-income pregnant and parenting women can be framed as a large barrier to overall economic and social mobility for families (Cambron et al. 2015,Jayakody & Stauffer 2000, Miranda & Patel 2005, Radey et al. 2020, Thornicroft & Patel 2014). The addition of employment-directed interventions to clinical interventions for depression has been shown to improve occupational outcomes and depressive symptoms but has not been widely investigated among low-income women (Lagerveld et al. 2012, Nieuwenhuijsen et al. 2014).
+Studies testing treatment for depression offer some evidence that mental health assistance and treatment have positive effects on employment success and earnings, but these studies are not specific to the population of low-income pregnant and parenting women. The findings may apply, but we need additional research to determine their generalizability. For the populations sampled, it appears that reductions in depressive symptoms allow for a greater likelihood of obtaining and sustaining employment, yet for reasons we have outlined we may expect different intervention effects for women.
+While a strong argument can be made for increasing access to and coverage of mental health treatment, the severe shortage in human resources (Patel et al. 2018) makes universal access difficult. For this reason, some have argued (Brownell et al. 2016, Forget 2013, Shaefer et al. 2018, Van Parijs 2004) that it would be beneficial to introduce broad-based poverty alleviation programs such as universal income policies as these could have a positive impact on the mental health of low-income pregnant and parenting women. In other words, these universal, broad-based poverty alleviation programs represent a pathway for indirect effects on maternal mental health. This strategy is based on the social causation hypothesis that poverty leads to mental ill health and thus suggests that investments in poverty alleviation programs can be framed as indirect methods of improving maternal mental health outcomes (Courtin et al. 2020, Topitzes et al. 2019).
+FUTURE POLICY DIRECTIONS
+Overall, these findings support the need for policies that integrate welfare and employment with mental health services for low-income pregnant and parenting women and thus suggest a bidirectional and interactive relationship between income and depressive symptoms for mothers. Yet the two systems of (a) welfare and employment policy and (b) mental health services and health care policy typically operate within different agencies and departments with very little overlap in programs and regulations. The evidence reported here supports the idea that depression can adversely affect employment and income and that improvement in depression positively affects economic opportunities. Specifically, this research suggests that one pathway to employment and higher incomes for low-income pregnant and parenting women is better mental health. Therefore, widely available assistance for mental health, especially for low-income pregnant and parenting women, could provide major contributions to programs designed to increase earnings and incomes.
+Policies designed to increase employment should acknowledge the individual characteristics and barriers faced by low-income pregnant and parenting women. Policies that increase total income for employed mothers are more likely to improve well-being than those that involve simply an exchange of welfare for work (Morris 2008). The concerns reported by working mothers about child well-being that, in turn, lead to increased depressive symptoms speak loudly to the need for public policies that create work supports (adequate child care in particular) and supports to manage the stress associated with balancing multiple demands from new roles, thus enabling mothers to balance work and parenting. Policies to address this issue include paid family leave, child care subsidies, earnings supplements, health insurance, universal access to mental health visits and support groups, and workplaces with sufficient flexibility to allow mothers to deal with family concerns and needs.
+Simply gaining employment is not a remedy that will alleviate economic or mental health burden. In fact, policies that emphasize immediate job entry for mothers with few skills lead not only to increased depressive symptoms but also to unstable employment (Morris 2008).
+MENTAL HEALTH OUTREACH FOR MOTHERS
+Limited evidence is available regarding interventions that address the complex burdens of depression and employment among low-income pregnant and parenting women (Moore et al. 2017). Yet, one such intervention, the Mental health Outreach for MotherS (MOMS) Partnership, has successfully reduced depressive symptoms among overburdened, underresourced pregnant and parenting women. Launched in New Haven, Connecticut, and now being replicated in five other states, the MOMS Partnership offers 8 weeks of cognitive behavioral therapy (CBT) for treatment of depressive symptoms. MOMS uses a model that engages mothers from the community and trains them to codeliver CBT-based interventions alongside clinicians. Importantly, this MOMS model is fully embedded in the TANF system in two states, thus demonstrating the feasibility of the innovative use of government TANF funds to broadly and simultaneously support maternal mental health and economic mobility.
+A recent pilot project in the Washington, DC, TANF system deployed the MOMS 8-week CBT program among two cohorts (n = 36) of pregnant and parenting female TANF participants. TANF staff, consisting of a social worker and a community mental health ambassador (a mother from the local community), were trained to deliver the CBT intervention to TANF participants. Participants completed baseline, midpoint, and endpoint measures to assess depressive symptoms, parenting stress, basic needs, employment, and acceptability. Fidelity of the intervention was tracked via audio recordings of sessions. Results examined from baseline to 8 weeks postintervention demonstrated the acceptability and feasibility of the MOMS approach. TANF participants
+reported being highly satisfied. Depressive symptoms and parenting stress were significantly reduced from the beginning to the end of the intervention, and mothers reported being more able to meet their family’s basic needs from the beginning to the end of the intervention. Additionally, employment (20 hours or more) increased significantly, by 30%, from the beginning to the end of the intervention. Moreover, TANF staff delivered the intervention with high fidelity (Smith et al. 2021).
+A major implication of the MOMS pilot project findings for policy is the value of integrating welfare and employment opportunities with mental health services for pregnant and parenting women. Furthermore, the findings suggest that in addition to the effects of the MOMS CBT group-based treatment and the use of a community mental health ambassador, group-based economic activities, such as those employed by MOMS, may also provide a means of affecting economic outcomes and depressive symptoms for low-income parenting women. In one study (Pronyk et al. 2008), participants reported that having an environment for social connection made them feel supported and connected; such benefits may in turn lead to improvements in maternal mental health. Traditional intervention research for people with common mental disorders living in poverty has focused on alleviating the burden of symptoms through psychotherapy and/or psychosocial programming (Lund et al. 2010). Research now needs to examine more fully the integration of economic and mental health interventions for low-income pregnant and parenting women.
+CONCLUSION
+Since our specific focus in this review is on the mental health of low-income pregnant and parenting women, most of our attention is on the links among poverty, depression, and social and economic mobility for women. It is important to see interventions in these specific areas as part of an ambitious set of policies to reduce poverty itself and to improve outcomes for low-income mothers.
+Our results make at least two important contributions to the literature on poverty, mental health, and parenting specific to women. First, we have summarized the gendered nature of the mental health and wealth relationship and the large body of work showing that poverty and mental health are in fact related in pregnant and parenting women (Lund et al. 2010). A gendered understanding of poverty and depression is crucial for exploring the differing impacts and resultant policy interventions. Women face the dual burden of widespread poverty and heightened risk for depression.
+Second, particular to low-income women who are pregnant and parenting, depressive symptoms have been shown to have negative impacts on the successful transition from welfare to work. Applicants to programs that aid the poor, such as the TANF program, experience depressive symptoms at much higher rates than do members of the general public. Yet the traditional focus in studies of barriers to employment among women has been on structural barriers, including access to child care and transportation, and assessment of individual barriers has often been limited to demographic factors, such as lack of schooling and work experience or physical health limitations. However, depression appears to be an important barrier to economic mobility, and the existence of depressive symptoms may prevent women from leaving welfare for work and remaining employed.
+The mechanisms that might explain this link between poverty and depressive symptoms, however, remain uncertain for low-income pregnant and parenting women (Burns 2015). Additional evidence is greatly needed to guide policy makers. For example, it is likely that raising income levels will affect depressive symptoms in women differently than in men. Women who have the personal resources to access networks of support will be disproportionately helped by raising their income, whereas women who lack access to social capital and supports are likely to be less affected
+by changes in their income and will require additional support to change employment trajectories, parenting practices, and, in turn, child outcomes.
+Research on the relationship between poverty, stress, and parenting must take into account the reciprocal relationships and interdependence between parents and children who are facing adversity together and the particular role of gender in intergenerational poverty and mental health (Gugushvili et al. 2019). Future research needs to focus not only on the overall effects on women but also on the interactive effects of poverty alleviation policies and programs on women and children together.
+Poverty is only one of a number of factors that affect parenting, so it should not be assumed that changes in income (especially minimal changes, such as those that typically result from government initiatives) will necessarily reduce low-income pregnant and parenting women’s depressive symptoms sufficiently to change parenting style (Belsky & Vondra 1989). For example, Fram’s (2003) study of mothers receiving welfare payments in the United States found that social support acted as a buffer against the effect of mothers’ stress and disciplinary practices characteristic of parenting style. Fram (2003) found that when a combination of factors supporting resiliency— more education, more earnings, and better neighborhoods—came together, parenting practices and child outcomes tended to be better. Despite the strong body of research linking poverty to poor child outcomes, there is equally good evidence to show that mothers living in poverty possess strong coping skills in the face of adversity.
+Because psychological factors play a critical role in the success of economic policy efforts, efforts to address depression in parenting women are a potentially important function of welfare and social services receipt that have often been overlooked. In this article, we have reported on findings from a pilot study to embed high-quality depression treatment for mothers into the Washington, DC, TANF system. The MOMS findings suggest that one pathway to employment and higher incomes for low-income pregnant and parenting women is better mental health. Therefore, widely available assessments and interventions for depression and other psychological distress through the TANF system could prove to be a scalable method to help mothers increase earnings and employment and reduce the cycle of intergenerational poverty.

Are national suicide prevention programs effective.txt

@@ -0,0 +1,163 @@
+Background
+Epidemiology
+The World Health Organisation (WHO) estimates that about a million people die by suicide every year, representing a “global” mortality rate of 16 per 100,000 or one death every 40 s making it the tenth leading cause of death worldwide [1]. Suicide rates in many developing countries have been steadily rising, and the overall worldwide suicide rate has increased during the last 50 years [2].
+Suicide and non-fatal suicidal behavior are significant public health issues worldwide requiring effective preventive interventions. However, there is still a need to identify what prevention strategies should be prioritized to achieve the biggest impact on a reduction of suicide attempts and suicide deaths [3].
+Suicidality is a problem caused by multiple factors, making it difficult to treat by individual medical, psychological, educational, social or political methods.
+Thus national suicide prevention programs (NSPP) were initiated in the 1990s aiming to take a holistic approach to combat suicide. There are 28 countries known to have national strategies for suicide prevention. Prevention programs are designed to identify vulnerable groups, improve the assessment and care of people with suicidal behavior, and improve surveillance and research. They also aim to raise awareness by improving public education. NSPPs attempt counter the stigma toward people exhibiting suicidal behavior and those who suffer from mental disorders. Institutions such as the World Health Organization (WHO) and the International Association for Suicide Prevention (IASP) have developed common guidelines and the following recommendations to set up suicide prevention strategies including NSPP [2, 4, 5]:
+1.	Preventive measures should address suicide and suicide attempts. The loss of human resources, socioeconomic burden, and costs for the healthcare of these individuals are considerable.
+2.	The support and rehabilitation of persons at risk can prevent some suicides. A holistic approach is necessary.
+3.	National governments are responsible for developing strategies to provide financial and technical support that involve society as a whole.
+4.	Measurable objectives and systematic studies must be forthcoming.
+National suicide prevention programs in Norway, Sweden, Finland and Australia
+Norway published the first national suicide prevention strategy in 1995, one that has been revised and updated several times. Aspects of the second and third prevention strategies are the focus of their program. Approaches to enhance the mindfulness of politicians, governmental departments and the general population were taken for this purpose. Medical and social welfare programs were optimized, and aftercare improved. An external board was assigned to evaluate individual projects and the entire program [5, 6]. Their findings were summarized in a publication by Soras in 2000 [7]. A follow-up project to the national plan “Measures against suicide 2000-2002” was evaluated and published by Mehlum and Reinholdt in 2001 [8].
+The aim of Sweden’s national suicide prevention program established in 1995 was a consistent drop in the number of suicides and suicide attempts, to reduce the factors encouraging suicidal behavior in children and youths, as well as the early detection of suicidal tendencies in endangered groups. They aspired to increase the level of awareness in the general population in the first, second and third prevention strategies [9].
+From 1986 to 1991, Finland enacted a research program on suicide and developed preventive strategies. One major goal was to engage high-risk groups. A suicide prevention program was implemented in 1992 as the first governmental program with activities involving all levels of prevention strategies [10]. Moreover, an external board was assigned to evaluate and improve this program [11].
+A national youth-suicide prevention program existed from 1995 to 1999 in Australia. This initiative then evolved into a national suicide prevention program involving first, second, and third prevention strategies. It was one of this program’s aims to observe a lower suicide rate and reduction in suicidal thinking and behavior. The program also intended a better psychological strain and mental health [12]. Following implementation of the original National Youth Suicide Prevention Strategy (NYSPS) in 1995, they did in fact observe a substantial decline in suicide in young men. A study by Page et al. [13] reported a minor discernible impact on suicide rates in those areas that had participated in local
+targeted suicide prevention activities in the period following the NYSPS.
+One of our main problems is how to evaluate suicide prevention programs. As Kerkhof and Clark [14] stated in their editorial, there are obvious limitations in studying effectiveness, e.g., there are no experimental designs that might be applied. So far, little research has been done investigating the effect of national suicide prevention programs, whereas there have been studies about local interventions programs to prevent suicide and suicide attempts; e.g. [15].
+The aim of the present study was to analyze the effectiveness of national suicide prevention programs taking a statistical approach involving the segmented regression analysis of interrupted time series data. We posed the following questions:
+1.	Does the implementation of a national suicide prevention program lead to a significant reduction in suicide rates?
+2.	Are there gender-related differences?
+3.	Are there age-related differences?
+Methods
+One of the major difficulties is verifying the success of NSPP, which we considered as a decrease in suicide attempt or suicide death rates within the population of a country. It is extremely hard to tell which parameters are actually responsible for success among the specific suicide-decreasing effects, spontaneous changes in long-term development, social changes and data inaccessibility due to privacy protection laws, as well as obtaining comparable data. All these factors make it difficult to prove the exact cause of a measured change, making it important therefore to clearly define the structure and approach of the analysis. For the purpose of a conscientious decision-making process three experienced psychiatrists/suicidologists (WF, HT, UL), as well as a statistician (CS), reviewed the existing literature as well as different statistical approaches. Within several face to face meetings, the described approach was defined, and the criteria for the selection of these programs were agreed.
+Criteria for selection of verum and control countries
+The verum countries
+Our main criterion for selecting the verum countries was the existence of a comprehensive national suicide prevention program for at least five years.
+Comprehensive statistical analysis should be available. Nations that have an NSPP are Australia, New Zealand, Finland, Norway, and Sweden. As the New Zealanders implemented their program in 1998 only for young people, we had to exclude them from our statistical
+analyses. The Netherlands, Great Britain, the USA, France, and Estonia have also implemented programs; they failed, however, to meet our inclusion criteria. We ultimately selected Finland, Norway, Sweden, and Australia for the verum group. All these countries have published their programs’ results comprehensively [6-13].
+Control countries
+We compared the verum countries with control countries selected according to the criteria below. They should not have an NSPP and should not differ in the following aspects (at the time of the study):
+1.	Culture and religion: e.g., suicide rates in Mediterranean countries are lower than in Northern European countries. Other examples are countries with a mostly Muslim population, which may be caused by the strong taboo about this problem in the society.
+2.	Historical and political factors: studies have shown that substantial historical developments such as political changes have influenced the dynamics of suicide statistics. Thus Eastern European countries, as well as Germany, could not be included.
+3.	Socio-economic structure: we decided that the control countries should have a western democracy with distinctive market-based economies similar to the verum countries.
+4.	Population size: Statistics from countries with small populations were not included because lower numbers of suicides, minimal increases or decreases can skew the statistics.
+5.	Quality of published statistics: data had to be well founded, reliable and accessible, and should have been collected during the same time period.
+Countries whose statistics were erratically collected were excluded (i.e., African states, China).
+Finally, Canada, Austria, Switzerland and Denmark were selected as control countries.
+Criteria for selecting the time period
+The verum countries implemented their programs in the 1990s. Long periods of observation were planned due to annual variability, which exerts strong effects, especially in countries with a smaller population. The second reason is the possibility that prevention programs gradually lead to success.
+For this study, we had to assess the suicide rates prior to the NSPP to detect any differences. For the verum countries, time 0 (T0) was defined as the year the NSPP was implemented. Countries differed in this respect, thus T0 for Finland is 1992, and 1994 for Norway. T0 is the year 1995 for the control countries. We decided on
+six years (T0 to T + 5) after the implementation of the NSPP to qualify as the period of analysis. All countries had to be statistically represented at the beginning of the analysis. For that reason, T-22 was established retrospectively as the starting point.
+Data were collected from World Health Organization statistics on suicides and demographics separated by age and sex [16].
+Statistics
+To estimate the impact of the NSPP on suicide rates in verum and control countries we applied a segmented regression analysis of interrupted time series data [17-21]. This method estimates changes in levels and trends controlling for baseline levels and trends, which is one of its major strengths. The observation period is divided into pre- and post-intervention segments for which separate intercepts and slopes are estimated. A linear relationship between time and outcome is assumed, and a least squares regression line is fitted to each segment of the independent variable. To take into account the autocorrelation among observations, we estimated the effect of the intervention using the ARIMA model (autoregressive integrated moving average) and tested for autocorrelation of the error terms via the Ljung-Box-test.
+The time series regression equation for our analysis is:
+Yt = P0 + P1 * time + P2 * phase + P3 * time_after_NSPP + et
+Yt is the outcome variable, in our model this is the number of suicides per 100.000 in year t, “time” is the number of years at time t from the start of the observation period starting with 1 at time point “t-22”; “phase” is an indicator variable, which is 0 for the time points before and 1 for the time points after the NSPP introduction, “time after NSPP” is how many years after NSPP introduction which is set to 0 for the years before the NSPP introduction and taking on the values of 1 to 5 for the years after NSPP introduction and et represents random variability at time t not explained by the model.
+The coefficient p0 estimates the baseline level of suicides per 100.000, p1 estimates the baseline trend before NSPP introduction, which is the change in the mean number of suicides per 100.000 occurring each year before the implementation. The coefficient p2 estimates the change in level in the mean yearly number of suicides per 100.000 immediately after the NSPP implementation and P3 estimates the change in trend in the mean number of suicides per 100.000 after its implementation.
+All analyses were conducted separately for men and women and split into four age groups each (< 24 years,
+25-44 years, 45-64 years and > 65 years). Furthermore, we analyzed the difference in suicide rates between verum- and control countries, again separately for men and women and the different age groups to estimate how the change in suicide rate in the verum countries differed from the change in the control countries. For all analyses, we used SPSS for Windows version 23. A significance level of <0.05 was considered significant.
+Results
+Analysis of verum countries - males
+Table 1 and Fig. 1 show the parameter estimates from the linear segmented regression model for all males in the verum countries. Right before the beginning of the observation period, an average of 23 per 100.000 males in the verum countries committed suicide per year. Before implementation of the NSPP, there was a significant year-to-year change in the mean number of suicides (p < 0.001). The immediate change in the number of suicides directly after the intervention was not significant (p = 0.536). However, the level change becomes significant two years after the intervention and remained significant for the following three years. The year-to-year trend in the mean number of suicides per 100.000 after the intervention changed significantly (p = 0.006).
+Observing the different age groups, we detected significant level changes in the group of males under age 24 after 5 years (p = 0.049) of NSPP Within the 25-to-44-year-olds, we noted significant level changes after 1 (p = 0.041), 2 (p = 0.001), 3 (p < 0.001), 4 (p < 0.001) and 5 (p < 0.001) years after the NSPP implementation. The trend change that occurred after the NSPP was implemented also reached significance (p = 0.014) in this age group.
+The group of 45-64-year-old males revealed significant level changes after 2 (p = 0.010), 3 (p = 0.001), 4 (p = 0.001) and 5 (p = 0.001) years of NSPP.
+Males older than 65 years showed significant changes in suicide rates after 3 (p = 0.011), 4 (p = 0.005) and 5 (p = 0.007) years of NSPP.
+We observed a significant baseline trend in all age groups except that of the males older than 65. These trends were positive for the males younger than 24 years and the group of 25-44-year-olds, which evolved into a negative trend after the NSPP implementation in both groups. However, this trend change only attained significance for the group of the 25-to-45-year-olds (p = 0.014). The trend for the group of the 45-64-year-old males was already slightly negative before the NSPP and was strengthened by it, an improvement that did not reach significance (p = 0.155).
+Analysis of verum countries - females
+Right before the beginning of the observation period, an average 8 of 100.000 women committed suicide per year in the verum countries. The baseline trend before implementation of the NSPP indicates that the suicide rates remained constant over the years up to the year of NSPP introduction. However, two years after its implementation our analysis showed a statistically significant level change in the suicide rate (p = 0.018). The same applies for years 3, 4 and 5 after implementation (Table 2, Fig. 2). The year-to-year trend in the mean number of suicides per 100.000 after implementation did not change significantly (p = 0.333).
+Assessing the different age groups, we identified significant level changes in the group of females aged between 45 and 64 years after 3 (p <0.021), 4 (p <0.011) and 5 (p < 0.012) years after the NSPP implementation.
+Females older than 65 years showed significant level changes immediately after NSPP implementation (p = 0.002) and after 1 (p < 0.001), 2 (p < 0.001), 3 (p < 0.001), 4 (p < 0.001) and 5 (p < 0.001) years of NSPP These two female groups’ baseline trends were significant. The trend was already negative before the NSPP in the group of 45-to-64-year-old females, while the trend for the females older than 65 was constant. The trend became negative in both groups, but the changes did not reach significance.
+Analysis of control countries - males
+The average number of suicides per 100.000 for all males in the control countries right at the beginning of the observation period was 28 per year. As we expected, there were no level changes or a significant trend after the period of NSPP implementation in the verum countries. However, although the trend change was not significant, it became clearly negative (Table 3, Fig. 3).
+The analysis of males according to age group showed no significant trend or level changes in either males <24 years and those aged 25-to-44 years. We noted a significant (p = 0.010) baseline trend of - 0.56 within the 45-to-64 age group, which changed by -1.1 after the time of NSPP implementation in the verum countries. However, this change was not significant (p = 0.303). Interestingly, males older than 65 years showed significant level changes after 2 (p = 0.040),
+3 (p = 0.005), 4 (p = 0.002) and 5 (p = 0.002) years. We also detected a significant trend change of - 2.6 (p = 0.045).
+Analysis of control countries - females
+The analysis of all females in the control countries revealed no significant level or trend changes (Table 4, Fig. 4). The average number of suicides per 100.000 at the beginning of the observation was 13 per year.
+The analysis of females according to age group showed no changes in trend or level for females < 24 years nor in females in the group of 25-to-44 year-olds. We observed a significant (p = 0.001) baseline trend of - 0.4 in the 45-to-64-year-old group. Similar to males, females older than 65 years showed significant level changes in suicide rates after 2 (p = 0.020), 3 (p = 0.012), 4 (p = 0.015) and 5 (p = 0.025) years.
+Comparison between verum and control countries
+To compare verum and control countries we calculated the difference in the two suicide rates for every year (i.e., the rate of the verum countries minus the rate of the control countries), thus we could analyze both rates in one ARIMA model. Taking the difference collapses the two time series into one and estimates a difference-in-differences effect enables us to make a statement about how the change in the verum countries differed from that in the control countries.
+One would expect a statistically significant negative level change a few years after implementation of the NSPP (e.g., the suicide rate in the verum countries would be expected to drop while that in the control countries would remain constant). However, we detected no significant level or trend change regarding the overall rates for all demographic groups including males or
+females and made the same observation when analyzing the males and females divided into different age groups.
+However, the difference in suicide rates right at the beginning of the observation period was significant for all male and female groups except the males <24, males aged 25-to-44 years, and females <24 (e.g., all males: difference of - 5.6, p = 0.018; all females: difference of - 5.2, p < 0.001). As mentioned above, segmented regression analysis controls for baseline level and trend.
+Discussion
+Overall, this study demonstrates that National Suicide Prevention Programs are effective, but this effect seems to correlate with age and sex.
+Segmented regression analyses of interrupted time series data have shown a statistical significant decline in suicide rates in the verum countries in males, with the
+strongest effects in groups aged 25-to-44 years and 45-to-64 years. We noted a significant effect in females aged 45-to-64 and > 65 years, although this effect was not as strong as it had been in males. We did not detect this effect in the control countries (except in those > 65 years of age). After analyzing the differences in suicide rates between verum and control countries, no significant level changes or trend changes appeared.
+Several working groups have investigated various suicide prevention strategies or programs.
+Major efforts have focused on the accessibility of suicide means. There is strong evidence that restricting the availability of methods (e.g., firearms) can reduce suicides [22-24]. Men are more likely to use guns as suicide method. That might partly explain the significant effects observed in males age 25-64 years. Another example is the detoxification of the English gas in the 60s
+which lead to clearly reduced suicide rates [25]. Similar results could be found in Saxony (Germany, “coal gas story”). A 74% reduction in suicide rates were shown due to the detoxification of the city gas [26].
+The establishment of suicide prevention centers like the “Samaritans”, “Befrienders International” or “Lifeline” caused a perceptible but nevertheless minor preventive effect [27, 28].
+Further approaches like “Tele-Help” or “Tele-Check” were associated with lower suicide numbers [29].
+Others have examined the influence of medication on suicidal behavior. Lithium, a mood stabilizer, is well established as a drug that reduces suicides [30].
+Advanced training for general practitioners was implemented in the 1980s by the Swedish government. Since general physicians became better able to detect depression than beforehand, suicide rates dropped considerably [31].
+The interpretation of statistical data and the causal combination with events or the course of suicide statistics give rise to a complex challenge. Multifarious, unforeseeable factors can play an important role in the appearance of suicidal behavior. Thus the genuine situation in different nations can only be compared under certain limitations.
+There are relatively few studies investigating the effectiveness of suicide prevention programs, and those reveal inconsistent outcomes [32-34]. Countries such as Finland and Scotland have reported a significant reduction in suicide rates [35], whereas others (e.g., Norway, Sweden or Australia) reported limited effects in certain subgroups.
+Our study results endorse the overall effectiveness of National Suicide Prevention Programs. A major reduction in suicide rates, especially in males over 25 years, is presumably related to all arrangements regarding preventing strategies of these programs rather than to one single strategy. There are a couple of hypotheses as to why we found no statistical differences when comparing verum and control countries:
+About 800.000 suicides occurred worldwide representing an annual age-standardized suicide rate of 11.4 per 100,000 population. We know that suicide rates are higher in males (15.0/100000) than in females (8.0/100000). It is acknowledged that three times as many men died by suicide as women; another possible explanation that this study could only reveal differences within the group of men.
+Suicide rates are highest in both males and females aged over 70 years. But several countries have different
+statistical patterns in their age related suicide rates. As the WHO report stated in some countries there is a peak in suicide rates in young adults that subsides in middle age and in other regions suicide rates increase steadily with age [2]. One could argue that our findings in age group 25-64 are partly related to such different patterns.
+Prevention programs aiming to help special age groups may play an important role. Within this study’s framework, we were not in a position to analyze other factors associated with changing suicide rates, such as access to and availability of health care providers. Furthermore, the observation period after NSPP implementation was quite short (five years). Certain strategies might well need longer to reveal their effectiveness.
+Despite the effort to decrease suicide rates via different approaches also the economic effects are remarkable. Vasiliadis et al. recently showed that suicide prevention programs such as the European Nuremberg Alliance against Depression (NAD) are cost-effective and may result in significant potential cost-savings due to averted suicide deaths and fewer life years lost [36].
+It is extremely challenging to investigate changes in implemented prevention strategies such as suicide rates within different countries. Matsubayashi and Ueda (2011) investigated the effect of national suicide prevention programs on suicide rates in 21 OECD nations [37]. Overall, they found that suicide rates decreased after the government initiated a nationwide suicide prevention program, as we did in this study; more so in men than in women. Remarkably, they detected the strongest
+effects in youth (< 24 years old) and the elderly (> 65 years old). They also noted a limited effect on the working-age population. They discuss those differences as a result of specific goals within the prevention programs, such as reducing the access to firearms. One could argue that a comparison of 21 countries may be too ambiguous, as major cultural, religious, socio-economic and political differences can play an important role. That is why we carefully selected countries that were fairly similar in those specific areas - a clear strength of this study.
+A very recent narrative analysis conducted by Zalsmann et al. [38] investigated the effectiveness of different suicide prevention strategies. Due to the heterogeneity of populations and methodology, formal meta-analyses could not be applied. They investigated different suicide prevention methods including school-based awareness program that reduced suicide attempts. They concluded that no one strategy is clearly superior to the others. Our results also support the idea that different approaches appear effective in different groups according to age and gender, for example. That might be another reason for the results found in this study.
+Several limitations of this study provide guidance for future research:
+•	Extensive programs have not been running long enough.
+•	The present study covered just four control and four verum countries, meaning that our results cannot be extrapolated to other countries.
+•	The length of our observational period after NSPP implementation is relatively short - later influences could not be excluded.
+•	Our study approach did not enable us to investigate whether specific components of an NSPP exert different influences on suicide rates.
+•	Our data did not provide information on whether other activities not implemented in a national strategy such as general welfare programs may also influence suicide rates. According to the WHO report [2], current data show a decrease in suicide rates in different countries even in those without an NSPP, which makes our findings not generalizable.
+Despite the encouraging drop in suicide rates, it is very important that future evaluations of suicide prevention programs include the number of suicide prevention interventions implemented successfully as well as the number of hospitalized suicide attempts. The systematic collection of specific data (including suicides and suicide attempts) is key. There are many countries that collect no such data at all or only very minimal data.
+Conclusion
+To the best of our knowledge, this is the first study investigating the effectiveness of national suicide prevention programs applying segmented regression analysis of interrupted time series.
+Our study implies that the implementation of a national strategy is an effective tool to reduce suicide rates. Special attention should be drawn to different approaches regarding age groups as well concerning females. Future research should investigate longer time periods and different aspects of prevention programs and what other factors may influence suicide rates.
+As stated in the WHO’s framework “Public Health Action for the Prevention of Suicide” [2], it is “imperative that governments - through their health, social and other relevant sectors - invest human and financial resources in suicide prevention.”
+Lewitzka et al. BMC Psychiatry (2019) 19:158
+Page 10 of 10
+Publisher's Note
+Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
+Received: 1 March 2018 Accepted: 14 May 2019
+Published online: 23 May 2019
+References
+1.	Turecki G, Brent DA. Suicide and suicidal behaviour. Lancet. 2016;387:1227-39.
+2.	WHO. Preventing suicide: A global imperative http://www.who.int/mental_ health/suicide-prevention/world_report_2014/en/. Accessed 20 May 2019.
+3.	Krysinska K, Batterham PJ, Tye M, Shand F, Calear AL, Cockayne N, Christensen H. Best strategies for reducing the suicide rate in Australia. Aust N Z J Psychiatry. 2016;50:115-8.
+4.	WHO 2012 http://apps.who.int/iris/bitstream/10665/75166Z1/ 9789241503570_eng.pdf. Accessed 20 May 2019.
+5.	IASP 2015 https://www.iasp.info/suicide_guidelines.php. Accessed 20 May 2019.
+6.	Norwegian Board of Health (1995). The National Plan for suicide prevention 1994-1998. Oslo https://www.med.uio.no/klinmed/english/research/centres/ nssf/articles/prevention/The_national_plan_for_suicide_prevention_1994-1998.pdf. Accessed 20 May 2019.
+7.	Sorâs I. Handlingsplan mot selvmord - hva viser evalueringen? Suicidologi. 2000;5:12-3.
+8.	Mehlum L, Reinholdt NP. Handlingsplan mot selvmord: Gode erfaringer skal fores videre. Suicidologi. 2001;6:16-8.
+9.	The Swedish National Council for Suicide Prevention. Support in suicidal crises: the Swedish National Program to develop suicide prevention. Crisis. 1997;18:65-72.
+10.	Jenkins R, Singh B. National Suicide Prevention Strategies. Psychiatr Fenn. 1999;30:9-30.
+11.	Beskow J, Kerkhof A, Kokkola A, Uutela A. Suicide prevention in Finland 1986-1996. External evaluation by an international peer group. Psychiatr Fenn. 1999;30:31-46.
+12.	Commonwealth of Australia. L.I.F.E. Living is for everyone. A framework for prevention of suicide and self-harm in Australia, areas for action.
+Department of Health and Aged Care 2000, Canberra. Available via https:// www.lifeinmindaustralia.com.au/about-us/the-life-framework. Accessed 20 May 2019.
+13.	Page A, Tylor R, Gunnell D, Carter G, Morrell S, Martin G. Effectiveness of Australian youth suicide prevention initiatives. Br J Psychiatry. 2011;199:423-9.
+14.	Kerhof JF, Clark DC. How to evaluate national suicide programs? Crisis. 1998;19:2-3.
+15.	Ono Y, Sakai A, Otsuka K, et al. Effectiveness of a multimodal community intervention program to prevent suicide and suicide attempts. A quasi-experimental study. PLoS One. 2013;(8):e74902.
+16.	http://www.who.int/whosis/en/, access at the time of data analysis: July 2005.
+17.	Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27:299-309.
+18.	Ma ZQ, Kuller LH, Fisher MA, Ostroff SM. Use of interrupted time-series method to evaluate the impact of cigarette excise tax increases in Pennsylvania, 2000-2009. Prev Chronic Dis. 2013;10:e169.
+19.	Gebski V, Ellingson K, Edwards J, Jernigan J, Kleinbaum D. Modelling interrupted time series to evaluate prevention and control of infection in healthcare. Epidemiol Infect. 2012;140:2131-41.
+20.	Penfold RB, Zhang F. Use of interrupted time series analysis in evaluating health care quality improvements. Acad Pediatr. 2013;13:S38-44.
+21.	Taljaard M, McKenzie JE, Ramsay CR, Grimshaw JM. The use of segmented regression in analysing interrupted time series studies: an example in pre-hospital ambulance care. Implement Sci. 2014;19(9):77.
+22.	Bronisch T. Der Suizid: Ursachen Warnsignale Prevention. C.H.Beck;1995, München.
+23.	Anestis MD, Anestis JC. Suicide rates and state Laws regulation access and exposure to handguns. Am J Public Health. 2015;105:2049-58.
+24.	Reisch T, Steffen T, Habenstein A, Tschacher W. Change in suicide rates in Switzerland before and after firearm restriction resulting from the 2003 “Army XXI” reform. Am J Psychiatry. 2013;170:977-84.
+25.	Kreitman N. The coal gas story. United Kingdom suicide rates, 1960-71.
+Br J Prev Soc Med. 1976;30:86-93.
+26.	Felber W. Die Entgiftung des Stadtgases und die Suizidrate in Sachsen. Symposium S-138: Die Reduktion von Suizidraten durch Beeinflussung der Suizidmethoden. Berlin: DGPPN-Kongress. p. 21-24.11.2007.
+27.	Leenaars A, Lester D. Impact of suicide prevention centers on suicide in Canada. Crisis. 1995;16:39.
+28.	Lester D. Evaluating the effectiveness of the Samaritans in England and Wales. Int J Health Sci. 1994;5:73-4.
+29.	DeLeo D, Carollo G, Dello Buono M. Lower suicide rates associated with a tele-help/tele-check servie for the elderly at home. Am J Psychiatry. 1995; 152:632-4.
+30.	Lewitzka U, Severus E, Bauer R, Ritter P, Müller-Oerlinghausen B, Bauer M. The suicide prevention effect of lithium: more than 20 years of evidence- a narrative review. Int J Bipolar Disord. 2015;3:32.
+31.	Rutz W, von Knorring L, WalinderJ. Long-term effects of an educational program for general practitioners given by the Swedish committee for the prevention and treatment of depression. Acta Psychiatr Scand. 1992;85:83-8.
+32.	De Leo D, Evans RW, editors. International suicide rates and prevention strategies. Gottingen: Hogrefe&Huber;2004.
+33.	Wassermann D. Evaluating suicide prevention: various approaches needed. World Psychiatry. 2004;3:153-4.
+34.	Wassermann GA, McReynolds LS, Musabegovic H, Whited AL, Keating JM, Huo Y. Evaluating project connect: improving juvenile probationer's mental health and substance use service access. Adm Police Ment Health. 2009;36:393-406.
+35.	Kerkhof AJ. The Finnish national suicide prevention program evaluated. Crisis. 1999;20:50,63.
+36.	Vasiliadis HM, Lesage A, Latimer E, Sequin M. Implementing suicide prevention programs: costs and potential life years saved in Canada. J Ment Health Policy Econ. 2015;18:147-55.
+37.	Matsubayashi T, Ueda M. The effect of national suicide prevention progams on suicide rates in 21 OECD nations. Soc Sci Med. 2011;73:1395-400.
+38.	Zalsman G, Hawton K, Wasserman D, et al. Suicide prevention strategies revisted: 10-year systematic review. Lancet Psychiatry. 2016;3:646-59.
+Ready to submit your research? Choose BMC and benefit from:
+•	fast, convenient online submission
+•	thorough peer review by experienced researchers in your field
+•	rapid publication on acceptance
+•	support for research data, including large and complex data types
+•	gold Open Access which fosters wider collaboration and increased citations
+•	maximum visibility for your research: over 100M website views per year
+At BMC, research is always in progress.
+Learn more biomedcentral.com/submissions
+kBMC

Association between socioeconomic status and the development of mental and physical health conditions in adulthood.txt

@@ -0,0 +1,119 @@
+Introduction
+Socioeconomic status, which captures social circumstances across the life course, is a powerful predictor of ill health. Studies have found increased morbidity and disability in individuals who are socioeconomically disadvantaged1-3 and the disease burden in this group is increasing with population ageing.4-6 However, to our knowledge, a comprehensive overview of the associations between socioeconomically patterned mental and physical health conditions is lacking.
+Previous investigations have explored the relationship between socioeconomic status and multimorbidity.3,7-14 These findings showed that having two or more diseases and developing multimorbidity was more common in people with low socioeconomic status. A limitation of these studies is the relatively restricted range of
+morbidities investigated (communicable diseases are typically not included) and a failure to capture the temporal sequence between specific diseases. Considering temporality in disease onset could yield new insights into the cascades of health conditions that characterise morbidity in people with socioeconomic disadvantage.
+To address these limitations, we examined the development of mental and physical health conditions among individuals with low and high socioeconomic status to determine temporal sequence and inter-relationships in the emergence of socioeconomically patterned conditions. We used a range of disease endpoints, adopting a data-driven approach, as we were not aware of any previous evidence-based test of hypotheses on disease cascades that characterised morbidity in people from different socioeconomic backgrounds.
+Research in context
+Evidence before this study
+Low socioeconomic status, which captures multiple aspects of disadvantage, is a known risk factor for several diseases.
+We searched PubMed for research on low socioeconomic status and morbidity, without language or date restrictions, up to Feb 10, 2019, and identified thousands of studies using the search terms “socioeconomic” in combination with “cancer”, “infection”, “cardiovascular”, “coronary heart disease”, “stroke”, and “psychiatric disorders”. Studies on socioeconomic status in relation to other diseases, such as “diabetes”, “endocrine disorder”, “respiratory disease”, “skin disease”, “neurodegenerative disease”, “dementia”, and “digestive disease” were also very common. Few studies examined “multimorbidity” and we found no research on temporal sequences in mental and physical diseases across all bodily systems according to socioeconomic status.
+Added value of this study
+To facilitate a more comprehensive evaluation of morbidity associated with socioeconomic disadvantage, we combined individual-level data from two large cohort studies and examined low socioeconomic status as a risk factor for a range
+of hospital-treated diseases. We determined temporal sequences in the emergence of diseases that were socioeconomically patterned. We repeated analyses in a third independent cohort. Across three indicators, low socioeconomic status was robustly associated with 18 (32-1%) of 56 specific diseases or health conditions, including 16 strongly interconnected conditions (hazard ratio >5 for each disease to be followed by another disease). This disease cascade started with psychiatric disorders, substance abuse, and self-harm and was followed later by diseases of the pancreas, liver, kidney, vascular and respiratory system, lung cancer, and dementia. Diabetes was associated with the cascade, but not with early psychiatric and behavioural disorders.
+Implications of all the available evidence
+Low socioeconomic status is a risk factor for a range of disorders, including mental and behavioural problems, which seems to set in motion a lifelong cascade of physical diseases. These findings suggest that policy and health-care practice addressing psychological health issues in social context and early in the life course might be an effective strategy for reducing socioeconomic inequalities in health.
+Methods
+Study design and population
+In this multi-cohort study, we used data from two Finnish prospective cohort studies: the Health and Social Support (HeSSup) study15 and the Finnish Public Sector (FPS) study.16 We tested the generalisability of our findings with an independent UK cohort study—the Whitehall II study.17 Ethical approval for these three studies was obtained from local committees on the ethics of human research. The derivation of the analytical sample used in each of these studies is shown in figure 1 and the appendix (p 2).
+In the HeSSup study, 21 486 of the men and women who responded to the survey between June 7, 1998, and May 23, 2000, or Jan 7, and Aug 12, 2003, had no missing data on residential area deprivation, and were successfully linked electronically to national hospitalisation and mortality registers until Dec 31, 2012.15 The FPS sample comprised 87 760 men and women who responded to at least one of four surveys done between March 1, 2000, and June 30, 2002, March 1, 2004, and June 30, 2005, March 1, 2008, and Nov 30, 2009, and Dec 1, 2011, and Nov 30, 2013, and had data on residential area deprivation.16 Study participants were linked to electronic health records until Dec 31, 2016.
+For our replication analyses, we used data from the Whitehall II study, which comprises 9838 government workers who participated in clinical examinations between Sept 10, 1985, and March 29, 1988, had no missing data on occupational position and covariates, and were linked electronically to national hospitalisation and mortality registers from Jan 1, 1997, when these records achieved a high level of national coverage, to March 31, 2017.18
+Assessment of socioeconomic status at baseline
+To explore the consistency of our results, we used three different indicators of socioeconomic status in our analyses. In the Finnish studies, we derived a score for residential area deprivation, similar to that developed by Townsend and colleagues,19 and a measure of educational attainment. The area deprivation score was obtained from Statistics Finland and is based on the proportion of adults with low education, the unemployment rate, and the proportion of people living in rented housing in each 250 m by 250 m grid area.20 Higher scores on the continuous index denote greater deprivation. We categorised these data as follows: low socioeconomic status (an area deprivation score higher than national mean), intermediate socioeconomic status (deprivation score from national mean to 0-5 SD below), and high socioeconomic status (the remaining data).
+Educational attainment, obtained from Statistics Finland via record linkage (for the FPS study) or from a survey (for the HeSSup study), was based on the following two categories: high (tertiary qualification, college or university) and low (all other qualifications, including none).
+In our replication analysis, we indexed socioeconomic status by a third indicator, the British civil service occupational grade.17 Broadly equivalent to the Registrar General’s indicators of occupational social class,21 this index of socioeconomic circumstances is related to salary, occupational prestige, level of responsibility at work, and future pension, and has three groups as follows: high (administrative occupations), intermediate (professional and executive occupations), and low (clerical and support occupations).
+Assessment of lifestyle risk factors at baseline
+Using predefined operationalisations, we chose the following baseline risk factors to determine the extent to which the associations between socioeconomic status and diseases were attributable to standard lifestyle factors: current smoking (yes vs no), risky alcohol use (consumption >210 g per week vs other), physical inactivity (yes vs no), and obesity (body-mass index >30 kg/m2 vs other).
+Follow-up for diseases, health conditions, and mortality Participants from the HeSSup study15 and FPS study16 were linked by their unique identification number to national registries of hospital discharge information (recorded by the National Institute for Health and Welfare) and mortality (recorded by Statistics Finland). These electronic health records include cause and date of hospitalisation or mortality and their coverage (all hospital types, including private hospitals, and records cover emergencies) reflects the comprehensive nature of Finland’s public health-care system. Additional information on site-specific cancers, diabetes, cardiovascular diseases (including hypertension), psychotic disorders, dementia, Parkinson’s disease, multiple sclerosis, epilepsy, asthma, chronic obstructive bronchitis, inflammatory bowel disease, liver disease, rheumatoid arthritis, gout, and renal failure was available via record linkage to the National Cancer Registry and the Drug Reimbursement Register of the Social Insurance Institution of Finland.
+Whitehall II study members were linked to the UK National Health Service (NHS) Hospital Episode Statistics (HES) database for hospital admissions and the NHS Central Registry for mortality. In studies of chronic diseases, the sensitivity and specificity of the HES database have been high.18,22
+In all cohort studies, the diagnosis for incident disease was coded according to the WHO International Classification of Diseases Tenth Revision (ICD-10). We focused on fifteen ICD-10 disease chapters that concern infectious and parasitic diseases (A00-B99), neoplasms (C00-D48), diseases of the blood (D50-D89), endocrine, nutritional, and metabolic diseases (E00-E90), mental and behavioural disorders (F00-F99), diseases of the nervous system (G00-G99), the eye (H00-H59), the ear (H60-H95), the circulatory system (I00-I99), the respiratory system (J00-J99), the digestive system (K00-K93), the skin (L00-L99), the musculoskeletal system (M00-M99), and the genitourinary system (N00-N99), injuries and poisoning (S00-T98), and external causes (V01-Y98).
+Statistical analysis
+Linked records captured 1204 ICD codes, including 56 major diseases or health conditions used in this analysis (for a complete list see appendix pp 2-13). Our primary analysis included two steps as follows: examination of associations between socioeconomic status (the exposure) and first new onset of health conditions after baseline (outcome) and mapping of
+Figure 1: Selection of participants for primary and replication analyses FPS=Finnish Public Sector. HeSSup=Health and Social Support.
+temporal sequences of interconnected health conditions in analyses stratified by socioeconomic status.
+First, having assessed the proportional hazards assumption (appendix pp 18-21), we examined associations between socioeconomic status and each of the 56 diseases in separate models using Cox proportional hazards regression. Follow-up continued until disease onset, death, or end of follow-up, whichever occurred first. Hazard ratios (HRs) computed for low socioeconomic status with high socioeconomic status as a reference were adjusted for the following potential confounding factors: age, sex, lifestyle factors (current smoking, heavy alcohol consumption, physical inactivity, and obesity), and cohort. In our analysis of socioeconomic status and new onset hospital-treated obesity, we did not control for baseline obesity. To identify diseases that were more common in participants with low socioeconomic status, related to socioeconomic differences that were likely to be meaningful for public health and unlikely to result from multiple testing,23,24 only socioeconomic status-disease endpoint associations that
+yielded predefined HRs equal to or greater than 1-223,24 and were statistically significant across the two different socioeconomic indicators (area deprivation and low education) were regarded as being sufficiently robust.
+Second, we determined potential temporal sequences at recorded diagnosis of diseases that were robustly associated with socioeconomic status by testing prospective associations between all socioeconomically patterned disease pairs separately in individuals with low socioeconomic status and high socioeconomic status. We used Cox proportional hazards regression and determined temporal order in testing the associations between disease pairs based on the mean age at diagnosis; a disease with an earlier onset was treated as the predictor and a disease with a later onset as the outcome. Followup started at recorded diagnosis for the first disease and continued until the date of diagnosis for the next disease, death, or end of follow-up, whichever occurred first. We adjusted HRs and 95% CIs for age, sex, and study.
+We constructed disease cascades from identified sequential interconnected disease pairs, starting from a single disease and continuing as far as interconnected disease pairs were available.25 We considered diseases and disease pairs to be interconnected if the HR for the association between them exceeded an arbitrary threshold of 5 in the socioeconomic status group, irrespective of the indicator used to define the group. In sensitivity analyses, we used alternative HRs of greater than 2-5 and greater than 10 as criteria for interconnectedness.
+For participants with high socioeconomic status, we analysed interconnected disease cascades, focusing on health conditions with HRs for indicators of socioeconomic status less than 1-0 (ie, diseases that were more common in high socioeconomic status groups than in low socioeconomic status groups). This more relaxed threshold was used because very few health conditions were more common in high socioeconomic status groups compared with low socioeconomic status groups.
+To examine the generalisability of the findings from the Finnish cohorts in the primary analysis across geographical regions and health-care settings, we tested robust associations between socioeconomic status and health conditions in a replication analysis using the Whitehall II cohort.
+All analyses were done using SAS version 9.4 and statistical code is provided in the appendix (pp 13-17).
+Role of the funding source
+The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. MK, JV, JP, and MJS had full access to all the data in the study. All authors had final responsibility for the decision to submit for publication.
+Results
+178 375 participants from the two cohorts of the primary analysis were eligible for inclusion (113 578 from the FPS
+study and 64797 from the HeSSup study). 109 246 (61-2%) participants responded to the baseline questionnaire, were successfully linked to registers of socioeconomic status and health, and were included in the analytic sample (figure 1). 83 066 (76-0%) of 109 246 participants were women. The mean participant age was 44-3 years (SD 11-0) and the range was 17-77 years. According to area-based deprivation, 36 216 (33-2%) participants were in the low socioeconomic status group. 52 990 (48-5%) participants were in the low education group (appendix p 23).
+During 1 110 831 person-years at risk, we recorded 245 573 hospitalisations in the 109 246 participants (figure 1). Compared with high socioeconomic status, low socioeconomic status was associated with an increased risk of 18 (32-1%) of the 56 health conditions for both indicators of socioeconomic status (HR >1-2; figure 2). By descending magnitude of association (ie, mean HR for the two indicators of socioeconomic status) these were self-harm, poisoning, psychotic disorders, arteriosclerosis, chronic obstructive bronchitis, lung cancer, dementia, obesity, disorders of substance abuse, pancreatitis, heart failure, anaemia, mood disorders, renal failure, diabetes, cerebral infarction, ischaemic heart disease, and disease of the liver. For disorders of substance abuse and ischaemic heart disease, the association was stronger in the first 3 years of follow-up than from year 3 onwards (appendix p 19). Minimally adjusted associations are also reported in the appendix (p 25).
+Three further health conditions were associated with area deprivation, but not low education, eight health conditions with low education, but not area deprivation, and 23 with neither area deprivation nor low education (figure 2). Four health conditions were more common in groups with high socioeconomic status (melanoma, spontaneous abortion, hypertension in pregnancy, and breast cancer). Of these four conditions, only breast cancer was associated with both indicators of high socioeconomic status (ie, low area deprivation and high education).
+Figure 3 shows interconnections between socioeconomically patterned health conditions and the mean age at diagnosis in participants with low socioeconomic status. The association between many disease pairs was stronger during the first 3 years of follow-up than from year 4 onwards (appendix p 22). 16 (88-9%) of 18 socioeconomically patterned health conditions were strongly interconnected as defined by HR greater than 5 for associations between disease pairs in the low socioeconomic status group, including both those with area deprivation and low education. Among participants
+with high socioeconomic status, we observed no strong interconnections between the four common diseases for this group (appendix p 26).
+Using mean age at recorded diagnosis, we were able to formulate a cascade of diseases in socioeconomically disadvantaged participants (figure 4). This cascade started
+with mental and behavioural disorders (psychiatric disorders, self-harm, and substance abuse) and was followed by pancreatitis, liver disease, anaemias, renal and heart failure, ischaemic heart disease, cerebral infarction, heart failure, arteriosclerosis, chronic obstructive bronchitis, lung cancer, or dementia. Diabetes was strongly connected
+with this cascade via association with renal failure, whereas hospital-treated obesity was not associated with any of the diseases in the cascade. In groups with low socioeconomic status defined using only one indicator, we observed additional connections between health conditions, particularly among participants with low education (appendix pp 27-28).
+When repeating our analysis of the 18 socioeconomically patterned health conditions among participants with low socioeconomic status using alternative thresholds (HRs 2-5 and 10) for connectedness (appendix p 29), the cascade of diseases starting from mental and behavioural disorders and including subsequent physical diseases remained apparent.
+In a subsidiary analysis of bidirectional associations, several physical diseases were associated with subsequent mental ill health—eg, heart failure (HR 3-18, 95% CI 1-31-7-71), cerebral infarction (3-68, 1-81-7-45), chronic obstructive bronchitis (3-63, 1-50-8-82), pancreatitis (7-65, 3-94-14-85), and renal failure (5-38, 2-23-13-02) predicted later mood disorders.
+The eligible population for the replication analysis was 14 121 men and women from the UK Whitehall II study.17 10 308 (73-0%) responded to the baseline survey and 9838 (69-7%) had no missing data on socioeconomic status or covariates and were successfully linked to electronic health records (figure 1; appendix p 24). In 186 572 person-years at risk (mean follow-up 19-0 years), we recorded 60 946 hospitalisations. All 18 associations between socioeconomic status and disease endpoints in the primary analysis were replicated (HRs >1-3; figure 5; minimally-adjusted HRs are presented in the appendix p 30). We found imprecision in the estimates for poisoning, selfharm, lung cancer, and arteriosclerosis as evidenced by the wide confidence intervals that included unity.
+Discussion
+In this study, we examined a range of mental and physical diseases and health conditions and found that low socioeconomic status was associated with 18 (32-1%) of the 56 diseases studied, independent of lifestyle factors and obesity and the indicator of socioeconomic status
+used (area deprivation, education, or occupational position). 16 (88-9%) of these 18 socioeconomically patterned diseases were interconnected, directly or indirectly, with mental health problems and substance abuse, including conditions such as pancreatitis, liver, renal, cardiovascular, and cerebrovascular diseases, chronic obstructive bronchitis, lung cancer, and dementia. With a less stringent threshold for interconnectedness, the cascade from mental disorders to physical illness was replicated and comprised all 18 diseases. When a higher threshold for interconnectedness was set—a minimum HR of 10 between diseases—mental health problems and substance use remained strongly connected with diseases of the liver, the cardiovascular and cerebrovascular system, and dementias, the latter emphasising the importance of socioeconomic patterns in diseases related to the CNS.
+Our findings are supported by several strands of evidence. The observed link between mental health and substance abuse, and between mental health and physical diseases, has been confirmed by meta-analyses measuring the impact of socioeconomically patterned adverse childhood experiences on mental disorders and chronic physical diseases in adulthood.26 The morbidity trajectories identified in our study included several of these chronic physical diseases, such as liver, respiratory, and cardiovascular diseases. Studies have shown that mental disorders increase the risk of physical diseases, both communicable and non-communicable, via a higher tendency to commit risky behaviours, reduced self-care, and complications in help-seeking.27,28 Additionally,
+psychotropic medications have adverse effects on many aspects of physical health that accumulate over time.27,28
+The link between mental and physical disorders could be strengthened by the bidirectional nature of this association. Although mental disorders are a risk factor for various physical diseases, the opposite direction of causality has also been observed. For example, socioeconomically patterned chronic conditions, such as cerebral infarction, heart failure, and chronic obstructive bronchitis, can increase the risk of mental disorders.28 We confirmed these associations and found associations between pancreatitis and renal failure and subsequent mood disorders.
+The range of health conditions in our study expands upon previous research on socioeconomic status and multimorbidity. A study of older Taiwanese people found low education to be associated with increasing trajectories of cardiovascular and chronic non-specific lung diseases12 and a UK study reported a link between low occupational grade and an increased risk of developing cardiometabolic multimorbidity (ie, two or three of diabetes, myocardial infarction, and stroke).3 A Canadian study identified an association between lower income and greater overall multimorbidity,13 a German study of primary care patients identified an association between low education and a higher number of diagnoses, particularly cardiometabolic diseases,14 and a study of Australian women found a relationship between low education and difficulties in managing income and increased self-reported multimorbidity in repeated questionnaire surveys.11
+We adjusted the association between socioeconomic status and health conditions for lifestyle behavioural factors, such as self-reported heavy alcohol consumption, smoking, physical inactivity, and obesity. This approach is conservative, as these factors are both confounders and potentially part ofthe causal pathway from socioeconomic disadvantage to disease. Prospective life-course research supports socioeconomic disadvantage as an origin of unhealthy lifestyle behaviours and subsequent morbidity. In a cohort study of Finnish children and adolescents, for example, differences in risk factors between socioeconomic groups at the beginning of follow-up were small, but large differences emerged in the third decade of life.20 In addition to risk behaviours, such as unhealthy diet, physical inactivity, and smoking, low socioeconomic status was associated with a poorer glycaemic profile in early adulthood and, like our findings, an excess prevalence of obesity, diabetes, fatty liver, and cardiovascular disease in middle age.20
+Our findings have important implications for research and public health policy. The pattern of mental health problems and substance abuse preceding socioeconomically patterned physical diseases is not reflected in global strategies to prevent diseases. The WHO Sustainable Development Goals and the Global Action Plan for the Prevention and Control ofNon-Communicable Diseases, for example, have their main focus on physical health;6 the 2013-2020 WHO Global Plan for the Prevention and Control of Non-Communicable Diseases29 and the Global Burden of Disease Collaboration do not include socioeconomic disadvantage as a modifiable risk factor.30 Moreover, treatment of psychiatric disorders, physical disease, and substance abuse is often split between health-care and social services.27,28 This approach is unlikely to be optimal for tackling problems with shared health determinants, including socioeconomic inequalities in morbidity. The 2019 Lancet Commission drew attention to the need to improve protection of physical health in people with psychiatric disorders;27 our findings suggest this is particularly important for people living in socioeconomic disadvantage.
+This study has several limitations. The response to baseline assessment varied between 61% in the primary analysis and 70% in the replication analysis. Sample attrition might lead to an overestimation or underestimation of the true associations between socioeconomic status and health. We measured morbidity mainly using electronic health records, which covered hospital-treated diseases. For some conditions, such as asthma, diabetes, and hypertension, additional nonhospitalised cases were identified via linkage to records of eligibility for special reimbursement for medication. However, we will have inevitably omitted undiagnosed conditions and less severe cases that are largely dealt with in primary care (eg, obesity). Therefore, the observed interconnectedness between diseases reflected the temporal order of treatments rather than causal
+associations between health conditions. The age distribution of participants at study induction meant that we did not have data on children or very old people. The generalisability of our findings beyond Finland and the UK and to other health-care systems is also uncertain and requires testing.
+However, by applying a data-driven approach to a wide set of diseases and health conditions, we refocus the field of socioeconomic inequality research from traditional analysis of specific diseases to the study of interconnected diseases. A large sample size, longitudinal design, minimal sample attrition after baseline because of follow-up via electronic health records, and the validation of our results across different indicators of socioeconomic status and health-care settings are further strengths.
+In conclusion, by mapping morbidity from electronic health records we showed that low socioeconomic status is a risk factor for a spectrum of interconnected diseases and health conditions. Our analyses of interconnected diseases highlight the importance of mental health problems and substance abuse in the cascade of socioeconomically patterned physical illnesses. These findings suggest that policy and health-care practice addressing mental health issues in social context and early in the life course might be effective strategies for reducing health inequalities.
+Articles
+(RG/16/11/32334), the Academy of Finland (311492), Helsinki Institute 16 of Life Science, and the Finnish Work Environment Fund (190424).
+References
+1	Dalstra JA, Kunst AE, Borrell C, et al. Socioeconomic differences in the prevalence of common chronic diseases: an overview of eight European countries. Int J Epidemiol 2005; 34: 316-26.
+2	Stringhini S, Carmeli C, Jokela M, et al. Socioeconomic status, non-communicable disease risk factors, and walking speed in older adults: multi-cohort population based study. BMJ 2018; 360: k1046.
+3	Singh-Manoux A, Fayosse A, Sabia S, et al. Clinical, socioeconomic, and behavioural factors at age 50 years and risk of cardiometabolic multimorbidity and mortality: a cohort study. PLoS Med 2018; 15: e1002571.
+4	Mackenbach JP, Stirbu I, Roskam AJ, et al. Socioeconomic inequalities in health in 22 European countries. N Engl J Med 2008; 358: 2468-81.
+5	WHO. Commission on Social Determinants of Health. Closing the gap in a generation: health equity through action on the social determinants of health. Final Report of the Commission on Social Determinants of Health. Geneva: World Health Organization, 2008.
+6	UN. Transforming our world: the 2030 agenda for sustainable development. New York, NY: United Nations, 2015.
+7	Frolich A, Ghith N, Schiotz M, Jacobsen R, Stockmarr A. Multimorbidity, healthcare utilization and socioeconomic status: a register-based study in Denmark. PLoS One 2019; 14: e0214183.
+8	van den Akker M, Buntinx F, Metsemakers JF, Roos S, Knottnerus JA. Multimorbidity in general practice: prevalence, incidence, and determinants of co-occurring chronic and recurrent diseases. J Clin Epidemiol 1998; 51: 367-75.
+9	Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study.
+Lancet 2012; 380: 37-43.
+10	Park B, Ock M, Lee HA, et al. Multimorbidity and health-related quality of life in Koreans aged 50 or older using KNHANES 2013-2014. Health Qual Life Outcomes 2018; 16: 186.
+11	Jackson CA, Dobson A, Tooth L, Mishra GD. Body mass index and socioeconomic position are associated with 9-year trajectories of multimorbidity: a population-based study. Prev Med 2015; 81: 92-98.
+12	Hsu HC. Trajectories of multimorbidity and impacts on successful aging. Exp Gerontol 2015; 66: 32-38.
+13	Canizares M, Hogg-Johnson S, Gignac MAM, Glazier RH, Badley EM. Increasing trajectories of multimorbidity over time: birth cohort differences and the role of changes in obesity and income. J Gerontol B Psychol Sci Soc Sci 2018; 73: 1303-14.
+14	Schafer I, Hansen H, Schon G, et al. The influence of age, gender and socio-economic status on multimorbidity patterns in primary care. First results from the multicare cohort study.
+BMC Health Serv Res 2012; 12: 89.
+15	Korkeila K, Suominen S, Ahvenainen J, et al. Non-response and related factors in a nation-wide health survey. Eur J Epidemiol 2001; 17: 991-99.
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+Kivimaki M, Lawlor DA, Davey Smith G, et al. Socioeconomic position, co-occurrence of behavior-related risk factors, and coronary heart disease: the Finnish Public Sector study.
+Am J Public Health 2007; 97: 874 79.
+Marmot MG, Smith GD, Stansfeld S, et al. Health inequalities among British civil servants: the Whitehall II study. Lancet 1991; 337: 1387-93.
+Kivimaki M, Batty GD, Singh-Manoux A, Britton A, Brunner EJ, Shipley MJ. Validity of cardiovascular disease event ascertainment using linkage to UK hospital records. Epidemiology 2017; 28: 735-39.
+Townsend P, Beattie A, Phillimore P. Health and deprivation: inequality and the North. London: Croom Helm, 1988.
+Kivimaki M, Vahtera J, Tabák AG, et al. Neighbourhood socioeconomic disadvantage, risk factors, and diabetes from childhood to middle age in the Young Finns Study: a cohort study. Lancet Public Health 2018; 3: e365-73.
+Elovainio M, Ferrie JE, Singh-Manoux A, et al. Socioeconomic differences in cardiometabolic factors: social causation or health-related selection? Evidence from the Whitehall II Cohort Study, 1991-2004. Am J Epidemiol 2011; 174: 779-89.
+Sommerlad A, Perera G, Singh-Manoux A, Lewis G, Stewart R, Livingston G. Accuracy of general hospital dementia diagnoses in England: sensitivity, specificity, and predictors of diagnostic accuracy 2008-2016. Alzheimers Dement 2018; 14: 933-43.
+Olivier J, May WL, Bell ML. Relative effect sizes for measures of risk. Commun Stat 2017; 46: 6774-81.
+Siontis GC, Ioannidis JP. Risk factors and interventions with statistically significant tiny effects. Int J Epidemiol 2011; 40: 1292-307 Jensen AB, Moseley PL, Oprea TI, et al. Temporal disease trajectories condensed from population-wide registry data covering 6-2 million patients. Nat Commun 2014; 5: 4022.
+Hughes K, Bellis MA, Hardcastle KA, et al. The effect of multiple adverse childhood experiences on health: a systematic review and meta-analysis. Lancet Public Health 2017; 2: e356-66.
+Firth J, Siddiqi N, Koyanagi A, et al. The Lancet Psychiatry Commission: a blueprint for protecting physical health in people with mental illness. Lancet Psychiatry 2019; 6: 675-712.
+Prince M, Patel V, Saxena S, et al. No health without mental health. Lancet 2007; 370: 859-77.
+WHO. Global action plan for the prevention and control of noncommunicable diseases 2013-2020. Geneva: World Health
+Organization, 2013.
+GBD 2017 Risk Factor Collaborators. Global, regional, and national comparative risk assessment of 84 behavioural, environmental and occupational, and metabolic risks or clusters of risks for 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2018; 392: 1923-94.
+e149
+www.thelancet.com/public-health Vol 5 March 2020

Association between suicide reporting in the media and suicide systematic review and meta-analysis.txt

@@ -0,0 +1,78 @@
+Introduction
+News reporting of suicide has increased substantially in recent decades.1-4 A number of studies have shown that media reports of suicide are associated with increased numbers of suicides.5-10 Media related imitation of suicide has been dubbed the Werther effect, based on a reported spike in suicides in young men in Germany and across Europe after the publication of Goethe’s The sorrows of young Werther in 1774, depicting the circumstances leading to the suicide of the male protagonist
+Werther.11 More than 150 studies have investigated the effects of suicide related to media reports.10 Most have used before and after comparisons or time series designs, testing whether media reporting was associated with subsequent changes in suicides at an aggregate level across a region of exposure. The Werther effect is discussed mostly in relation to non-fictional news stories,8 particularly stories about deaths of celebrities by suicide,6 and stories with a dramatic or romanticised depiction of suicide, or featuring an explicit and detailed description of a suicide method.12-14
+In acknowledgment of the Werther effect, mental health and suicide prevention organisations worldwide, including the World Health Organization, have developed guidelines for responsible reporting of suicide by the media with a specific focus on news and information media.15 16 These guidelines are now a standard component of many national and regional suicide prevention strategies.16 Typically included in the guidelines are specific suggestions about ways to minimise harm (eg, by avoiding glorification of suicide, discussions of specific suicide methods, and repeated reporting about the same suicide). The guidelines also recommend including information on the role of treatable mental illness, where and how to seek help for suicidal thoughts, and a message of hope that suicide is preventable. But specific information on individual deaths by suicide continues to be published; the suicide of the actor Robin Williams is an example of the guidelines not being fully followed.17 18
+Controversies around suicide and the media remain, despite a global focus on avoiding the Werther effect and compelling associations in the literature. Research shows that not all media coverage of suicide is associated with subsequent increases in suicides, resulting in a debate lasting decades on the impact of media reporting of suicide on subsequent suicides.9 10 13 14 19 In several countries that have implemented media guidelines, journalists and media professionals have pushed back, arguing that the body of evidence is not compelling enough to warrant changes to the way suicide is reported.10 20
+Meta-analyses can better quantify the combined evidence of a Werther effect across published studies, but these studies are scarce. One meta-analysis of 10 studies examined media reporting on deaths of celebrities by suicide and found an average increase of 2.6 suicides per million people (95% confidence interval 0.9 to 4.3) in the month after the reports of death.6 In the largest meta-analysis so far, Stack9 combined findings from 55 studies examining non-fictional reports of suicide as a predictor of suicide, and found that only 36% identified an apparent Werther effect. This meta-analysis did not, however, define clear inclusion and exclusion criteria; consider the quality of the studies; account for potential duplication of results; and, crucially, involve the abstraction of quantitative data on suicides (as is normally the case). The outcome of the meta-analysis was a binary variable of increase versus no increase in suicides.
+Media coverage of celebrity deaths by suicide is a small proportion of all suicide reporting8-10 13 14 and the guidelines make recommendations about all forms of reporting of suicide.16 Meta-analyses on the effects of general reporting of suicide (that is, any reporting related to suicide) are lacking. General reporting of suicide might involve deaths of celebrities or other individuals, or might include more general discussions on the topic of suicide. These studies typically use broad search terms to identify media reports (eg, suicide or various suicide methods).
+The aim of this systematic review and meta-analysis was to examine and quantify the findings from the literature on the
+Werther effect. We aimed to evaluate the effects of three types of media reporting on suicide on the subsequent incidence of suicide. The primary objective was to summarise the evidence on the association of media reporting of deaths of celebrities by suicide on total suicides over a short period of time (up to two months). The secondary objectives were to summarise the association of media reporting of information about the specific methods used by the celebrities on suicides by the same method, and the association of general reporting of suicide on the total number of suicides. We hypothesised that reporting of the deaths of celebrities by suicide would be associated with an increased incidence of suicide in the general population, and that increases by the same method would be strongest. We did not have a clear hypothesis for general reporting of suicide because of the variety of content, some of which might be harmful and some protective.13 14 For our meta-analysis, we use the term “intervention” to refer to media reporting of suicide. The study was conducted according to the meta-analyses of observational studies in epidemiology (MOOSE) guidelines.
+Methods
+Search strategy
+We defined news and information media as all non-fictional accounts of suicide on TV, in print, in online news, or in educational non-fiction media (eg, non-fiction books or films). Studies on the effects of searching for suicide related information online (eg, Google searches) were not eligible because these studies do not distinguish between positive (eg, for help services) and negative (eg, pro-suicide websites) searching.21 We searched PubMed/Medline, Embase, PsycInfo, Scopus, Web of Science, and Google Scholar for relevant studies from their inception to September 2019. These databases show modest to strong overlap in coverage.22 Google Scholar was used specifically to identify grey literature.23 We used the search terms suicide (suicid*) AND imitation (Werther; Papageno; copycat; imitat*; contagio*; suggesti*); AND media (media; newspaper*; print; press; radio*; televis*; film*; book*; documentar*; internet; cyber*; web*).
+The titles and abstracts of the retrieved articles were screened for relevance, and the full text versions of studies that might meet the inclusion criteria were reviewed. The reference lists of the full text articles were also screened for relevant studies, and a cited reference search was conducted for all relevant primary articles with Google Scholar. English and non-English language articles were included. Non-English articles often had English abstracts, and we used Google Translate and consulted with fluent language speakers to assess the inclusion criteria and extract the data.
+Study selection
+Studies were eligible for inclusion if they used a before-and-after design, compared single or multiple times before-and-after media reports related to suicide, or an interrupted times series design; if they used death by suicide as the outcome variable; and if they reported non-fictional media stories (that is, stories in news and information media).
+Exclusion criteria
+We excluded studies that did not have original data. We also excluded studies that examined associations in subgroups of the population because the findings might not be representative of the total population. For our analysis of media reporting on the method of suicide, we excluded studies reporting on an
+emerging new suicide method if the incidence of the respective suicide method at baseline (that is, before onset of media reporting) was low (<5%). These studies also typically measured possible effects over a longer than usual period of time. We excluded studies that provided only associations for a follow-up period of more than two months, because this is beyond the typical time frame for studying imitation effects, and might be based on mechanisms that are different from imitation.24 Also excluded were studies with data before the second world war; those with media interventions that were not about suicide; those which applied non-eligible designs; those that were at critical risk of bias; or those that duplicated data from another study.
+If studies had duplicated data (data on the same celebrities in the same setting reported in more than one study), we included one study. We selected this using a hierarchical approach based on: (1) the lowest risk of bias; (2) covering the longest period of time or the largest number of celebrities; and (3) the most recent. The 31 studies selected were included in the qualitative and quantitative synthesis (supplementary appendix).
+Data extraction
+We extracted these data from the studies: study location; study period and length; length of the observation period after media reporting; unit of analysis at which outcome data were measured (eg, daily or weekly); how the media intervention was measured (eg, binary variable representing the presence or absence of reporting or a continuous variable representing the number of news stories); whether the study reported on deaths of celebrities by suicide or general reports of suicide; number of interventions (eg, number of media reports over time); type of media (print media v other forms of media, such as television, online, or mixed media); any outcome reported related to the specific suicide method used in a reported suicide (exclusively or in addition to total suicides); whether the analysis was adjusted or unadjusted for confounders (in addition to any adjustments for seasonal or long term time trends); which confounders were measured and adjusted for; type of estimate extracted (rate ratio or expected and observed suicides); study design (single arm before and after comparison, multiple arm before and after comparison, interrupted time series)25-28; analysis technique; method to control for time trends; and source of the outcome data. In a single arm before-and-after comparison, suicides were observed in one group before and after the intervention. In a multiple arm before-and-after comparison, suicides were observed in multiple groups because there were multiple sites for one intervention or one site but multiple interventions occurring at different times.25
+Additional information was obtained for studies of deaths of celebrities by suicide: number of celebrities; type of celebrity (eg, entertainer); and level of recognition of the celebrity (local, international). For level of recognition, we used information from the study and online sources (eg, Wikipedia). Local celebrities were famous in one country or region (eg, a local politician) and international celebrities were known in a western or global context or were described in the original publication as international. A mixed code was used for celebrities with different levels of recognition. For studies looking at increases in the incidence of suicide by the same method as reported in the media, we recorded the suicide method.
+We obtained rate ratios and standard errors from each study by one of the following methods:
+•	Extracting directly a rate ratio and either a standard error, 95% confidence interval, t value, or other estimate to calculate a standard error
+•	Using the number of expected and observed suicides to calculate rate ratios and standard errors
+•	Extracting the observed number of suicides in the before and after intervention periods (along with the corresponding times) and calculated rate ratios and standard errors
+•	Obtaining a coefficient and standard error from a linear regression model that was converted to a rate ratio with the study’s population at the mid-point
+•	The authors of the original study providing us with rate ratios and standard errors.
+For each study, we recorded how the estimate was derived (obtained directly from the study, combined estimates using meta-analysis, or reanalysis of the data by the authors).
+We aimed for one quantitative outcome, but two studies (table S1) reported multiple quantitative estimates because the results were presented separately for different news sources. Hence we combined these into one estimate using random effects meta-analysis (see below).
+The search strategy was performed by two of the authors (TN and MB). Decisions on excluding studies after full text review were made by TN and separately by MJS. Discrepancies were discussed and resolved. Quantitative data were abstracted by MJS and discussed with TN. Metadata of studies were obtained by TN and MB initially, and separately by MJS. Discrepancies were discussed and resolved among the team.
+Risk of bias
+Risk of bias was assessed for each study based on the Robins-I tool.29 This tool was originally designed for non-randomised cohort studies, and does not directly apply to our study designs. The general concept, however, is applicable to interrupted time series designs,30 and the authors of Robins-I have published on issues that will be looked at in a future version for studies of interrupted time series.31 We developed a specific adaption for this study with six domains of bias: bias as a result of confounding issues; bias in classification of interventions; bias because of preparatory phases; bias because of missing data; bias in measurement of the outcome; and bias in selection of reported results.
+Studies were considered at low risk of bias if all domains were coded as low risk; at moderate risk if at least one domain was coded moderate but none as serious; at serious risk if at least one domain was assessed as serious but none as critical; and at critical risk if any domain was coded as critical. Like an earlier study that applied the Robins-I tool to natural experiments,30 we found that the first domain, risk of bias as a result of confounding, generally determined the overall risk of bias. This domain comprised coding for subdomains if the number of pre-intervention times was sufficient to allow characterisation of the series; appropriate analysis techniques were used to account for time trends and time patterns; seasonality was accounted for; and possible confounders were measured and controlled for. Risk of bias because of selective reporting was also relevant for some studies. We assessed if the outcome measurement and analyses were clearly defined and consistent in the methods and results sections of the studies, and if there was some risk of selective reporting from multiple analysis methods, multiple follow-up times, or multiple subgroups. The full quality assessment plan is in the supplementary appendix.
+As recommended in the Robins-I tool, studies with up to moderate risk were included in the primary and secondary analyses, and studies at serious risk were included in sensitivity analyses only. Studies at critical risk of bias were excluded.
+Assessments of the risk of bias were based only on the data we abstracted. If the authors provided a reanalysis of their data, for example, only the reanalysis was assessed for risk of bias, not the original study. Similarly, if a study reported total suicides as a side outcome, only the components relevant to the abstracted data (total suicides) were assessed. Our quality ratings, therefore, do not always apply to the original studies. Risk of bias was assessed independently by TN and MJS, and discrepancies were discussed and resolved.
+Quantitative data synthesis
+We described the studies using descriptive statistics. For our primary analysis, we estimated the pooled rate ratio for the effect of media reporting on deaths of celebrities by suicide on total suicides. We also conducted two secondary analyses. In the first (secondary analysis A), we estimated the pooled rate ratio for reporting about the method used in a suicide by a celebrity on suicide by the same method. In the second (secondary analysis B), we estimated the pooled rate ratio for general reporting of suicide on total suicides. The primary and two secondary analyses were restricted to studies at moderate risk of bias. In sensitivity analyses, we repeated these analyses adding studies at serious risk of bias.
+All pooled rate ratios were estimated with a random effects model, with standard errors calculated by the Knapp-Hartung method.32 Heterogeneity of effect sizes was assessed with the I2 statistic and Cochran’s Q test. For I2, values around 25% indicated low heterogeneity, around 50% moderate heterogeneity, and around 75% high heterogeneity.33 Publication bias was assessed visually by contour enhanced funnel plots34 and quantitatively with Egger’s regression test for asymmetry.35
+Sources of heterogeneity
+Meta-regression was used to identify the factors that might contribute to heterogeneity. We conducted univariate meta-regressions for each variable and combined significant variables (P<0.05) into a multivariate model. The meta-regressions were estimated with a random effects model with standard errors calculated by the Knapp-Hartung method. We combined the coefficients algebraically (that is, a linear combination of coefficients presented on the exponential scale) so that we could show rate ratios and 95% confidence intervals in each category of a variable. For the studies in the primary analysis, we examined the period published (up to 2005, 2006-10, 2011-15, 2016 or later), follow-up time (< 14 days, >15 days) location (Asia, Europe, North America-Australia), design (multiple arm before-and-after comparison, interrupted time series analysis), length (per 1000 days), period of analysis (day, week, month), adjustment for confounders (no, yes), celebrity recognition (local, international, mixed), celebrity type (entertainer, other), and number of celebrities (1, >2). We used similar variables for the studies in secondary analysis A, along with a variable on the method of suicide reported (hanging v other methods) but combined several categories where only one study was available for analysis. We did not conduct a meta-regression for studies in secondary analysis B because heterogeneity was low. All analyses were conducted in Stata 16.0. This study was registered with PROSPERO (https://www. crd.york.ac.uk/PROSPERO/, registration No CRD42019086559, 18 January 2019).
+Patient and public involvement
+There were no funds or time allocated for patient and public involvement so we were unable to involve patients. We have invited patients to help us develop our dissemination strategy.
+Results
+Study characteristics
+We retrieved 8823 references and 1496 remained after removal of duplicates (fig 1). After screening the titles and abstracts, the full texts of 143 studies were assessed and 112 were excluded: 26 because suicide was not an outcome; 30 because of strong data duplication with other studies; 17 because the intervention (media story) was not about suicide; 19 because of reports on emerging suicide methods; five because suicide was analysed only in a population subgroup; and five had data from before the end of the second world war. Also excluded were: two case studies; two studies measuring the outcome for longer than the maximum follow-up; one study about a fictional intervention; and two studies with annual outcome data. After quality assessment, three studies were excluded because of a critical risk of bias. The remaining 31 studies were included in our review: 23 were from database searches, three from Google Scholar, and five from cross reference searches.
+Study characteristics are summarised in table 1 and table 2 (and table S1). The 31 studies were published between 1974 and 2019 and examined the period 1947 to 2016. Nineteen studies examined the total number of suicides as the outcome and two examined increases in suicides by the same method reported in the media; 10 studies reported both. Twenty two studies examined media reporting of deaths of celebrities by suicide and nine studies evaluated general reporting of suicide. Studies were from Asia (Taiwan, Hong Kong, South Korea, and Japan), Europe (Austria, Germany, Hungary, the Netherlands, Slovenia, France, and Israel), North America (United States and Canada), and Australia. Most studies (n=20) used an interrupted time series design, 10 a multiple arm before-and-after design, and one a single arm before-and-after design. Seven studies had follow-up of 1-7 days, eight had 8-14 days, 12 had 15-30 days, and four had 31-60 days. The median follow-up time was 21 days (range 1-60 days).
+Quality assessment
+We classified 24 studies as being at moderate risk of bias because of confounding issues and seven at serious risk of bias. We judged 22 studies as being at low risk of bias because of classification of interventions, six at moderate risk, and three at serious risk. All 31 studies were at low risk of bias because of preparatory phases. Twenty eight studies were at low risk of bias because of missing data, two were at moderate risk, and for one the risk was unknown. Thirty studies were at low risk of bias because of measurement of the outcome, and one was at serious risk. Twenty nine studies were judged to be at moderate risk of bias because of selection of reported results, and two were at serious risk. Overall, 20 studies were assessed as moderate risk and 11 as serious risk of bias (table S2).
+Quantitative data synthesis
+Figure 2, figure 3, and figure 4 show the forest plots for the primary and secondary analyses. For the primary analysis (fig 2), on the impact of media reporting of deaths of celebrities by suicide on total suicides, 14 studies met the inclusion criteria. The pooled rate ratio was 1.13 (95% confidence interval 1.08 to 1.18, P<0.001) over a median follow-up of 28 days (range
+7-60 days). For the secondary analysis A (fig 3), on reporting of method of suicide of celebrities on suicides by the same method, 11 studies met the inclusion criteria. The pooled rate ratio was 1.30 (95% confidence interval 1.18 to 1.44, P<0.001) over a median follow-up of 28 days (range 14-60 days). For the secondary analysis B (fig 4), on the impact of general reporting of suicide on total suicides, five studies met the inclusion criteria and the pooled rate ratio was 1.002 (95% confidence interval 0.997 to 1.008, P=0.25) for a one article increase in the number of reports on suicide. The median follow-up was 1 day (range 1-8 days).
+Heterogeneity
+Estimates of heterogeneity were large and significant for the primary analysis (I2=83.5%, P<0.001) and the secondary analysis A (I2=72.1%, P<0.001) but not for the secondary analysis B (I2=0.02%, P=0.40). We therefore undertook meta-regressions for the first two sets of studies to identify possible sources of heterogeneity.
+In univariate analyses of the 14 studies in the primary analysis, differences in the pooled rate ratios between subgroups were observed for three variables (table 3): publication date (P=0.04, I2=29.1%), celebrity type (P=0.009, I2=47.9%), and number of celebrities under investigation (P=0.009, I2=47.0%). Weak evidence that reporting of deaths of celebrities by suicide was associated with suicides was found for studies published before 2005 (rate ratio 1.12, 95% confidence interval 0.99 to 1.26, two studies) but clear evidence of a positive association was found for the three other times (2006-10: 1.13, 1.04 to 1.23, three studies; 2011-15: 1.06, 1.02 to 1.10, four studies; >2016: 1.16, 1.11 to 1.22, five studies). Suicides by entertainers showed a positive association between reporting and suicide (rate ratio 1.17, 1.12 to 1.23, six studies) as did studies about other types of celebrities (1.08, 1.04 to 1.12, eight studies). Studies about one celebrity (1.17, 1.12 to 1.23, seven studies) and multiple celebrities (1.08, 1.04 to 1.12, seven studies) showed positive associations between reporting of deaths of celebrities by suicide and suicide.
+Because celebrity type and number of celebrities were collinear, we entered only publication date and number of celebrities in a multivariate meta-regression. We found no differences between subgroups in the pooled rate ratio for either variable (table S3) but the overall I2 for the model was lower, indicating low to moderate heterogeneity compared with the primary analysis (I2=34.6%). In the meta-regressions of the nine studies in the secondary analysis A, none of the factors was associated with the reporting of method of suicide on total suicides (table S4).
+Publication bias
+Figure 5 shows the contour enhanced funnel plots for the three analyses. For the primary analysis, the funnel plot was asymmetrical with more study specific rate ratios falling to the right of the pooled rate ratio line than the left. Few rate ratios were within the P value greater than 10% contours of statistical significance. Studies appeared to be missing from the region between the null value and the pooled rate ratio. Egger’s regression test for funnel plot asymmetry was significant (P=0.01). The funnel plots for the two secondary analyses were symmetrical around the pooled rate ratio, and Egger’s test was not significant for either analysis (P=0.23 for secondary analysis A and P=0.13 for secondary analysis B).
+Sensitivity analyses
+We undertook sensitivity analyses that included studies at serious risk of bias (fig S1). For media reporting of deaths of celebrities by suicide on total suicides, 20 studies met the inclusion criteria. The pooled rate ratio was 1.10 (95% confidence interval 1.06 to 1.14, P<0.001, I2=93.4%) over a median follow-up of 28 days (range 7-60 days). Investigation of heterogeneity failed to identify new factors that could account for differences between the studies. Heterogeneity remained large and persisted for all variables (table S5). Egger’s regression test for funnel plot asymmetry (fig S2) was close to significance (P=0.06). For reporting of the suicide method used by a celebrity on suicide by the same method (12 eligible studies), the pooled rate ratio was 1.32 (95% confidence interval 1.19 to 1.47, P<0.001, I2=74.9%) over a median follow-up of 28 days (range 14-60 days). We were unable to find sources of heterogeneity (table S6). Egger’s test was not significant (P=0.10). For general reporting of suicide (nine studies, median follow-up 7 days, range 1-30 days), the pooled rate ratio was 1.002 (95% confidence interval 0.999 to 1.005, P=0.11, I2=0.02%) for an increase of one article. Egger’s test was not significant (P=0.60). Because three studies in the primary analysis were about the same celebrity (Robin Williams), we performed a final sensitivity analysis where we excluded two of the studies, retaining the study with the lowest risk of bias.36 The pooled rate ratio was 1.10 (95% confidence interval 1.06 to 1.15, P<0.001, I2=61.8%) over a median follow-up of 28 days (range 7-60 days).
+Discussion
+Main findings
+To our knowledge, this systematic review and meta-analysis is the most comprehensive to date of the effects of media reporting of suicide on subsequent suicides. The evidence indicates an increase in total suicides in the period after the reporting of a death of a celebrity by suicide. When the suicide method used by the celebrity was reported, evidence of a corresponding increase in the number of suicides by the same method was found. This effect appeared to be larger than for increases in total suicides, although suicides by a specific method typically only account for a limited proportion of all suicides. General reporting of suicide did not appear to be associated with increases in total suicides but the evidence was based on a small number of studies, mainly from the same region of the world.
+At least three mechanisms might explain the increases in the number of suicides associated with reporting of suicide: identification with the deceased person, which might occur more frequently when the reported suicides are about individuals with high social standing37 38; increased media reporting of suicide leading to normalisation of suicide as an acceptable way to cope with difficulties7; and information on suicide methods, which might influence the choice of suicide method by a vulnerable individual.38 Our findings support several of these mechanisms. Firstly, reporting on deaths of celebrities by suicide appears to increase total suicides, suggesting that the phenomenon goes beyond the influence of knowing the suicide method used by the celebrity. Secondly, some evidence exists of stronger effects in studies focusing on suicide by entertainers, compared with other celebrities, consistent with their strong public identity, which has been previously described for entertainment celebrities in particular.39 Studies that focused on increases in suicide after one (rather than several) suicide by a celebrity often reported on entertainers, suggesting that these celebrities were well known and of interest to the public. Thirdly, the
+finding of a pronounced increase in suicide by the same method as that of a celebrity suggests that transfer of information about the method might be another relevant factor in the association. Media reporting on a suicide method increases the cognitive availability of this method,7 and individuals considering suicide might be more likely to subsequently select the method used by celebrities. The evidence suggests that suicide by hanging, for example, especially among men aged 45-64 years, increased after the suicide of Robin Williams by the same method.36
+Support for the effect of media coverage of suicide also comes from individual level studies that typically used outcomes such as suicidal thoughts rather than suicidal behaviour. Harmful effects on mood, self-esteem, and suicidal thoughts, especially in those who have previously contemplated suicide, have been identified.40-42 Individuals with suicidal thoughts, particularly new thoughts and a suicide plan, have an increased risk of suicidal behaviour.43 The increases in total suicides, and greater increases in suicides by the same method reported in the media, as identified in our meta-analysis, suggest that media stories on deaths of celebrities by suicide might do both: increase suicidal thoughts and contribute to planning suicide with a specific method. Suicidal thoughts are a common occurrence. A recent survey in the United States estimated that 9.4 million adults (4% of the population) had seriously considered taking their own life in the previous 12 months, and 2.7 million (1% of the population) made plans to do so,44 suggesting media reports of suicide have the potential to negatively influence many vulnerable people who might be swayed by news items.
+We found that the size of the association for suicides after the reporting of deaths of celebrities by suicide was smaller than in a previous meta-analysis that included fewer studies and did not assess the risk of bias comprehensively (based on 10 studies).6 Reasons for the smaller association might include the broader literature search, exclusion of duplicate data, and exclusion of studies at critical risk of bias. A rate ratio of 8-18% increase in suicide, however, highlights that media exposure relating to deaths of celebrities by suicide has a strong influence on the incidence of suicide in a population. In contrast, the global financial crisis of 2009 was associated with a 6% increase in suicide (although over a longer period of time).45 The estimated increase for reporting on deaths of celebrities by suicide might also underestimate the effects of media reporting on well known celebrities. Some of the studies included focused on individuals with questionable prominence, including mid-level or regional politicians and others not likely to be known by most of the population.11 Estimates were higher for well known celebrities, such as Robin Williams. Unlike in fixed effects meta-analyses of drug trials in defined populations, no true single effect exists for the association of media reporting on suicide with the number of subsequent suicides. Associations will probably vary depending on factors such as the prominence of the person in the media reports, the population’s connection with that person, and the extent to which the death is reported responsibly by the media in the region where the study is conducted. The World Health Organization has emphasised that media professionals should be cautious when reporting on suicides in general and on deaths of celebrities by suicide in particular.16
+For general reporting on suicide, taking into account all media reporting of suicide, no association with increases in the number of suicides was found. These studies usually evaluated the effect of the number of news articles on suicide on the next day or in the next week whereas studies of reporting on deaths of celebrities reported the presence or absence of a death of celebrity by suicide. The studies on general reporting of suicide
+also tended to use wide ranging search strategies to identify a broad variety of media reports related to suicide. This search strategy might have resulted in media reports associated with suicides but might have been distorted by inclusion of other reporting types that do not cause harm. Previous research suggests that not all reporting on suicide is associated with increases in the number of suicides.13 14 The risk appears to vary with reporting characteristics.13 14 Increases are particularly likely for a subset of media reporting that describes suicide methods13; depicts suicide as inevitable14; or publicises false public myths about suicide.13 Some media reports on suicidal thoughts feature stories of hope and healing, rather than suicide attempts or deaths, and might help to prevent suicides (the so-called Papageno effect).13 46 47 None of the studies in our meta-analysis considered the qualities of media reports based on media recommendations, and the variability in reporting qualities is likely large. Hence, a resulting underestimation of the effects for media stories that are inconsistent with media recommendations is possible. Future research should aim for a clear definition of the reporting to separate associations for different types of reports.
+Our meta-analysis generally took a conservative approach by limiting the analysis to studies at moderate risk of bias and focusing entirely on total suicides (rather than subgroups) as the outcome for the primary analysis. If, for example, a study reported on the effects of news media reporting on the incidence of suicide in teenagers, the data extracted for our meta-analysis were for the total population, even if the study put a focus on its specific findings for the subgroup of teenagers. This approach was to ensure that selective reporting of findings in subgroups did not bias our estimates.
+Like previous reviews,6 9 we found strong heterogeneity in risk estimates across studies on reporting of suicide by celebrities in particular. A large part of the heterogeneity was a result of the type of celebrity or number of celebrities analysed, suggesting that individuals best known to the public are those most likely to trigger more suicides. None of the characteristics explained the method specific increases in suicides. The remaining unexplained heterogeneity suggests that factors out of scope of our analysis might impact on the risk of increases in suicides after suicide reporting, including overall trends in the incidence of suicides in a country or over a period of time when media reporting occurs; socioeconomic conditions that might influence suicide reporting and imitation effects; and precise measurement of social identification with celebrities and other individuals who die by suicide.
+We observed a number of significant and positive effect sizes and an absence of non-significant effect sizes in some regions of the funnel plots for studies of reporting on deaths of celebrities by suicide. This lack of symmetry could indicate publication bias. Many factors can contribute to asymmetry in funnel plots, however, and precise interpretation is difficult when the underlying evidence is based on observational data.48 49 Unpublished studies could have shown no association. If true, this means that our meta-analysis will have overestimated the association between media reporting of suicide by celebrities and subsequent suicides. A sensitivity analysis including studies with serious risk of bias indicated a similar pattern to the overall findings, suggesting that effect estimates for lower risk studies were similar.
+Strengths and limitations
+The strengths of our meta-analysis included its wide ranging systematic search strategy; screening of more studies than in
+previous quantitative meta-analyses on this topic; thorough check for duplicate data; and the comprehensive quality assessment of the primary studies. Our approach was intentionally conservative, focusing on studies with a low to moderate risk of bias and only with estimates related to total suicides in the population. This study also looks at the research on the effects of general reporting of suicide on subsequent numbers of suicides, although only five studies were available for this analysis.
+Limitations included our inability to test causality because of the before-and-after and interrupted time series designs of the original studies, high levels of heterogeneity that could not be fully accounted for, and possible publication bias. Further, it was not possible to generate absolute risk estimates because the included studies mostly did not report the baseline risk of suicide in their respective settings. Despite the wide ranging search strategy, non-English language studies in the international literature might not have been indexed in the databases we searched. Our analysis covered only a proportion of suicide related media items. Studies on the effects of media items covering the spreading of novel suicide methods, such as charcoal burning in parts of Asia,50 were not included because of the low prevalence of these methods at baseline. Studies on fictional suicides were not included to avoid a further increase in heterogeneity between studies. Hence we cannot draw conclusions on these types of studies; our meta-analysis included studies with a narrower focus on interventions related to the reporting of suicide and suicidal behaviour. Finally, the only outcome considered in this meta-analysis was suicide. Although this outcome is of highest relevance to suicide prevention, media reports can impact on other domains as well, including help seeking behaviour and stigmatisation that have not been looked at in this meta-analysis.7
+Conclusions
+In this large and up-to-date systematic review and meta-analysis, we looked at the impact of suicide reports in news and information media on subsequent numbers of suicides. Our results support the continued use and promotion of guidelines on responsible media reporting of suicide, which are the best available interventions to address and prevent imitation effects in the population.15 16 Collaboration between suicide prevention experts and media professionals in implementing these guidelines is an essential part of any suicide prevention strategy. Caution should be exercised in reporting suicides by celebrities in particular. The media will continue to report on newsworthy suicides but have a social responsibility to mitigate the likelihood of the Werther effect.

Association of school absence and exclusion with recorded neurodevelopmental disorders, mental disorders, or self-har.txt

@@ -0,0 +1,118 @@
+Introduction
+Poor school attendance due to absence (authorised or unauthorised) from available sessions or exclusion (where a headteacher forbids a student to attend for a fixed number of sessions or permanently) leads to multiple immediate and long-term socioeconomic disadvantages. It is associated with a range of negative outcomes across the life course, including poor educational
+attainment, unemployment, and poverty.1-5 Several smallscale studies in the UK, USA, and Australia, with sample sizes ranging from less than 100 to 13 000, suggest that absence from school is more common in children with a mental disorder, specifically depression, anxiety, and disruptive behaviour disorders, through school refusal, truancy, or the condition itself.6-11 Studies from the UK12,13 and the USA14 report an association between
+Research in context
+Evidence before this study
+We searched PubMed for papers published in English between database inception and July, 20, 2021, using the search terms ((children) OR (adolescents)) AND ((school attendance) OR (school absence) OR (exclu*) OR (truan*) or (school disengagement) OR (School Refusal)) AND ((depression) or (anxiety) or (adhd) or (autism) or (learning difficulty) or (schizophrenia) or (bipolar) or (self-harm) or (eating disorder) or (drugs) or (alcohol) or (conduct disorder)). We found 13 small-scale cross-sectional surveys that used questionnaires to assess mental disorders and one national electronic cohort study linking education and secondary healthcare datasets. School absence and exclusion were found to be associated with neurodevelopmental disorders, depression, anxiety, disruptive behaviour, substance misuse, or self-harm, but current evidence is sparse and based on small numbers.
+Added value of this study
+Our population-based, electronic cohort study was larger than most previous studies, including more than 400 000 pupils, and linked routinely collected primary and secondary healthcare data to educational data. Previous studies have been based on secondary care data only and probably missed disorders,
+such as anxiety, that are more commonly managed in primary care. Our study encompasses a wide range of clinically diagnosed and recorded mental and neurodevelopmental disorders up to the age of 24 years, and so includes conditions, such as bipolar disorder and schizophrenia, that are less frequently studied in this context are are more often diagnosed in late adolescence and early adulthood. Furthermore, the large size of this study allows for the inclusion of people with less common diagnoses, such as eating disorders. We found strong associations across all disorders and self-harm with absenteeism and exclusion from school. Odds ratios for both outcomes increased with the number of comorbidities and deprivation.
+Implications of all the available evidence
+Poor attendance affects the educational attainment of children and future social and developmental outcomes. Children with mental or neurodevelopmental disorders or who self-harm are more likely to miss school through absenteeism and exclusion than their peers. Exclusion or persistent absence are potential indicators for current or future poor mental health that are routinely collected and could be used to target assessment and early intervention.
+neurodevelopmental disorders (ie, ADHD and autism spectrum disorder [ASD]) and self-harm with persistent absenteeism. Similarly, school exclusion appears to be strongly associated with ADHD, ASD, and mental disorders, particularly depression, in UK-based and international studies.13,15 In these mostly cross-sectional studies, diagnoses were assessed by use of questionnaires or interviews. However, children and young people (<24 years) with these disorders are more commonly from disadvantaged families and might be less likely to participate in research surveys.16-18 They also have higher levels of attrition at follow-up16-18 for reasons including impairments affecting the young person or their parent and impacting survey completion or a related absence when surveys are done in a school setting. Furthermore, birth cohort studies often include insufficient numbers of children with mental health conditions to support indepth analysis of rarer conditions.
+For the sail Databank see	In this study, we capitalised on electronic linkage
+https://saildatabank.com/ between routinely collected primary and secondary health-care data on clinical diagnoses and data on school attendance and exclusions at a population level. Our hypothesis was that school absences and exclusions are associated with a broad range of diagnosed and recorded neurodevelopmental and mental disorders and self-harm by 24 years of age within our cohort of pupils, even after adjusting for sex, age at the start of the academic year, and deprivation. Once established (and previous literature is scarce), this hypothesis would lead to further questions for more detailed study.
+Methods
+Study design and participants
+In this nationwide, retrospective, electronic cohort study, we drew our cohort from the 5 341 392 individuals in the Welsh Demographic Service Dataset to include individuals aged 7-16 years (16 years being the school leaving age in the UK) enrolled in state-funded schools in Wales in the academic years 2012/13-2015/16 (between Sept 1, 2012, and Aug 31, 2016) who had primary and secondary care linked data and no conflicting data in the education dataset that pointed to a many-to-one correspondence between the anonymised linkage field and the internal pupil identification number. Ethics approval was granted from the Secure Anonymised Information Linkage (SAIL) Information Governance Review Panel, an independent body consisting of a range of government, regulatory, and professional agencies, in line with ethical permissions already granted to the analysis of data in the SAIL Databank (approval number 0808).
+Procedures
+We linked data on an individual level via the Adolescent Mental Health Data Platform, an international data platform that supports mental health research in children and young people. For our study, the Adolescent Mental Health Data Platform used datasets from the SAIL Databank, a repository of routinely collected health and education datasets for the population of Wales.19,20 All data are treated in accordance with the Data Protection Act 2018. Individuals within the datasets are assigned a
+unique anonymised linkage field that replaces any identifiable information, such as names, and enables anonymised linkage across the different datasets.
+The datasets in the SAIL Databank that we used were: the Welsh Demographic Service Dataset (a demographics register of people registered with general practitioner [GP] practices in Wales) on Nov 1, 2018; the Office for National Statistics deaths register on March 28, 2019; the Welsh Index of Multiple Deprivation 2011 (an official measure of small area [defined as containing approximately 1500 individuals] deprivation in Wales, based on employment opportunities, income, education, health, community safety, geographical access to services, housing, and the physical environment; quintile 5 represents the most deprived areas) on Nov 1, 2018; the Welsh Longitudinal General Practice Database (on Aug 20, 2018) and the Patient Episode Database for Wales (on Jan 31, 2019), which contain attendance and clinical information for all GP interactions and hospital inpatient and day case activity in Wales, respectively; and the Welsh Government Education Dataset (appendix 2 p 6). The Welsh Government Education Dataset includes records for all children registered at mainstream state schools in Wales or educated in settings other than school. It contains information on attendance, exclusions, eligibility for free school meals, and receipt of a statement of special educational needs (SEN). Attendance records were available from the academic year 2007-08 to the academic year 2015-16. Each school reported, per pupil, the number of authorised and unauthorised absences for that year out of a total number of possible sessions per year. Exclusion records (categorised as permanent, fixed, or lunchtime) were available from the academic year 2012-13 to the academic year 2015-16. A child might have SEN status if they have a learning difficulty or disability (including neurodevelopmental or mental disorders) that requires special education provisions to be made for them.21
+We queried primary and secondary care datasets to extract recorded neurodevelopmental and mental disorders and self-harm using code lists from the ICD (version 10) for secondary care and read codes (version 2)22 in primary care. The codes were collated from published articles and code lists or were compiled in collaboration with clinicians (appendix 2 p 7). Neurodevelopmental disorders (ie, ASD and ADHD), learning difficulties, and conduct disorder were extracted for our cohort of pupils from their birth until they reached 24 years of age because these conditions often arise early in development and are diagnosed at a young age. Other mental disorders (including depression, anxiety, eating disorders, bipolar disorder, schizophrenia, alcohol misuse, and drugs misuse) or self-harm were extracted for our cohort of pupils between the ages of 10 years and 24 years. We categorised all F ICD-10 codes and E read codes not included in other category code lists, such as those for mania, into the other psychotic disorders category. Each pupil had a flag per each disorder categorised as a binary
+variable (recorded present or absent). The age at first diagnosis was extracted for each pupil and disorder. Where a pupil could not be linked to primary or secondary care datasets, this was flagged as linked or unlinked.
+We extracted SEN status for each pupil as a binary variable (present vs absent) to understand the extent to which it, in addition to a disorder, affected outcomes. We counted the number of morbidities per person to assess the effect of comorbidities (defined as two or more of the studied disorders recorded for the same individual, not necessarily concurrently).
+Outcomes
+We defined absenteeism as a binary variable, categorised as 1 when a pupil missed more than 10% of sessions in 1 year and categorised as 0 otherwise. The choice of 10% was based on a report from Estyn (the quality inspectorate of education in Wales), which described that, of pupils who were absent for more than 10% of sessions, fewer than 80% achieved the level expected of them by age 11 years in mathematics, science, and either English or Welsh as a first language and fewer than 40% achieved level 2 (equivalent to five GCSEs at grades A*-C) at 16 years of age.23,24 10% is the level used in England to define persistent absenteeism,25 although the level used in Wales is 20%.26 Exclusion (a record of any type of exclusion in a specific academic year) was also categorised as a binary variable (yes vs no).
+Statistical analysis
+Data were retrieved from the SAIL Databank by use of IBM DB2 9.7 SQL. Statistical analysis was done by use of R (version 3.3.3), accessed through RStudio (version 1.2). We analysed the association between the outcome variables (absenteeism and exclusion) and the existence of a record of each neurodevelopmental or mental disorder and self-harm, up to 24 years of age, using generalised estimating equations (R library geepack).27 Generalised estimating equations with exchangeable correlation structures using binomial distribution with the logit link function were used to calculate odds ratios (OR) for absenteeism and exclusion, adjusting for sex, age at the beginning of the academic year, and deprivation. We used a long data format with one row per pupil and year. The experimental unit (id) was the pupil, and the repeated measurements were ordered by academic year (wave). 95% CIs for proportions and percentages were estimated by the Wilson score method with continuity correction. We did not explore causality; therefore, the time dependence of measurements per person (up to four measurements at different academic years) was modelled with a correlation function as a source of variance, which was marginalised over so that the variance of the estimated covariates were calculated efficiently.28 We used an exchangeable correlation structure, in which any two measurements for the same pupil had the same correlation. We analysed each
+recorded disorder separately using a sub-cohort consisting of those presenting with these disorders together with pupils in our cohort with no record of any of these disorders (our controls). We tested multiple models, sequentially adding age (week of birth), sex (male vs female), and deprivation quintile (quintile 5 representing the most deprived) as covariates.29,30 We tested the goodness of fit for each of these models by calculating the Quasi-likelihood model information criteria.31 We stratified the population by condition and analysed the association between the outcome variables and sex, age, and deprivation separately. For the main analysis, we pooled pupils with ADHD and ASD (under neurodevelopmental disorder) and pooled pupils with depression, anxiety, eating disorders, schizophrenia, bipolar disorder, and other psychotic disorders (under any mental disorder).
+In the first sensitivity analysis, we ran the main analysis on both the subgroup that had linked healthcare data and the full dataset. Neurodevelopmental and mental disorders typically show comorbidity over the life course. We assessed the sensitivity of our analysis to comorbidities by conducting three extra sensitivity analyses. First, we compared the model results between those with one morbidity and those without any of the morbidities studied. Second, we compared the model results between those with more than one morbidity and those without any of the morbidities studied. Finally, separately, we ran the model with the number of comorbidities as a covariate.
+We conducted several other sensitivity analyses. For some children, certain types of mental disorders are an entry point for SEN status within the education system. We assessed sensitivity to SEN status by exploring the interaction between any of the disorders studied and SEN status. We also ran our models in the subpopulation of participants first diagnosed or with their first record
+before 17 years of age (ie, while of school age). Furthermore, we compared individual-level yearly rates of absences (number of sessions absent/total number of possible sessions per year) for those excluded versus for those not excluded and calculated the Pearson correlation coefficient between individual-level yearly rates of absences and individual-level yearly rates of exclusions.
+Role of the funding source
+The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.
+Results
+437 412 individuals had education and demographic data and were aged 7-16 years during the 2013-16 academic years, of whom 213 816 (48-9%) were female and 223 596 (51-1%) were male (figure 1). Of these 437 412 individuals, 22 775 (5-2%) had no linked hospital or primary care records. We considered health data for these individuals as missing at random because missingness was not based on health or education status.
+In the group with health-care data, 212 848 (51-3%) of 414637 pupils were boys and 201789 (48-7%) were girls (appendix 2 p 1). In the group without health-care data, 10 748 (47-2%) of 22 775 pupils were boys and 12 027 (52-8%) were girls (appendix 2 p 1). Compared with those with health-care data, a higher proportion of individuals with missing health-care data resided in quintile 2 areas on the Welsh Index of Multiple Deprivation and a lower proportion resided in quintile 5 areas (appendix 2 p 1). We repeated the main analysis in the group with linked primary and secondary health-care data (n=414637) and in the larger group (n=437 412). Results were equivalent (appendix 2 pp 25-26) so we removed those without linked health-care data from all main and other sensitivity analyses (list-wise deletion). Each pupil contributed 1-4 years of data. The distribution of morbidity, sex, and deprivation by number of years of data contributed is shown in appendix 2 (pp 2-4). Demographics and morbidity were not correlated with the number of years of data contributed so we also viewed this as missing at random.
+Of the 414 637 pupils with primary and secondary care data comprising our study population, 201 789 (48-7%) were female and 212 848 (51-3%) were male. Their mean age on Sept 1, 2012, was 10-5 years (SD 3-8). Ethnicity data were not available. 57 930 (14-0%) pupils had at least one of the disorders studied (a neuro-developmental or mental disorder or a record of selfharm) by the age of 24 years, and 42 734 (10 -3%) while of school age. 356 707 (86-0%) of414 637 individuals had no record of self-harm or any of the disorders studied. The numbers of diagnosed individuals, age at first diagnosis, and SEN status before 17 years of age are detailed in appendix 2 (p 7).
+118 140 (28-5%) had recorded absenteeism during at least 1 school year, of whom 5901 (5-0%) had a neuro-developmental disorder, 17 724 (15-0%) had a mental disorder, and 5164 (4-4%) had a record of self-harm. 20 507 (17-4%) of the 118 140 with recorded absenteeism were diagnosed while of school age, ofwhom 5762 (28 -1%)
+had a neurodevelopmental disorder, 12 164 (59 -3%) had a mental disorder, and 4450 (21-7%) had a record of selfharm.
+The proportion of absentee pupils with no record of any of the disorders studied remained stable in primary school (7-11-year-olds) at around 12-5% and increased
+in secondary school (11-16-year-olds) to around 18% for 16-year-olds (figure 2). For the raw counts used to create figure 2, please see the appendix (pp 16-20). Across all ages, a higher proportion of pupils with a neuro-developmental disorder, mental disorder, or self-harm record were absent from school compared with pupils without a record (figure 2). In the last 2 years of primary school (10-11-year-olds), pupils with a subsequent diagnosis of schizophrenia or drugs misuse had the highest rate of absenteeism at around 30-33% (figure 2). In the last 2 years of secondary school (ages 15-16 years), pupils with a record of bipolar disorder, schizophrenia, alcohol misuse, drugs misuse, or self-harm had the highest rate of absenteeism at around 40-55% (figure 2).
+Goodness-of-fit tests (appendix 2 p 27) showed that including sex, age, and deprivation as covariates sequentially improved the fit, so we present both unadjusted and adjusted results (table 1). Having a record of a neurodevelopmental disorder (OR 2-1, 95% CI 2-0-2-2), mental disorder (2-9, 2-8-2-9), or self-harm (4-0, 3-8-4-1) was associated with absenteeism (table 1). Adjusted ORs (aORs) ranged from 2-0 (95% CI 1-9-2-0) for pupils with a neurodevelopmental disorder to 4-2 (3-4-5-3) for those with schizophrenia and 5-5 (4-2-7-2) for those with bipolar disorder (table 1).
+Of those with a record of neurodevelopmental disorders, learning difficulties, conduct disorder, depression, other psychotic disorders, or drugs or alcohol misuse, boys were less likely to be absent than were girls (appendix 2 p 9). For those with a record of anxiety, eating disorders, bipolar disorder, schizophrenia, or self-harm, sex was not significantly associated with absenteeism (appendix 2 p 9). For pupils with a record of neurodevelopmental disorders, conduct disorder, depression, anxiety, eating disorders, drugs or alcohol misuse, or self-harm, age was associated with absenteeism, with slight increases in ORs per year (appendix 2 p 9). The sample sizes for bipolar disorder and schizophrenia were too small to assess the association of deprivation quintile with absenteeism; however, the odds of being absent increased with increased deprivation (5th vs 1st quintile) for all other variables apart from other psychotic disorders, ranging from 1-5 (95% CI 1-3-1-9) for conduct disorder to 2-8 (2-2-3-6) for alcohol misuse (appendix 2 p 9).
+15 199 (3-7%) of 414637 pupils had been excluded from school at least once, 243 (0-1%) of whom were excluded permanently. 1979 (13 -0%) of 15 199 had a neurodevelopmental disorder, 3161 (20-8%) had a mental disorder, and 1518 (10 -0%) had a record of self-harm. 4568 (30 -1%) were diagnosed while of school age, of whom 1925 (42 -1%) had a neurodevelopmental disorder, 2048 (44-8%) had a mental disorder, and 1291 (28-3%) had a record of self-harm. Children aged 7-11 years with no record of the studied diagnoses or self-harm were very unlikely to be excluded (1174 [0-5%] of 233 191). They were more likely to be excluded if they had a record of ASD (211 [4-7%] of4464) or conduct disorder (193 [8 -0%] of 2415). Exclusions generally
+became more common among older children (figure 3). For those with no disorder or self-harm, the proportion of exclusions increased to 2-9% (2770 of95 977) among those aged 15 years, before decreasing to 2-2% (2064 of94 172) in the last year of secondary school (age 16 years). Notable increases in exclusion rates were seen among pupils aged 14 years with ADHD (374 [15-1%] of 2483), conduct disorder (207 [14-5%] of 1433), drugs misuse (205 [24-2%] of 848), alcohol misuse (150 [14-6%] of 1026), and selfharm (443 [10-7%] 4135), although exclusion rates tended to decrease in the final year of secondary school (figure 3). The proportion of pupils with severe mental illness who were excluded was also high, with 16 (17-4%) of 92 with bipolar disorder excluded at age 15 years and 16 (18-4%) of 87 with schizophrenia excluded at age 14 years (figure 3). For the raw counts used to create figure 3, please see the appendix (pp 21-24).
+Goodness-of-fit tests (appendix 2 p 27) again showed that including sex, age, and deprivation as covariates sequentially improved model fit. Having a neuro-developmental disorder, a mental disorder, or a record of self-harm were all associated with being excluded from school (table 2). After adjusting for sex, age, and deprivation, pupils with a record of drugs misuse had the highest odds of being excluded (table 2). To note, alcohol misuse, self-harm, schizophrenia, and bipolar disorder also had high aORs (table 2).
+Across disorders, apart from bipolar disorder, boys were significantly more likely to be excluded than were girls (appendix 2 p 13). Boys with a record oflearning difficulties, anxiety, eating disorders, schizophrenia, other psychotic disorders, or self-harm had an OR for being excluded between 2 and 3 (appendix 2 p 13). Being older was associated with a higher odds of exclusion for individuals with a record of most variables studied (OR range 1-09-1-19), except for bipolar disorder, schizophrenia, other psychotic disorders, drugs misuse, and alcohol misuse (appendix 2 p 13). The sample sizes for bipolar disorder and schizophrenia were too small to assess the association of deprivation quintile with exclusion; however, the odds of exclusion were higher in the most deprived areas than in the least deprived areas for all variables apart from other psychotic disorders, with the OR varying from 1-4 (95% CI 1-1-1-9) for those with conduct disorder to 3-3 (2-7-4-0) for those with anxiety (appendix 2 p 13).
+Pupils in our cohort had up to eight morbidities in total. 41 018 had one morbidity, 12 096 had two, 3495 had three, and 1321 had four or more (appendix 2 p 10). Absenteeism was more likely in pupils with comorbidities than in pupils with one morbidity, except in the case of bipolar disorder (table 1). Pupils with a single diagnosis of an eating disorder, schizophrenia, or other psychotic disorder were not at higher risk of being absent compared with their healthy peers (table 1). When the number of comorbidities was modelled as a covariate, the OR of being absent was between 1-2 and 1-4 for each additional
+comorbidity, except for bipolar disorder for which the OR was 1-0 (appendix 2 p 11). SEN status did not reduce the ORs for being absent in those with anxiety, eating disorders, schizophrenia, or alcohol misuse (appendix 2 p 12). For those with ADHD, ASD, learning difficulties, conduct disorder, depression, bipolar disorder, drugs misuse, other psychotic disorders, or a record of selfharm, having SEN status reduced the OR for absenteeism to 0^59-0^89 compared with not having SEN status (appendix 2 p 12). The results for pupils with a record before 17 years of age were similar to those of the main cohort, except for pupils with a record of alcohol or drugs misuse, bipolar disorder, or depression who had slightly higher odds of being absent, and for pupils with schizophrenia who had slightly lower odds (table 1).
+In the group of pupils with more than one morbidity, aORs for being excluded were consistently higher than for those with one morbidity (table 2). When the number of comorbidities was modelled as a covariate, the OR of being excluded was between 1-2 and 1-8 per each additional comorbidity (appendix 2 p 14). SEN status was associated with decreasing ORs for being excluded for those with neurodevelopmental disorders, conduct disorder, depression, bipolar disorder, other psychotic disorders, drugs misuse, and self-harm (appendix 2 p 15). aORs for being excluded differed between the subpopulation diagnosed while at school and the main
+cohort, depending on the variable (table 2). Of note, there was little difference for neurodevelopmental disorders, the aOR for exclusion was lower for people with schizophrenia when diagnosed while at school, and the aORs for exclusion were somewhat higher for individuals with drugs or alcohol misuse recorded while of school age (table 2).
+The individual-level yearly absence rate (number of sessions missed/total number of possible sessions per year) was higher in the group with a record of exclusion than in the group without a record of exclusion (appendix 2 p 5). The correlation between individuallevel yearly absence rates and individual-level yearly exclusion rates was r=047 in the full cohort and r=045 for those with recorded exclusions.
+Discussion
+Our study, which involved more than 400 000 pupils, highlights that children and young people diagnosed with a neurodevelopmental disorder or mental disorder, or who have a record of self-harm, before 24 years of age are much more likely to miss school than their peers, even after adjusting for age, sex, and deprivation. Our data and study size enabled us to include disorders typically not included in studies of school-aged children, such as rare disorders and disorders that typically present after individuals have left school (eg, schizophrenia), that
+might confer antecedent clinical vulnerabilities.32 School absenteeism and exclusion rates were higher after 11 years of age for all children but disproportionally more so in those with a record of a disorder or self-harm, even if it was recorded during school age. This finding could reflect a reduced direct influence of parents on older
+children’s attendance or the smaller size of primary schools compared with secondary schools. Generally, individuals with more than one recorded morbidity were more likely to be absent or excluded than were those with only one morbidity, which was exacerbated with each additional disorder. Within the diagnosed populations,
+girls with neurodevelopmental disorders, learning difficulties, conduct disorder, depression, other psychotic disorders, or drugs or alcohol misuse were more likely to be absent than were boys, and boys were more likely to be excluded than were girls across all studied disorders apart from bipolar disorder. This finding aligns with the view that boys externalise mental distress through their behaviour, which in turn impacts the school environment and results in their exclusion, whereas girls, and especially those with emotional disorders or delayed diagnosis of neurodevelopmental disorders, tend to be more anxious and withdraw from social contact.32 Age was found to be associated with both outcomes in relation to most disorders. We also found associations between both outcomes and deprivation within most disorders studied. Having SEN status reduced the likelihood of being absent or excluded, most notably for
+those with records of neurodevelopmental disorders or bipolar disorder, compared with those with a record but no SEN status, potentially highlighting the positive impact of recognition, diagnosis, and educational interventions.
+Our findings strengthen those found previously in much smaller population-based studies. In the ALSPAC study29 of a UK birth cohort, by 8 years of age, 19% of children with ADHD and 31% of those with conduct disorder were excluded from school compared with 1-9% and 2-8% of those without ADHD or conduct disorder, respectively. In another study ofa UK cohort (BCAMHS),30 psychiatric symptoms (assessed through validated questionnaires) were a significant predictor of exclusions.
+Our study is based on routinely collected data encompassing a wide range of clinically diagnosed and recorded disorders. It benefits from well documented,
+often validated, and curated lists of ICD-10 and read (version 2) codes to ascertain each of the disorders. Arguably, diagnoses made by clinicians for those in contact with services provide more complete case ascertainment than do surveys or cohort studies, which are susceptible to selection bias due to low recruitment and high attrition in populations with psychiatric disorders. However, a common feature of all database studies of routinely collected data is the underestimation of the number of disorders in the population as not all those affected consult their GP, or conditions might not be recognised or recorded.33 Additionally, there is no validated measure of the clinical problems recorded, which prevents any estimation of severity, and administrative data are vulnerable to random errors in data entry.
+This study’s novelty lies in its linkage of education, health (including primary care), and deprivation datasets for a whole population (Wales) at an individual pupil level over 4 school years for a wide range of disorders. Linking health and education data on this scale allows us to gain valuable insights on the education of children with neurodevelopmental disorders, mental health disorders, or self-harm. Because many older adolescents with common mental disorders are managed in primary care, it is important to include this data source. A whole population dataset enabled us to include pupils with rarer conditions such as schizophrenia and bipolar disorder. Linking diagnoses up to the age of 24 years allowed for assessment of conditions more frequently diagnosed after school leaving age (eg, schizophrenia), for which their antecedents or premorbid presentation, such as cognitive or social deficits, apathy, or selfmedication with drugs, might affect attendance and exclusion. We did not take physical comorbidities into account, although we note the strong association between poor mental and physical health,34 because some absences would have been due to physical morbidity and medication rather than the mental or neurodevelopmental disorder, which would have complicated the interpretation of our findings.
+Our estimates might underestimate the effect of mental health difficulties on exclusions and absenteeism. Younger children will have had less time for evidence of their diagnosis to be recorded, especially for those conditions that tend to appear later in adolescence, and some young people who are diagnosable will not present to services. There is some evidence30 to suggest that, for each diagnosed child, there could be a number that have multiple symptoms but do not meet the criteria for a diagnosis. These children might well have issues at school that could lead to poor attendance or exclusion. Some children, especially those with ADHD, ASD, or learning difficulties, might not have been included in our dataset because they are in schools for children with special educational or behavioural needs or are homeschooled. Different pupils contributed different numbers of years to the analysis. We are satisfied that there were
+no demographic or mental health-related differences between these pupils. 5 -2% of pupils with education data in the 2013-16 academic years did not have any linked health data and were removed from our analyses.
+There are various processes through which school attendance might be associated with neurodevelopmental disorders, mental disorders, and self-harm. These processes include disruptive behaviours resulting in exclusion, physical comorbidities or somatic symptoms (eg, stomach pain and headaches) leading to authorised absence, symptoms associated with anxiety and depression leading to school refusal, family problems, and peer problems (eg, bullying). If absence from school results in social isolation and poorer academic performance, it could exacerbate mental health and attendance issues if the cycle is not disrupted. Our study cannot infer causal relationships and further research should focus on the direction of the association, which could be bidirectional for individual disorders and outcomes. Clinical record data might not be ideal to use in these future studies because the documentation in clinical records will not represent an accurate measure of the time of first onset of the symptoms or disorder. However, even without an understanding of the direction or mechanisms of the association, the demonstration of an association using real life outcomes and data is important.
+Poor school attendance affects the educational attainment of children and future social, developmental, employment, and physical health outcomes. Many governments, including UK Governments, have recognised the importance of regular school attendance and have issued related guidelines, which include penalty notices for the carers or parents of persistently absent children and the use of incentives to encourage high attendance.35 Exclusions from schools in England and Wales are intended to be used in serious breaches of behaviour policies—for violence, sexual abuse, the supply of illegal drugs, or the use of weapons.36,37 Currently, rates of exclusion in England are rising, raising concerns about school-based policies to improve behaviour and support teachers. Similar initiatives are in place in the USA and elsewhere.38
+Linking routinely collected health and education data has the potential to improve services for children39 by identifying those in need, alongside gaps in provision. Our analysis clearly shows that children with neuro-developmental disorders, mental disorders, and selfharm spend less time at school. As such, exclusion or persistent absence is a potential indicator for current or future poor mental health that is routinely collected by schools and local education authorities and could be used to target assessment and early intervention.40 There is growing interest in school-based prevention and early intervention programmes that focus on improving the school climate for reducing adolescent mental health problems,41,42 which has relevance now as children return to school following closures and blended learning in
+response to the COVID-19 pandemic. Other interventions have included psychological interventions that focus primarily on anxiety and depression symptoms.43 Schoolbased mental health provision and integration with mental health services has been highlighted as a major strategic priority in the UK.44 This approach could benefit young people, as supported by our finding that having a SEN status decreases the odds of being absent or excluded, even if it does not remove the risk completely. Attendance and exclusion data, which are already collected by schools, could provide useful information about where to focus sparse resources. School-based mental health prevention strategies might also help to build resilience, enabling pupils to develop strategies for managing and improving their mental health and wellbeing, and to understand when and how to seek additional help.
+Future research could further explore whether improvements in school attendance over time serve to reduce the incidence of mental disorders and whether the timing of diagnosis is an important factor in the risk for absenteeism or exclusions. This can be done by looking at causal relationships between mental health and school outcomes using a longer follow-up period. Other avenues for research include evaluating the effect of physical comorbidities on school outcomes and the differential associations of pairs of disorders with school outcomes.
+To conclude, people up to 24 years of age who have mental or neurodevelopmental disorders or self-harm have poorer attendance at school than their peers who do not have disorders or self-harm. Exclusion or persistent absence is a potential indicator of current or future poor mental health that is routinely collected by schools and local education authorities and could be used to target assessment and early intervention.
+Articles
+16	Frojd SA, Kaltiala-Heino R, Marttunen MJ. Does problem behaviour 31 affect attrition from a cohort study on adolescent mental health?
+Eur J Public Health 2011; 21: 306-10.	32
+17	Wolke D, Waylen A, Samara M, et al. Selective drop-out in longitudinal studies and non-biased prediction of behaviour	33
+disorders. Br J Psychiatry 2009; 195: 249-56.
+18	Saiepour N, Ware R, Najman J, Baker P, Clavarino A, Williams G.
+Do participants with different patterns of loss to follow-up have	34
+different characteristics? A multi-wave longitudinal study.
+J Epidemiol 2016; 26: 45 49.
+19	Ford DV, Jones KH, Verplancke JP, et al. The SAIL Databank: building a national architecture for e-health research and	35
+evaluation. BMC Health Serv Res 2009; 9: 157.
+20	Lyons RA, Jones KH, John G, et al. The SAIL databank: linking multiple health and social care datasets. BMC Med Inform Decis Mak 36 2009; 9: 3.
+21	Sadler K, Vizard T, Ford T, et al. Mental health of children and young people in England, 2017. Nov 22, 2018.https://digital.nhs.uk/ data-and-information/publications/statistical/mental-health-of-children-and-young-people-in-england/2017/2017 (accessed	37
+Nov 12, 2021).
+22	NHS Digital. Read codes. 2018. https://digital.nhs.uk/services/ terminology-and-classifications/read-codes (accessed Nov 2, 2021).	38
+23	Estyn. Effective practice in improving attendance in primary
+schools—June 2015. June 12, 2015. https://www.estyn.gov.wales/ thematic-report/effective-practice-improving-attendance-primary-schools-june-2015 (accessed Nov 2, 2021).	39
+24	Estyn. Attendance in secondary schools—September 2014.
+Sept 1, 2014. https://www.estyn.gov.wales/thematic-report/ attendance-secondary-schools-september-2014 (accessed	40
+Nov 2, 2021).
+25	Department of Education. A guide to absence statistics.
+March, 2019. https://assets.publishing.service.gov.uk/government/ 41 uploads/system/uploads/attachment_data/file/787314/Guide_to_ absence_statistics_21032019.pdf (accessed Nov 2, 2021).
+26	Welsh Government. Absenteeism from secondary schools, 2018/19.
+Aug 29, 2019. https://gov.wales/sites/default/files/statistics-and- 42 research/2019-08/absenteeism-from-secondary-schools-september-2018-august-2019-318.pdf (accessed Nov 2, 2021).
+27	Hojsgaard S, Halekoh U, Yan J. The R package geepack for	43
+generalized estimating equations. J Stat Softw 2006; 15: 1-11.
+28	Liang KY, Zeger SL. Regression analysis for correlated data.
+Annu Rev Public Health 1993; 14: 43-68.
+29	Paget A, Parker C, Heron J, et al. Which children and young people 44 are excluded from school? Findings from a large British birth cohort study, the Avon Longitudinal Study of Parents and Children (ALSPAC). Child Care Health Dev 2018; 44: 285-96.
+30	Ford T, Parker C, Salim J, Goodman R, Logan S, Henley W.
+The relationship between exclusion from school and mental health: a secondary analysis of the British Child and Adolescent Mental Health Surveys 2004 and 2007 Psychol Med 2018; 48: 629-41.
+Pan W. Akaike’s information criterion in generalized estimating equations. Biometrics 2001; 57: 120-25.
+Pine DS, Fox NA. Childhood antecedents and risk for adult mental disorders. Annu Rev Pyschol 2015; 66: 459-85.
+John A, Marchant AL, Fone DL, et al. Recent trends in primary-care antidepressant prescribing to children and young people: an e-cohort study. Psychol Med 2016; 46: 3315-27.
+van der Lee JH, Mokkink LB, Grootenhuis MA, Heymans HS, Offringa M. Definitions and measurement of chronic health conditions in childhood: a systematic review. JAMA 2007;
+297: 2741-51.
+Welsh Assembly Government. Strategies for schools to improve attendance and manage lateness. 2011. https://dera.ioe.ac. uk/2945/3/110308section3en.pdf (accessed Nov 2, 2021).
+Department for Education. Exclusion from maintained schools, academies and pupil referral units in England. September, 2017. https://assets.publishing.service.gov.uk/government/uploads/ system/uploads/attachment_data/file/641418/20170831_Exclusion_ Stat_guidance_Web_version.pdf (accessed Nov 2, 2021).
+Welsh Government. Exclusions from schools and pupil referral units (PRU). April 1, 2015. https://gov.wales/exclusion-schools-and-pupil-referral-units-pru (accessed Nov 2, 2021).
+US Department of Education. Key policy letters signed by the Education Secretary or Deputy Secretary. Oct 7, 2015. https://www2. ed.gov/policy/elsec/guid/secletter/151007.html (accessed Nov 2, 2021).
+Downs J, Gilbert R, Hayes RD, Hotopf M, Ford T. Linking health and education data to plan and evaluate services for children. Arch Dis Child 2017; 102: 599-602.
+Kearney CA, Graczyk P. A response to intervention model to promote school attendance and decrease school absenteeism. Child Youth Care Forum 2014; 43: 1-25.
+Shinde S, Weiss HA, Varghese B, et al. Promoting school climate and health outcomes with the SEHER multi-component secondary school intervention in Bihar, India: a cluster-randomised controlled trial. Lancet 2018; 392: 2465-77.
+Bonell C, Blakemore S-J, Fletcher A, Patton G. Role theory of schools and adolescent health. Lancet Child Adolesc Health 2019; 3: 742-48.
+Fleming T, Dixon R, Frampton C, Merry S. A pragmatic randomized controlled trial of computerized CBT (SPARX) for symptoms of depression among adolescents excluded from mainstream education. Behav Cogn Psychother 2012; 40: 529-41. Department of Health & Social Care, Department for Education. Transforming children and young people’s mental health provision: a green paper. December, 2017. https://assets.publishing.service.gov. uk/government/uploads/system/uploads/attachment_data/ file/664855/Transforming_children_and_young_people_s_mental_ health_provision.pdf (accessed Nov 2, 2021).
+34
+www.thelancet.com/psychiatry Vol 9 January 2022

Association-of-Logics-hip-hop-song-18002738255-with-Lifeline-calls-and-suicides-in-the-United-States-Interrupted-time-series-analysisThe-BMJ.txt

@@ -0,0 +1,71 @@
+WHAT IS ALREADY KNOWN ON THIS TOPIC
+The increase in suicides after media stories about suicides by celebrities is referred to as the Werther effect
+Much less is known about the protective effects of media stories of hope and recovery in the context of suicidal crises
+Some evidence from randomised controlled trials shows a beneficial effect of media narratives of hope and recovery on suicidal thoughts and help seeking intentions
+WHAT THIS STUDY ADDS
+During 34 days of wide scale public exposure to Logic’s song “1-800-273-8255,” Lifeline received 9915 excess calls (95% confidence interval 6594 to 13 236):
+6.9% over the expected number
+In the same period, 245 fewer suicides (95% confidence interval 36 to 453) occurred: 5.5% below the expected number
+A media event intended to tell a “suicide prevention story” was associated with both an increase in calls to the National Suicide Prevention Lifeline and a simultaneous reduction in suicides in the United States
+2018), Lifeline received an excess of 9915 calls (95% confidence interval 6594 to 13 236), an increase of 6.9% (95% confidence interval 4.6% to 9.2%, P<0.001) over the expected number. A corresponding model for suicides indicated a reduction over the same period of 245 suicides (95% confidence interval 36 to 453) or 5.5% (95% confidence interval 0.8% to 10.1%, P=0.02) below the expected number of suicides.
+CONCLUSIONS
+Logic’s song “1-800-273-8255” was associated with a large increase in calls to Lifeline. A reduction in suicides was observed in the periods with the most social media discourse about the song.
+Introduction
+Repetitive reporting on suicide deaths or potentially lethal actions has been shown to trigger further suicides, known as the Werther effect.1 A recent meta-analysis found that news reporting on celebrity suicides—often highly repetitive over the following weeks2—was associated with a 13% increase in suicides.1
+Some other suicide related narratives might have preventive effects—media stories of people who managed to cope with suicidal crises without dying by suicide have been associated with reductions in subsequent suicides.3 The possible protective effects of stories of hope and recovery from suicidal crises is referred to as the Papageno effect.3 In contrast with studies on the Werther effect, most studies on the Papageno effect have used experimental designs. These trials typically use suicidal thoughts rather than suicide death as the outcome. Consistent with the research evidence for the Papageno effect, some of these studies indicate that media narratives of hope and recovery from a suicidal crisis are associated with reduced suicidal thoughts, particularly in people with some risk factors for suicidal behaviour.4-6 A noted limitation of these studies is that findings about suicidal thoughts do not necessarily generalise to suicidal behaviours and, most importantly, suicides.
+Suicide prevention and education efforts must harness positive media to educate the general public and high risk groups about suicide prevention without doing harm to individuals at risk. But a major dilemma for research in this area has been that stories of hope and recovery receive much less media coverage than stories of suicide death.
+On 28 April 2017, the American hip hop artist Logic released his song “1-800-273-8255,” prominently featuring the number of the US National Suicide Prevention Lifeline (referred to as Lifeline). The narrative of the song is centred around someone calling the 1-800 number for Lifeline and then telling the counsellor that they don’t want to live anymore. The accompanying music video, which was released four months later and has since received more than 419 million views on YouTube, depicts a young black man struggling with discrimination and bullying from peers and adults for being gay. He prepares for his suicide, but ultimately takes his phone and calls Lifeline, which marks a turning point towards improvement and mastery of his crisis.7
+The release of the song in April 2017 marked the start of a series of media events promoting the story of hope and recovery featured in the song, along with the phone number of Lifeline. The song was performed at the MTV Video Music Awards in late August 2017 to 5.4 million viewers and ultimately marked a breakthrough for “1-800-273-8255.”8 The song, which was labelled a “suicide prevention anthem” by the media, entered the top 10 of the Billboard Hot 100 music charts in the US, remaining there for several weeks and ranking as high as number 3 in September 2017.9 10 The song’s release was also associated with a nearly 10% uptick in online Google searches for Lifeline in the 28 days after its release.11 By the end of 2020, the song had surpassed one billion streams on Spotify.
+Logic’s song likely represents the broadest and most sustained suicide prevention messaging directly connected to a story of hope and recovery in any location to date and is thus a serendipitous event for research. To assess whether the song was associated with help seeking or suicides, we conducted a time series analysis examining the associations between Logic’s song and daily calls to the Lifeline number as well as daily suicides in the US.
+Methods
+Public attention to Logic’s song
+Three known distinct events directed strong public attention to Logic’s song: the release of the song on 28 April 2017, Logic’s performance at the MTV Video Music Awards on 27 August 2017, and his performance at the Grammy Awards on 28 January 2018. All these events gave widespread public attention to the message of the song—that help from Lifeline is available and effective. To obtain estimates for the timespan of public attention related to each of the events as a proxy for assessment of the impact period, we retrieved all original tweets geolocated to the US that contained the search terms “Logic” and “1-800-273-8255” from Brandwatch (www.brandwatch.com). Our approach was similar to previous studies estimating exposure periods for suicide related media events.2 Brandwatch is a data reseller that stores the entire historical Twitter stream of more than 350 million tweets per day, giving us access to all public tweets, retweets, and replies across the total observation period. More than 90%
+of tweets can be successfully matched to a country of origin.
+This search allowed us to generate an exhaustive dataset with all mentions specifically related to Logic’s song, excluding tweets produced by accounts that Twitter considered malicious bots, from 1 March 2017 to 30 April 2018, covering the entire period before the release and during the song’s presence in the Billboard Hot 100.
+We visually inspected the daily time series of tweets to identify peaks in tweeting behaviour qualitatively. Consistent with our study preregistration, we were mainly interested in the three events that were most relevant to dissemination of Logic’s song (song release April 2017, MTV Video Music Awards August 2017, and Grammy Awards January 2018), but explored the time series of tweets for further peaks as well. Subsequently, we visually assessed the duration of any peak to capture the period until the day that attention wore off. In addition, we performed a post hoc change-point analysis to assess whether findings from visual inspection differed from quantitative assessment of change-points in the time series data (see supplementary text S2 for details on the methodological approach for identifying peaks and their duration).
+Lifeline calls and suicide data
+We obtained the total number of calls to Lifeline across the US directly from Lifeline. Call data were provided as daily aggregates for the period 1 January 2010 to 31 December 2018. National suicide data were obtained from the National Center for Health Statistics (part of the US Centers for Disease Control and Prevention). Suicide was defined using ICD-10 (international classification of diseases, 10th revision) underlying cause of death codes X60-X84, Y87.0, and U03. Data were provided as daily aggregates for the period 1 January 2010 to 31 December 2018.
+Statistical analysis
+Seasonal autoregressive integrated moving average models to estimate baseline trends in calls and suicides were fitted to the data up to 6 April 2017. This cut-off date was selected to allow for a three week preparatory period before the release of the song on 28 April, consistent with the observation that the first tweets about the song were posted as soon as three weeks before the release. The selection of models was aided by the SPSS Expert Modeler function, version 26 (IBM), choosing models with the lowest bayesian information criterion value, highest stationary R2 value (that is, variance attributable to the fitted time series model), and, when possible, a non-significant Ljung-Box Q statistic (indicating whether residuals could be assumed white noise, with stated degrees of freedom). The models derived from the baseline data were subsequently fitted to the full set of data for each series.
+Based on the time periods of strong social media attention on the song, we investigated the temporary
+association between each of the identified song related events and calls to Lifeline and suicides. We used dummy variables to model these associations as discrete pulses (that is, we modelled them as sudden changes from the baseline, starting and ending with the previously identified duration of the event of interest). These pulses were coded as binary variables, with a value of 0 before the onset of the event of interest, 1 during the event of interest (for 30 days, for example) and 0 thereafter. After fitting our models, we used model estimates to calculate the number of excess calls and suicides for each event (see supplementary text S1 for details of the statistical model. Supplementary table S7 provides an annotated syntax for the time series analyses ).
+As a further step, planned in our preregistration, we repeated the analyses of Lifeline calls and suicides using a single dummy variable to combine the effect of the three main media events that captured the most public attention.
+Possible confounding exogenous events
+Because of possible confounding by the release of 13 Reasons Why, a Netflix show that sparked strong criticism for violating media recommendations for safe portrayals of suicide,12-14 we included a dummy variable (coded 1 from the release date of 13 Reasons Why (31 March 2017 to 30 June 2017, and 0 otherwise).12 Notably, previous research found that the show was associated with a noticeable increase of 5.5% in suicides (95% confidence interval 5.5% to 21.1%) in the US among 10 to 19 year olds in the three months after its release.12 Our use of a three month period was consistent with social media data indicating that the show received the strongest attention in that period.12
+To identify any further events that might be associated with Lifeline calls and suicides, we used a list of Wikipedia entries of suicides by well known people between 7 April 2017 (immediately before the song’s release) and 31 December 2018 (end of observation period). We accessed and assessed tweet volumes for all the identified (American and international) celebrities to identify the suicides that received strong public attention so that we could adjust for the occurrence of these confounding events in the model (supplementary table S1). Variables were subsequently added for the suicides of Chris Cornell (18 May 2017), Kate Spade and Anthony Bourdain (5 and 8 June 2017, respectively),15 Chester Bennington (20 July 2017), and Avicii aka Tim Bergling (20 April 2018).2 Consistent with research on the association between celebrity suicides and subsequent suicide prevalence in the general population, suicides of lesser known celebrities (Chris Cornell, Tim Bergling) were coded as dummy variables, with value 0 before their deaths, 1 for the 30 days after their deaths, and 0 thereafter. For Chester Bennington, Anthony Bourdain, and Kate Spade, a 60 day period was used,15 because these suicides continued to receive considerable public attention in the second month after their deaths (supplementary table S1).
+Finally, World Suicide Prevention Day is held annually on 10 September to promote awareness of suicide prevention.16 In the US, World Suicide Prevention Day is part of the annual National Suicide Prevention Week. We included dummy variables for the seven day period of these events in 2017 (10-16 September) and 2018 (9-15 September).
+Sensitivity analyses
+We performed three (not preregistered, exploratory) sensitivity analyses. First, we used daily unique calls to Lifeline (as opposed to total calls) to assess whether patterns were similar after removing repeat callers. Second, we changed the pre-intervention period to end by the day before the song’s release (27 April 2017) to investigate whether this affected key findings. Third, we conducted an additional analysis combining all song related media events (including events that emerged only from visual inspection) into a single variable and assessed its associations with calls and suicides.
+Patient and public involvement
+No patients or members of the public were directly involved in this study because of time constraints in planning, owing to the long period between the song’s release and the setting up of this research. We did, however, speak to patients about the study and we asked a member of the public to read our manuscript after submission.
+Results
+Public attention as indicated by tweets
+Logic’s song generated 81 953 tweets by 55 471 unique users, posted between 1 March 2017 and 30 April 2018 (fig 1). Daily tweets reached three peaks corresponding to the three main events—the song’s release in April 2017, the MTV Video Music Awards in August 2017, and the Grammy Awards in January 2018. Two smaller peaks were identified; based on a qualitative assessment of a sample of specific tweets in those periods: one peak occurred around the time of the song’s video release (17 August 2017) and the second one alongside media reports of an increase in calls to Lifeline associated with the song (aired on CBS on 10 October 2017).17 All peaks emerged rapidly, reaching their maximum within one day of the event. The duration of all five peaks was estimated, using the first day of the increase as the start of the impact period and ending the day the peak had worn off.
+Figure 1 shows the estimated impact periods for each event (see supplementary figure S1 for a large version of this figure showing the time series of tweets, Lifeline calls, and suicides, and the identified impact periods).
+Attention to the song was strongest immediately after Logic’s performance at the 2017 MTV Video Music Awards, with an average of 1324 daily tweets over a 28 day period. More time limited peaks were seen after the song’s release (1151 tweets a day for three days) and after the 2018 Grammy Awards (1883 tweets a day for another three day period). Overall, 56.3% of tweets
+about Logic’s song between March 2017 and April 2018 were posted in the 34 day high impact period covering these three media events (table 1).
+Association with Lifeline calls
+Descriptive information about Lifeline users (gender, age, and demographic distribution of calls) is not
+routinely collected and thus was not available for the entire dataset; supplementary text S3 provides a breakdown of a not fully representative subsample).
+A statistically significant association was found for calls to Lifeline for the 34 day period covering the three main events, but not for the two minor events (the video release and news about the apparent effect of the
+song). The strongest increase was seen immediately after the MTV Video Music Awards. The increase was smallest for the period after the song’s release. This pattern was consistent with the observed public attention to these events, which was most pronounced for the MTV Video Music Awards (fig 1).
+The observed number of calls during the periods of song related media events exceeded the range of forecasted calls (based on baseline data) for the song’s release (5.3%, 95% confidence interval 0.53% to 10.0%, three day period), the performance at the MTV Video Music Awards (8.5%, 5.1% to 11.9%, 28 day period), and the performance at the Grammy Awards (6.5%, 1.7% to 11.2%, three day period) (table 2, fig 1). No significant associations were found for the two smaller spikes in public attention. A combined analysis across the three main media events indicated an excess of 9915 calls (95% confidence interval 6594 to 13236), corresponding to an increase of 6.9% (95% confidence interval 4.6% to 9.2%, P<0.001). Supplementary table S2 provides the parameter estimates for all model components.
+Association with suicides
+The observed number of suicides in the periods of Logic related media events were within the range of forecasted values, using the model fit to the baseline data. Estimates for the three major media events pointed to a decrease in suicides, but these estimates were not significantly different from the expected number of suicides (table 3, fig 1). Combining the data for all three major events into a single variable yielded an observed number of suicides that was below the range of the model forecasts. Models including a discrete pulse for these events indicated a significant reduction in suicides, amounting to a decrease of 245 suicides (95% confidence interval 36 to 453 suicides). This corresponded to a reduction of suicides of 5.5% (0.8% to 10.1%, P=0.02) in the 34 day period. Supplementary table S3 provides parameter estimates for all model components.
+A sensitivity analysis using daily unique Lifeline calls (rather than total daily calls) showed similar patterns to the original analysis (supplementary table S4). A further sensitivity analysis, using a pre-intervention cut-off date of 27 April 2017, showed no deviations
+from the original analysis (supplementary table S5). A third sensitivity analysis, combining all Logic related media events (including those not individually associated with daily Lifeline calls) into a single model, indicated that the association with Lifeline calls remained significant, whereas there was no significant association with suicides (supplementary table S6).
+Discussion
+This interrupted time series analysis found that Logic’s song “1-800-273-8255” was associated with a noticeable increase in calls to Lifeline (an additional 9915 calls or increase of 6.9%) during the 34 day period when public attention to the song was substantial. Over the same period, there was some evidence of a reduction in suicides, amounting to 245 fewer suicides (decrease of 5.5%).
+These findings are consistent with a possible Papageno effect3 and are important from a suicide prevention perspective. Media campaigns for suicide prevention have received a groundswell of support internationally, but evaluations are scarce and often limited in terms of scope.18 Our finding of a substantial increase in actual help seeking and a possible decrease in suicides during the period of high public attention to Logic’s song support the real world effectiveness of this intervention. Previous peaks in calls to Lifeline were almost always associated with harmful media events, such as celebrity suicides.19 These events were often associated with increases in suicides,1 20 indicating that both increases in calls to Lifeline and increases in suicides might reflect considerable distress in the community from these media events. The overall patterns found for suicides by celebrities included as covariates in this analysis were largely consistent with previous research. Most notably, the suicides of Kate Spade and Anthony Bourdain,15 which received the strongest and most sustained attention across all celebrity deaths in the post-intervention period, were clearly positively associated with Lifeline calls, and increases of suicides were close to the boundary of significance. In contrast, the patterns observed for Logic’s song, consistent with the song’s narrative, indicated an increase in help seeking behaviour accompanied by a slight reduction in suicides.
+The effectiveness of the song on calls to a helpline is a novel finding. The results show that it is possible to promote help seeking for suicidal crises in the absence of negative news, and indicate that suicides could potentially be reduced with prevention focused campaigning, such as Logic’s song. Although the reduction in suicides was small, this finding shows that the song did not result in harmful effects on suicide occurrence, which would have been indicated by an increase in suicides. This is important because some prevention messaging that aimed to reduce suicides was ultimately associated with increases in suicides.12 These narratives, however, typically focus on suicide deaths and attempts, not on hope and recovery from suicidal thoughts and feelings.
+The amount of exposure generated by prevention messages about hope and recovery seems to be crucial in efforts to yield positive effects on behaviours including help seeking and potentially suicides. In accordance with the most sustained and strongest public attention to Logic’s song, as indicated by tweet volume, the uptick seen for Lifeline calls was strongest after his performance of the song at the MTV Video Music Awards in 2017 followed by his performance at the Grammy Awards in 2018, and then the song’s release. This pattern underscores the importance of reaching large proportions of the target population to achieve behavioural effects. Compared with other prevention events, public attention for Logic’s song, as reflected in tweet volume, was more sustainable, but the event did not necessarily result in more tweets overall. Based on query terms used for World Suicide Prevention Day 202016 in the US, for example, the day resulted in a total of approximately 94 000 tweets from 58 000 users, including more than 30 000 on 10 September 2020 alone. Overall social media attention was comparable in total magnitude to Logic’s song (82 000 tweets from 55 000 users), but the attention was concentrated on a single day, whereas Logic’s song was highlighted repeatedly over several specific and diverse media events. This amount of attention was, however, still considerably smaller than for some harmful media events in the recent past—for example, in the three months after the release of 13 Reasons Why, a TV show that violated key recommendations for safe portrayals of suicide12-14 and was associated with an increase in teenage suicides,12 there were tweets from 870 056 individual users about the show.12 This is more
+than 15 times larger than for Logic’s song. Although Logic’s song sets an important, extraordinary example for widely disseminated and quite sustainable suicide prevention messaging, exposure is still considerably stronger for some conceivably harmful media events.
+Different pathways and mechanisms might be at play in any reduction in suicides associated with Logic’s song. The present analysis indicates that periods that were strongly associated with an uptick in calls showed a simultaneous decrease in suicides. We have not established whether calling behaviour affected people who did (or did not) die by suicide after the song’s release. People who are at risk of suicide are often socially withdrawn, and some might not consistently use helplines. The proportion of socially isolated, suicidal callers, however, has been found to be particularly high among frequent callers of crisis lines,21 indicating that the threshold for calling a telephone crisis centre is lower than for accessing help services that require on-site visits. The song, however, might have triggered other or additional routes of action beside calling Lifeline. Finding romantic love and positive communication with his family, for example, are major contributors to the improvement in suicidal thoughts and feelings of the protagonist seen in Logic’s music video, and some people might have felt inspired to reach out for help from other sources. In-depth qualitative research might help shed light on the question of whether and how precisely people with suicidal thoughts and feelings were influenced by the song.
+Strengths and limitations of this study
+A strength of this study is the length of the time series, with daily data from 2010 to 2017 available to model the expected pre-intervention suicide counts. The model controlled and adjusted for several exogenous variables (namely, concurrent side events, such as suicides of well known celebrities, and National Suicide Prevention Week). Trends, temporal fluctuations, and seasonality were adjusted for in the SARIMA models. A further strength of this study is the use of daily data, resulting in precise modelling in accordance with periods when social media attention to Logic’s song was strong, as indicated by the volume of tweets about it. Estimating exposure based on social media data is more objective than merely estimating exposure times in the absence of any supporting data. This refined
+approach is consistent with studies indicating that public attention to social media discourse in terms of tweets is short lived, with lifespans of a tweet normally not extending beyond a few hours.22
+The main limitation of the study is that it was based on ecological data, so it was not possible to ascertain whether the people calling Lifeline or not dying by suicide had been exposed to Logic’s song and related media events, or what their motivations might have been for calling or not dying from suicide. The observational nature of this research does not allow us to establish causality. Furthermore, only total daily call and suicide data were available, without any stratified data for various demographic factors, such as gender, age, or location of residence. Specifically for suicides, data stratified by age or gender would have resulted in small numbers on some days. This is a limitation because adolescents and young adults are over-represented among the viewership of the MTV Video Music Awards.23 Although this age discrepancy does not equally apply to the other events related to the song—the Grammy Awards 2018 was watched by 17% of Americans aged 50 or older24—or to the music preferences for rap or hip hop music in the US,25 a somewhat stronger exposure of young people seems plausible. Notably, the demographic profile (as well as primary presenting problems) of callers in September 2017, the month with strongest attention to the song, did not grossly deviate from other months in the observation period after the song’s release, thus indicating that possible age effects, if present at all, might have been limited (see supplementary appendix).
+Other limitations of our approach include the inability to assess longer term associations beyond the periods of strong public interest. As previously noted, 56.3% of tweets about Logic’s song between March 2017 and April 2018 were posted in the 34 day high impact period defined in this study, indicating that the bulk of attention was covered by the selected period. This is consistent with the generally short lived public attention to media events and related public discourse, which requires constant repetition to become sustainable.22 26 Furthermore, the different peaks identified were dissimilar not only in duration but also in the maximum number of tweets on a given day. The present impact periods have been modelled as discrete pulses, consistent with the assumption that visible, large changes in attention (rather than overall number of tweets) might most likely affect behavioural outcomes, such as help seeking and suicide. This is consistent with related evidence that other media data, such as those from Google Trends, generally struggle to predict general suicide trends,27 whereas sudden strong changes in Google search behaviours, as seen during major events such as the release of 13 Reasons Why or the covid-19 pandemic, do seem to be useful in estimating suicide trends.28 29 Studies are needed to assess how long the effects of suicide prevention messaging generally last and what absolute amount and duration of attention, as reflected in social media,
+is necessary to yield any observable effects. Social media data, such as tweets, are only a proxy of public attention, however, and might not always reliably reflect actual exposure. Finally, the determination of suicide deaths is a challenging task, and suicide deaths are sometimes under-reported.30 Even though a certain degree of misclassification is possible, rapid fluctuations in classification accuracy of national suicide data, which could impact the present analysis, seem unlikely.
+Conclusions
+This analysis suggests that Logic’s song “1-800-2738255” was associated with a noticeable increase in calls to Lifeline and a simultaneous small reduction in suicides during peak social media discourse about the song. The latter outcome is worth underscoring—the occurrence of a widely disseminated song and video was associated with more than 200 fewer suicides than expected. These findings emphasise the potential population health benefits of working creatively and innovatively with other sectors, such as the music and entertainment industries, to promote new impactful stories of help seeking that resonate with broad audiences, leave a visible footprint on social media, and are safe in terms of not featuring potentially lethal actions but rather coping and mastery of crises.13 Interventions that follow these principles could help create behavioural change to increase help seeking and prevent suicide.

Associations-of-adverse-childhood-experiences-and-social-support-with-selfinjurious-behaviour-and-suicidality-in-adolescentsBritish-Journal-of-Psychiatry.txt

@@ -0,0 +1,60 @@
+Non-suicidal self-injury (NSSI), suicidal ideation and suicide attempt are major public health problems in adolescents worldwide,1,2 and they represent some of the strongest and most consistent predictors of future suicidal behaviour across both in-patient and general populations.1,3 To sustain improvements in management and prevention initiatives, research continues to strive to better comprehend the complex interplay between many of the recognised psychosocial risk factors. Thus far, a substantial body of research has demonstrated significant independent effects between adverse childhood experiences (ACEs) and social support on self-injurious behaviour (SIB) and suicidality.4,5 Yet current knowledge surrounding these relationships is predominantly derived from Western countries and from adult or clinical popula-tions,4,5 with only a few studies undertaken in community adolescent populations in China.2,6 There is also little research on the interaction effects between ACEs and social support on NSSI, suicidal ideation and suicide attempt in adolescents, despite ACEs and social support being highly correlated.7 Finally, despite evidence differentiating boys and girls in terms of the prevalence and effects of different ACEs,8 the perceptions and utilisation of social support9 and the presentation of NSSI, suicidal ideation and suicide attempt,1 few studies have been undertaken to examine gender differences in the interaction between ACEs and social support on NSSI, suicidal ideation and suicide attempt. This is particularly important in nonWestern populations where there is a dearth of research despite the cultural context in China, which continues to demonstrate inherent gender discrimination.10,11 Therefore, our study first sought to
+investigate the independent effects of ACEs and social support on NSSI, suicidal ideation and suicide attempt in Chinese community adolescents; second, sought to examine the interaction effects between ACEs and social support on NSSI, suicidal ideation and suicide attempt and third, sought to ascertain whether there are any apparent gender differences in either independent or interaction effects for these relationships.
+Method
+Study sample and procedures
+Three provinces, namely Anhui, Henan and Guizhou, were chosen as our study fields for data collection. These provinces are broadly representative of the average population within China in terms of economic development and demographic composition, and are also where our adolescent health research network is located, thus facilitating the data collection. In each province, one region (Bengbu in Anhui province, Zhengzhou in Henan province and Guiyang in Guizhou province) was randomly selected. In each region, eight general junior and senior schools (four from rural areas) were randomly chosen to recruit participants. As 4 schools were combined junior and senior schools, only 20 schools were selected for inclusion in the survey. In total, 15 278 students aged 10-20 years, from grades 7-12, were contacted for this health survey and asked to complete an anonymous questionnaire. Informed consent was sought from parents/guardians, and 1.5% of
+the recruited participants or their parents/guardians opted out of the study. The design and data collection procedures were approved by the Ethics Committee of Anhui Medical University (2012534). The survey was conducted from November 2013 to January 2014.
+Measurement of sociodemographic profile, psychological symptoms, ACEs and social support
+Sociodemographic profile and psychological symptoms
+Demographic data for each participant was recorded, including age, gender (boys or girls), urban/rural residency, parents’ education level (less than junior middle school, junior middle school, senior middle school, college or more) and self-perceived economic status of the family (poor, moderate or good). Psychological symptoms, including emotional, behavioural and social adaptation symptoms, were evaluated by the psychological domain of the Multidimensional Sub-health Questionnaire of Adolescents12 (Cronbach’s a = 0.920 in this study).
+ACEs
+ACEs were defined as having experienced childhood maltreatment and/or household dysfunction. Childhood maltreatment was evaluated by the Child Trauma Questionnaire (CTQ),13 a widely used 28-item measure that assesses 5 different forms of childhood trauma (physical abuse, sexual abuse, emotional abuse, physical neglect and emotional neglect). The CTQ was translated and validated in Chinese.14 The participants were asked about abusive childhood experiences before 16 years of age. Responses ranged from ‘never true’ to ‘rarely true’, ‘sometimes true’, ‘often true’ or ‘very often true’. Respondents were defined as exposed to a category if they responded ‘very often’, ‘often’ or ‘sometimes’ to any item in that category. A Cronbach’s a coefficient of 0.737 was obtained for the CTQ in the current study. Household dysfunction questions were derived from the Centers for Disease Control and Kaiser Permanente Adverse Childhood Experiences Study in the USA.8 Household dysfunction was assessed through endorsement of the following experiences: (a) the divorce/separation of parents, (b) a parent serving time in jail, (c) having witnessed domestic violence, (d) having lived with someone who was mentally ill or suicidal or (e) having lived with someone with an alcohol or drug problem. Respondents were classified as exposed to household dysfunction if they responded; yes; to any item. The Cronbach’s a coefficient for the household dysfunction questionnaire was 0.705 in the present study. Because of the high interrelatedness of various types of ACEs (all P <0.01), an ordinal number of ACEs categories score was created by summing the dichotomous ACEs items (range, 0 (unexposed) to 6 (exposed to physical abuse, sexual abuse, emotional abuse, physical neglect, emotional neglect and household dysfunction)) to investigate the graded association of ACEs and both SIB and suicidality.8 The total score was then converted into four categories of summed score (0, 1-2, 3-4 and 5-6), with zero experiences selected as the referent for analysis purposes.
+Social support
+Social support was assessed by the 17-item Adolescent Social Support Scale,15 which includes three dimensions: objective support, subjective support and support availability. Participants reported whether an item from the scale was in ‘inconformity’, ‘little inconformity’, ‘uncertainty’, ‘little conformity’ or ‘conformity’. The scale scores, with a possible range of 17-85 (low to high social support), had a good internal consistency in the present study, with a significant Cronbach’s a coefficient of 0.940. The total score was divided into three levels (high, P75-P100; moderate, P25-P75; and low, P0-P25) for analysis.
+Measurements of NSSI, suicidal ideation and suicide attempt
+NSSI
+All participants received a screening questionnaire for NSSI, asking ‘In the past 12 months, have you ever harmed yourself in a way that was deliberate, but not intended to take your life?’. A list of eight NSSI methods were specified: hit yourself, pulled your own hair, banged your head or fist against something, pinched or scratched yourself, bitten yourself, cut or pierced yourself and burned yourself. Participants were then asked, ‘Have you ever done something with the intention of hurting yourself other than what was presented?’.6 For those who confirmed that they had engaged in NSSI, the frequency of NSSI was investigated. NSSI was dichotomised (frequency of NSSI of three or more versus fewer than three as yes or no, respectively) for analysis. The internal consistency reliability of NSSI was 0.749 in the current study.
+Suicidal ideation and suicide attempt
+Suicidal ideation and suicide attempt refer to the ‘middle school questionnaire’ of the 2013 Youth Risk Behaviour Surveillance System in the USA.16 Suicidal ideation was defined as a ‘yes’ in response to the question ‘Have you ever thought about killing yourself in the past 12 months?’. Suicide attempt was defined as a ‘yes’ in response to the question ‘Have you ever tried to kill yourself in the past 12 months?’.
+Statistical analysis
+Of the 15 278 school adolescents recruited, 458 (3.0%) were excluded from the study because of absence from school on the day of the survey or unwillingness to respond to the questionnaire, or high levels of missing data or obviously fictitious or inconsistent responses. Thus, a total sample of 14 820 (97.0%) participants was analysed.
+Sociodemographic data, ACEs, social support, NSSI, suicidal ideation and suicide attempt were described in both the total population and for boys and girls separately. Gender differences were assessed with the x-test for categorical variables and one-way analysis of variance for continuous variables. Binomial logistic regression models were used to examine the associations of NSSI, suicidal ideation and suicide attempt with ACEs and social support individually, and then in combination. In the models, adjustment was made for age, gender, regional area, school, urban/rurality, mother’s education level, economic status of family and psychological symptoms. In examining the association of NSSI with ACEs and social support, we also used the thresholds of NSSI score >1 and >5 for sensitivity analysis.
+Gender differences in the associations were examined via two odds ratios (Ratio of two odds ratios, RORs). 7 All analyses were conducted with SPSS software, Windows version 16.0 (SPSS Inc., Chicago, IL).
+Results
+Characteristics of participants
+Of the 14 820 participants, the mean age was 15.4 years (s.d. 1.8), and 50.2% were girls. Scores showed that 45.7% of the sample had experienced childhood emotional abuse, 20.3% had experienced physical abuse, 13.3% had experienced sexual abuse, 64.2% had experienced physical neglect, 58.5% had experienced emotional neglect and 42.5% had experienced household dysfunction; in total, 89.4% had experienced one or more ACEs and 46.3% reported three or more types of ACEs. Compared with boys, girls had significantly more psychological symptoms, fewer ACEs and a higher level
+of social support (P <0.001). Boys had significantly increased exposure to physical and sexual abuse, physical and emotional neglect, household dysfunction and NSSI (P < 0.001). Girls had significantly greater exposure to emotional abuse, suicidal ideation and suicide attempt (P <0.001). The details of gender differences and sociodemographic factors can be seen in Table 1.
+Effect of ACEs and social support on NSSI, suicidal ideation and suicide attempt, and gender difference
+Table 2 shows the number and percentage of participants according to NSSI category, among different levels of ACEs and social support. There were significant trends toward increased NSSI with higher ACEs and lower social support. Multiple adjusted odds ratios for NSSI were significantly increased with higher ACEs and lower social support (model 2 in Table 2). Even when ACEs and social support were included in the model simultaneously, there were main effects of ACEs score and level of social support on NSSI (model 3 in Table 2). Using the thresholds of NSSI > 1 and NSSI > 5 for separate data analysis, we found that the associations of ACEs score and level of social support (Supplementary Table 1
+148
+https://doi.org/10.1192/bjp.2018.263 Published online by Cambridge University Press
+available at https://doi.org/10.1192/bjp.2018.263) were similar to those found with NSSI > 3 (Table 2).
+Supplementary Table 2 shows data on suicidal ideation and suicide attempt in relation to ACEs and social support. There were significant trends toward increased suicidal ideation and suicide attempt with higher ACEs and lower social support. Multiple adjusted odds ratios for suicidal ideation and suicide attempt were significantly increased with higher ACEs and lower social support, respectively. When ACEs and social support were put in the model simultaneously, the main effects of the ACEs and social support on suicidal ideation and suicide attempt remained, with the exception of social support on suicide attempt in boys.
+No gender differences were found in the independent effects of ACEs or social support on NSSI, with the exception of the lowest social support having a stronger effect in girls than in boys (Table 3). In the data analysis for NSSI > 1 and NSSI > 5, gender differences in ACEs score or social support on NSSI (Supplementary Tables 3 and 4) were similar to those found with NSSI > 3 (Table 3), whereas the NSSI > 5 data showed a borderline significance for the lowest social support having a stronger effect in girls than in boys. There were no gender differences in the effects of
+ACEs, Adverse childhood experiences; NSSI, non-suicidal self-injury.
+a.	Unadjusted model.
+b.	Adjusted for age, regiona l areas, schoo l , urban/rura l ity, mother’s education l eve l , economic status of fami l y, psychol ogica l symptoms, ACEs score and l eve l of socia l support.
+c.	Calculated by adjusted odds ratio.
+ACEs or social support on suicidal ideation (Supplementary Table 5), but the effects of high ACEs score and low or moderate social support on suicide attempt were significantly stronger in girls than in boys (Table 4).
+Further analysis was conducted to examine whether specific types of ACEs demonstrated gender differences in the effect on NSSI, suicidal ideation and suicide attempt (Supplementary Tables 6-8). Each type of ACEs was significantly associated with NSSI, suicidal ideation and suicide attempt in boys and girls,
+except for emotional and physical neglect with suicide attempt in boys (Supplementary Table 8). Emotional abuse increased the risk of NSSI, suicidal ideation and suicide attempt more than other types of ACEs among both genders. When exposed to emotional abuse, girls were more likely to engage in NSSI than boys (Supplementary Table 6). When exposed to physical abuse, girls were also more likely to report a suicide attempt than boys (Supplementary Table 8). Conversely, suicidal ideation was more common among boys than girls when exposed to emotional
+neglect (Supplementary Table 7). No further gender differences were found.
+Interaction effect between ACEs and social support on NSSI, suicidal ideation and suicide attempt, and gender difference
+Although ACEs and social support were highly correlated in the present study (Supplementary Table 9), there were no interaction effects between ACEs and social support on NSSI (total sample, P = 0.062; boys, P = 0.521; girls, P = 0.115), and suicidal ideation (total sample, P = 0.087; boys, P = 0.061; girls, P = 0.075). However, there was an interaction effect of ACEs and social support on suicide attempt in the total sample (P = 0.001) and in boys (P = 0.002), although it was not significant in girls (P = 0.334).
+Discussion
+A large-scale, school-based survey was conducted to examine the independent and interaction effects between ACEs and social support on NSSI, suicidal ideation and suicide attempt in adolescents. Our data reveals that ACEs and social support have main effects on both SIB and suicidality independently, as well as interaction effects between ACEs and social support on suicide attempt but not NSSI or suicidal ideation. Similar relationships are found across genders; however, girls were more likely to engage in NSSI and suicide attempt when social support was low, and an interaction effect between ACEs and social support on suicide attempt was found only in boys.
+The effects of ACEs on NSSI, suicidal ideation and suicide attempt
+Previous research has shown a relationship between ACEs, NSSI and suicidality.4 An investigation in 957 undergraduate students in Canada suggested that those experiencing more adverse familylife events and higher perceived relational trauma were more at risk of engaging in NSSI behaviours.18 A prospective study in youths aged 14-26 years also suggested that physical abuse, emotional abuse and emotional neglect were associated with subsequent risk of suicidal behaviour.19 Our study has extended this literature by demonstrating gender-specificity effects in the relationships, and by addressing the limitation of the lack of equivalent research within Chinese community populations. However, the relationships observed between all maltreatment types and NSSI, suicidal ideation and suicide attempt conflicts with other studies20 in clinically referred youth, which indicate that only indirect childhood maltreatment (i.e. witnessing domestic violence) is significantly associated with NSSI, whereas direct forms of maltreatment (physical or sexual abuse) are not. Similarly in adult samples, varied effects were found with different forms of childhood adversity, as childhood abuse was not significantly associated with NSSI and suicide attempt after adjusting for the correlation with low maternal or paternal care.21 Thus, various study samples, different definitions of ACEs and diverse control variables should be considered to interpret the results of the study.
+The tradition of son preference remains prevalent in China.10 This may contribute, at least in part, to the poorer mental health outcomes (including depression, anxiety, low self-esteem, sensitivity to negative life events and interpersonal pressure) previously observed in female Chinese school children.22 Yet, no study has attempted to identify gender differences in the relationships between ACEs, SIB and suicidality in Chinese adolescents. Although the risk of NSSI, suicidal ideation and suicide attempt increased in line with a greater number of ACEs categories
+generally, our study suggests that girls who experienced a higher number of ACEs (categories) seem to be more vulnerable to suicide attempt (but not NSSI or suicidal ideation) than boys, which is in line with prior studies.23,24 Moreover, girls were found to be more susceptible to NSSI and suicide attempt when they encountered particular forms of maltreatment, primarily emotional or physical abuse. This substantiates findings from a series of studies, including Isohookana et al,23 who examined psychiatric in-patients aged 12-17 years and found that a higher number of ACEs was associated with an increased risk of NSSI and suicide attempt in girls, but not in boys. Garcia et al24 also showed that significant correlations were found between childhood trauma scores and psychotic symptoms, depressive symptoms and poorer functionality, but only in women, whereas childhood trauma was associated with poorer social cognition in both males and females. Further studies are needed to ascertain whether there are gender differences in the relationship between ACEs, SIB and suicidality.
+The effects of social support on NSSI, suicidal ideation and suicide attempt
+Social support was found to have an independent effect on NSSI, suicidal ideation and suicide attempt, even after adjusting for sociodemographic risk factors, psychological symptoms and ACEs. Previous findings5,25 have also indicated that individuals with higher social support have a significantly lower odds of engaging in NSSI, suicidal ideation and suicide attempt, which is consistent with our study. One study in a clinical sample of adolescents suggested that perceptions of school support were independently and negatively associated with suicidal ideation, especially among adolescents who also reported perceptions of lower parent support.5 Furthermore, lower perceived parental support was independently associated with greater odds of history of suicide attempt. One study in community and in-patient mental health settings25 also showed that children and adolescents who had some form of social support had a 26% decrease in the odds of engaging in NSSI when compared with their counterparts who lacked social support. Collectively, this supports Ayub’s26 suggestion that social support may play a significant role in the prevention of suicidal thoughts and behaviours, and that psychologists should include family and friends in their approaches to treating suicidal youth. This may be especially important in China, where mental illness is frequently blamed on the family and the individual.11 Moreover, considering that Chinese students are often burdened with tremendous academic pressure,27 support within the school environment may be particularly beneficial in this context.
+The inverse association between social support and NSSI was found to be stronger in girls than in boys. This may contribute, in part, to the explanation of why the risk of suicidal behaviours is greater in girls despite having fewer ACEs and higher social support than boys. Traditionally within China, boys are more likely to be socialised to be independent than girls, which may help to account for the increased sensitivity to lower social support observed among females in our sample. Other studies examining developmental trajectories of suicidal ideation showed that support from family and friends differentiated suicidal ideation trajectories for both boys and girls.28 However, an ecological study in adults in 75 regions of 23 European countries revealed inverse relationships between social support and suicide rates for both genders, with some indication of a stronger relationship among men.29 Future studies may look to examine whether different types of social support (e.g. peer or parental support) are more influential for males or females.
+Interaction effects between ACEs and social support on NSSI, suicidal ideation and suicide attempt
+Our study extends existing knowledge by demonstrating the interaction effect between ACEs and social support on suicide attempt, particularly in boys, but failed to establish a similar relationship for either NSSI or suicidal ideation. The relationship between ACEs, social support and SIBs and suicidality is complex. Christoffersen et al suggested that social support is a partial mediator between traumatic life-events and NSSI in young adulthood. A study of women in the USA found that the link between intimate partner violence and suicidal behaviour was moderated by social support.30 Extending this to the adolescent population, it would be reasonable to suggest that social support may also moderate the relationship between ACEs and SIBs in this population. That said, our study failed to establish interaction effects between ACEs and social support on NSSI and suicidal ideation. This may be partly attributable to the confounding effect of psychological symptoms, because these have found to be significantly related to ACEs, social support and SIB.2, ,7 In this study, statistical interaction effects of ACEs and social support on suicidal ideation were found after removing psychological symptoms from the multivariable model (results not shown). One possible alternative explanation may be that interaction effects are only apparent among those who are most seriously affected (such as those who have actual suicide attempts), as this group may be most likely to encounter the highest level of ACEs and lowest social support. Further studies on the interaction effects between ACEs and social support on NSSI, suicidal ideation and suicide attempt will be needed to further elucidate this complex interaction.
+Strengths and weakness of the study
+This study is the first to examine gender differences and interaction effects between ACEs and social support on NSSI, suicidal ideation and suicide attempt within Chinese adolescents. The study sample is representative, covering urban and rural areas in China, and the response rate of the participants was high. The large sample from urban and rural areas provided enough statistical power to examine gender differences, with multivariate adjustment analysis. However, the study has limitations. First, the study was crosssectional, therefore it is difficult to establish a causal relationship. Nonetheless, our findings pertaining to the association between ACEs and social support with SIB and suicidality were similar to those in previous cohort studies.19,28 Second, because of the use of self-reported questionnaires for data collection purposes, it is possible that recall bias may exist. This may ultimately influence the strength of the observed relationships, and our results may represent a more conservative estimation than is truly present. Third, the focus was solely directed toward the number and type of ACEs in this study, but it may be important to understand an individual’s subjective experience of the events. Finally, the study focused on adolescents in traditional school environments, therefore the findings did not represent adolescents who were absent from school, which is important because studies have shown that ACEs, lower social support and suicidality are more prevalent in individuals with lower educational achievement and socioeconomic status.31 Caution should be exercised in the application of the findings to the whole population of adolescents in China.
+Implications
+The findings indicate that ACEs and poorer social support are independently associated with an increased risk of both SIB and suicidality in school adolescents. Moreover, an interaction effect was observed between ACEs and social support on suicide attempt,
+which, by implication, suggests that the combination of these factors may be particularly detrimental in increasing the likelihood of behavioural enactment. In light of this, intervention and prevention strategies focused on enhancing perceived social support as a fundamental feature, particularly among female adolescents with a history of ACEs, may go some way toward mitigating the negative trajectory of ACEs in this population.
+Educational settings are likely to represent an important conduit through which to improve the quality and accessibility of social support available to vulnerable adolescents. On a targeted level, awareness-raising initiatives should aim to integrate a training element specifically focused on the psychoeducation of teachers and other school staff. This should centre around both improving their recognition of the signs and symptoms of mental illness, particularly among individuals with a known history of ACEs, and on developing basic training in mental health literacy to facilitate appropriate emotional responses.32 On a universal school level, interpersonal skills training for pupils, provided in the context of an educational setting, may help to improve skills in social and peer interactions with a view to enhancing the quantity and quality of social support available to adolescents.33 Finally, as part of a broader systems approach, schools may also seek to improve perceptions of social support through increased participation in school and community life.33 This can be achieved through the strengthening of ties between schools and communities to facilitate greater cohesion and engagement in extracurricular activities, which will ultimately contribute to the construction of wider social networks for adolescents.

Behavioral Sci The Law - 2018 - Ortiz - Traditional and new media s influence on suicidal behavior and contagion.txt

@@ -0,0 +1,51 @@
+1	| INTRODUCTION
+The effect that traditional media (e.g. newspapers, television) can have on suicide contagion has been well researched, and several countries, such as Australia, Canada, and the UK, have suggested guidelines for media reporting on suicide in order to reduce the risk of suicide contagion. Since the end of the 20th century, however, several new forms of media have emerged, and have quickly been adopted as a primary form of social interaction, especially for adolescents and young adults. This article will briefly review the existing literature regarding the media's effect on suicide and suicide contagion. Special populations will be discussed, including those who appear to be more susceptible, and those who may contribute more heavily, to suicide contagion. Current recommended guidelines on media reporting of suicide will be examined. A review of the current literature will explore the effect of new forms of media on suicide and suicide contagion. Finally, updated recommendations will be made for mitigating the effects of both traditional and new media on suicide contagion.
+“Suicide contagion” can be defined as the process by which one suicide becomes a compelling model for successive suicides (Gould, 2001). Other similar terms used are “imitation,” “copy-cat,” or “cluster” suicide. However, there seems to be no clear consensus on the exact definitions of these terms, and even the underlying concept or theory of “suicide contagion” itself has yet to be defined consistently (Cheng, Li, Silenzio, & Caine, 2014). In this paper, “suicide contagion” will refer to the spread of suicidal thoughts, behaviors, and deaths, specifically as a result of exposure to suicide-related material in traditional or new media, although this is certainly not the only way one can be exposed to suicide. “Suicide cluster” will refer to a larger-than-expected number of suicides, or attempted suicides, occurring more closely in time and geographical proximity than would typically be expected. Clusters can be viewed as statistical deviations from the norm, or as groups of suicides that are somehow linked, as a kind of contagion (Robertson, Skegg, Poore, Williams, & Taylor, 2012).
+The term “the Werther effect” has also been used to describe the phenomenon of suicide contagion. A 1974 article for the American Sociological Review first demonstrated that media attention to a suicide could lead to a spike in the number of suicides in the geographical area of the reporting (Phillips, 1974). Phillips coined the term “the Werther effect” to describe suicides that occur as a result of exposure to the media reporting of another suicide. The term refers to the book The Sorrows of Young Werther by Goethe, published in 1774, in which the protagonist, Werther, falls in love with a woman who is beyond his reach. He consequently decides to end his own life by sitting at his desk and shooting himself, wearing boots, a blue coat, and a yellow vest (von Goethe, 2012). After the book was published, several suicides occurred across Europe with significant evidence that at least some were influenced by the novel: victims were found dressed in similar clothing, they used the method as described in the book, or the book was found at the scene of the death (Jack, 2014). This is commonly regarded as the earliest known occurrence of suicide contagion related to a form of media.
+3	| EVIDENCE REGARDING TRADITIONAL MEDIA
+There have been several reviews of studies looking at the effect of traditional media on suicide contagion. Two comprehensive reviews, both titled “Suicide and the media,” were published in 2001. Gould reviewed studies on suicide-related effects of nonfictional (e.g. newspaper and television news reports) and fictional media (e.g. television soap operas, fictional movies), all published between 1967 and 1992. She concluded that there was substantial evidence to support the idea that publication of nonfictional stories in the media (i.e. newspapers) is associated with an increase in the rate of subsequent suicides. She also found that the magnitude of this increase is proportional to the amount, duration, and prominence of the media coverage. Regarding fictional media, the studies were found to have more contradictory results; however, Gould concluded that there was “ample evidence of an imitative effect of these broadcasts” (Gould, 2001).
+Pirkis and Blood reviewed much of the same literature as Gould, and they also found a relationship between nonfictional media portrayal of suicide and subsequent increase in actual suicides. They noted that this association fulfilled many of the criteria for causality, such as temporality and specificity. The effect appeared to vary as a function of time, with a peak at three days and attenuation by two weeks. Greater numbers of media items and more prominent “high impact” stories were associated with a stronger effect. They also found that the effect was greatest when the initial suicide victim was similar to the observer in terms of age or sex, or when the decedent was revered in some way by the observer, as with celebrities. Finally, there was evidence that, when there were explicit descriptions of the method of suicide included in the media, there was an increase in actual suicidal behavior employing that method (Pirkis & Blood, 2001a). In terms of fictional media, the authors examined studies that considered the relationship between fictional portrayal of suicide (e.g. film, television, music, plays) and subsequent suicidal behavior. They found the evidence for this effect to be more equivocal and not sufficient to make a case for causality, while acknowledging the need for more research in this area (Pirkis & Blood, 2001b).
+4	I SPECIAL POPULATIONS
+Studies related to the effect of reporting on celebrity suicides have mixed results. While there does appear to be an association between reports on celebrity suicide and an increase in rate of suicides, there have been inconsistent findings with different celebrities. One early study of media and suicide contagion found that suicide rates increased by 40% in the month after Marilyn Monroe died by suicide in 1962, but there was no significant change in the yearly rate of suicide. The increase was largely attributed to a spike in male suicides, suggesting a “reaction to loss” element. The next year saw an 42% increase in female suicide, which raised the issue of identification with the person whose suicide is publicized. However, after Ernest Hemingway suicided in 1961, there was no evidence of increase in suicide rates. The author hypothesized that if identification with the model is limited to the act of suicide alone, then the drive to imitate the behavior may be reduced (Motto, 1967). One study found a significant rise in the national suicide rate after celebrity suicides were covered on the front page of the New York Times (Gould, 2001; Wasserman, 1984). Another study found that even stories reporting on noncelebrity suicides were associated with a significant increase in the national suicide rate, although the increase was not as large as that seen after celebrity suicides were published (Gould, 2001; Stack, 1990). No significant increase in suicide was found following the death of grunge rock star Kurt Cobain, which may have, at least in part, been a result of the significant efforts by his widow, Courtney Love, to present his suicide in a negative light and minimize glamorization of his death (Gould, 2001). Other studies have suggested that celebrity type, region of the study, and other specific observer/model variables may have a significant impact on the contagion effect (Niederkrotenthaler et al., 2012; Stack, 2002; Yang et al., 2013).
+The rate of suicide between ages 10 and 24 in the United States increased dramatically between 2000 and 2015. Among young adolescents aged 10 to 14, there was a 35% increase in suicide rates, teenagers aged 15 to 19 showed a 22% increase, and suicides among young adults aged 20 to 24 increased by 21% (CDC, 2016). Adolescent populations appear to be more susceptible to imitating suicidal behaviors after being exposed to suicide, either in the media or through knowing the original victim through school or personally (Swanson & Colman, 2013). Suicide clusters occur predominantly among people aged 15 to 24, and one adolescent suicide is known to be a risk factor for additional suicides. In addition, the evidence suggests that when clusters occur, the deaths represent statistical increases over the expected suicide rate; that is, they are not just suicides that would have eventually occurred anyway (Robertson et al., 2012), suggesting that exposure to suicide activity may lead vulnerable adolescents to attempt suicide when they otherwise might not have.
+5	| EXISTING GUIDELINES
+Many countries have developed guidelines promoting responsible reporting of suicide by the media. They tend to have similar content but differ in the way in which they have been developed and implemented (Pirkis, Blood, Beautrais, Burgess, & Skehan, 2006). Recommendations apply to online content including citizen-generated media coverage, social media sites, blogs, and online content from traditional media organizations’ websites (Reporting on Suicide, 2015) (Table 1).
+The effectiveness of these guidelines has not been rigorously evaluated given methodological obstacles and low base rate for completed suicide (Sudak & Sudak, 2005). There have been studies showing that suicide rates decreased after implementation of guidelines; however, correlation could not be proven (Pirkis et al., 2006).
+Several studies have attempted to assess adherence to suicide reporting guidelines across various countries and forms of media, including new media, with disappointing results (Abbott, Ramchand, Chamberlin, & Marcellino, 2017; Easson, Agarwal, Duda, & Bennett, 2014; John et al., 2016; Utterson, Daoud, & Dutta, 2017; Young, Subramanian, Miles, Hinnant, & Andsager, 2017). A recent study examining the fidelity of media reporting on Robin Williams’ suicide found that 55% of articles surpassed the 80% threshold for “high fidelity” and 85% of articles applied at least 70% of the guidelines. The most commonly overlooked recommendation was “tell others considering suicide how they can
+get help,” which was missing from over 70% of the articles (Creed & Whitley, 2017). Suggestions have been made to increase adherence to guidelines, including involving media organizations in the development of suicide reporting strategies to increase buy-in and increasing efficiency of disseminating recommendations to media organizations (Scherr, Arendt, & Schafer, 2016). While more research is needed, the development of new instruments to systematically evaluate news reports could advance our knowledge in the near future (John et al., 2014, 2016; Nutt, Kidd, & Matthews, 2015).
+Another notable void is the lack of accountability for those violating the existing guidelines or laws on media reporting of suicides. While there are many countries, such as Australia, Canada, the UK, and the USA, who have developed guidelines for suicide reporting, New Zealand stands alone as the country with criminal laws governing what can be said publicly about a suicide. These legal restrictions have been developed to prevent suicide contagion in New Zealand. Starting in 1988, it became a criminal offense to report details of suspected suicides without the coroner's ruling that it was safe to publish the details. In 2006, this law was tightened further with the Coroners Act, which restricts reporting or publicly discussing specific aspects of individual suicides. In recent years, the 2016 Coroners Amendment Act narrows reporting restrictions to the details most likely to lead to copycat behavior. Despite these well-laid-out legal restrictions, there has been no record of anyone ever having been prosecuted or fined for breaching them (New Zealand Ministry of Justice, 2006, 2016).
+6	I NEW MEDIA
+Suicide rates in the USA have been slowly but steadily increasing over the last several years. In 2006, the suicide rate was less than 11 per 100,000 individuals. That number has increased by over 20%, to 13.26 per 100,000 as of 2015 (American Foundation for Suicide Prevention, AFSP, 2016). While the cause of this increase is yet unknown and likely multifactorial, one undeniable fact is that, within this same time span, the Internet has provided the opportunity to disseminate a plethora of information across the globe in seconds. Given the convincing evidence for the existence of a contagion effect, it is prudent to examine the role that new media may be playing in this concerning trend.
+In this article, “new media” is defined as media that facilitates communication and requires an Internet connection, such as websites, social media, blogs, forums, video games, electronic messaging applications, online television and movie streaming services, and others, in addition to SMS text messaging available on cellular phones.
+7	| NEW MEDIA, SUICIDE, AND SUICIDE CONTAGION
+Research exploring the relationship between the Internet and suicide has steadily increased in the last two decades. The most common topics published between 1997 and 2015 include searching for suicide-related information online, online suicide interventions and their effectiveness, conversations about suicide and suicide methods on Internet forums, and online behaviors and how people use the Internet. Less than 4% of the publications during the same time period focused on suicide contagion or suicide pacts made online (Krysinska et al., 2017). Most of the research has attempted to learn more about the individuals who are searching for suicide content, and much less has been published about the Internet as a source of suicide contagion (Robertson et al., 2012).
+Literature regarding the Internet and any potential influence it may have on suicidal behavior or self-harm in adolescents and young adults shows mixed results. Studies on general Internet use gave strong evidence, according to a systematic literature review of publications from 2011 through January 2015 by Marchant and colleagues. This review found that high Internet use (more than 2-5 h per day) and Internet addiction correlated with more suicidal ideation and self-harm behaviors, but causality was unclear. Studies exploring particular online media, such as social media, forums, or blogs, were smaller and provided lower quality evidence. Distressed online posts appear to be related to suicidal ideation and behavior in young people, but there is little evidence to suggest that social media use itself increases risk (Marchant et al., 2017).
+Despite the strong link between exposure to suicidal content in traditional media and subsequent increases in suicide rates, the relationship has been more difficult to characterize when content is consumed online. Suicide-related videos and images often have a high number of views and comments, and one high quality study found that comments on videos may perpetuate self-harm behavior among teens and young adults (Marchant et al., 2017).
+Twitter posts in the UK were monitored after a British soap opera aired an episode centered on an assisted suicide; although mentions of the word “suicide” did increase after the show aired, there was no evidence of increased communication of suicidal intent. On the contrary, suicide-related communication on Twitter tended to be more protective (Scourfield et al., 2016).
+A recent Japanese study found evidence of suicide contagion after celebrity suicides that resulted in a strong reaction to the death on Twitter. Actual suicide rates increased when the celebrity death, often a young entertainer, generated a large number of tweets. Interestingly, no increase in suicide rate was noted when the death received little attention on Twitter, even if there was considerable coverage in traditional media (Ueda, Mori, Matsubayashi, & Sawada, 2017).
+Social media and text messaging have further complicated the already onerous task of identifying suicide clusters, which can impede timely intervention. Electronic communication technology has made the traditional definition of a cluster more fluid, with geography becoming less important as information is spread further and more rapidly, significantly expanding the size of an individual's circle of influence (Robertson et al., 2012).
+One more recent, and very concerning, trend is the so-called “Blue Whale” game that appears to exist on social media networks around the world. The game is made up of “mentors,” giving children and adolescents 50 tasks to complete and send photo proof of. These tasks range from watching horror films and listening to “psychedelic” music to self-mutilation, such as carving symbols or words into their arms or cutting their lips, to dangerous activities such as climbing on a roof. The game concludes with the 50th task, which is to carry out their own suicides. Hundreds of suicides are suspected of being linked to this game around the world, in countries such as Russia, France, Romania, and Brazil; however, not all of these are confirmed (Ferreira de Sousa, de Deus Quirino Filho, de Cassia, Bezerra dos Santos, & Rolim Neto, 2017).
+In the spring of 2016, the Internet-based entertainment company Netflix released a series co-produced by actress and pop singer Selena Gomez called 13 Reasons Why, based on the 2007 young adult book of the same name by Jay Asher. The series is based on the aftermath of the suicide of a 15-year-old girl who left 13 audio recordings on cassette tapes, each addressed to a person that she felt contributed in some way to her decision to commit suicide. The final episode includes a graphic three-minute long depiction of the suicide (Howard, 2017). Internet search-engine queries on suicide increased by 19% in the 19 days following the release of the series. There was evidence of increased suicide awareness, but the searches also indicated an increase in suicidal ideation. It is unclear whether or not any of the searches preceded actual attempts (Ayers, Althouse, Leas, Dredze, & Allem, 2017). One small study in an urban teaching hospital in New Jersey found that there was a statistically significant increase in psychiatric presentations to their emergency department in the days following the release of the series; however, there was no change in the number of presentations related to suicidal ideation or attempts, and no significant increase in the number of psychiatric admissions (Salo et al., 2017).
+Given the evidence covered so far, it is no surprise that the series attracted much attention from the media, concerned parents, schools, and the health community. There was much debate over the public health implications of the series. Critics claim that the show targets a teen audience, but presents mature topics in an “adult way.” They allege that the series romanticizes suicide, does not focus enough on mental illness, villainizes parents and school officials, depicts the suicide as a rational choice and a form of revenge, and makes the victim appear to be a role model. The graphic depiction of the suicidal act at the end of the series has caused great concern that it may lead to a contagion effect, and multiple prolonged scenes with physical and sexual violence that appear in other episodes may trigger vulnerable individuals (Jacobson, 2017). Some also feel that the problem is not that the series was created, but that many young people “binge watched” it without parental guidance or knowledge that it existed (Knopf, 2017). Proponents of the show feel that it raises awareness of the adolescent suicide problem and highlights many problems that have been linked to suicide risk: social media harassment; sexting; sexual, emotional, and physical assault; and alcohol and other drug abuse (Tasman, 2017).
+Of note is that the second season of the show has already begun filming. The National Association of School Psychologists (NASP) has issued recommendations on how to manage effects from the first, or second, season. They recommend that parents watch the series with their children, discuss the material, and actively listen to any concerns they may have without judgement. It is also important for parents to remember that discussing suicide does not increase the risk of suicide or “plant the idea,” but avoiding the subject when a person is vulnerable could lead to tragic consequences. Parents are encouraged to seek professional help from a school-employed or community-based mental health provider if they think their child, or someone they know, is at risk (Knopf, 2017). Another interesting suggestion is for mental health professionals to consider volunteering to speak at schools about adolescent suicide and other important concerns raised by the show (Tasman, 2017).
+9	| USING NEW MEDIA TO IDENTIFY AT - RISK INDIVIDUALS
+New media has provided access to enormous amounts of information and insight into people's lives and thoughts as never before. Recent research has focused on finding ways to use this information to identify vulnerable populations as early as possible, which would allow for proactive interventions, instead of intervening reactively as is most common currently.
+Much attention has been paid to search engines and finding trends in suicide-related search terms. Web searches for suicide-related terms tend to retrieve more protective than harmful websites; however, resources with harmful characteristics tend to be ranked higher in the results (Till & Niederkrotenthaler, 2014). In addition, a considerable proportion of the results expressed mixed or neutral attitudes toward suicide, and a small percentage were clearly
+pro-suicide (Thornton, Handley, Kay-Lambkin, & Baker, 2017). Suicide search trends have been found to correlate with actual suicide rates in Taiwan (Yang, Tsai, Huang, & Peng, 2011), and elaborate search-engine algorithms have been created to identify search terms and trends that more accurately detect high risk individuals who are searching for suicide-related information (Arendt & Scherr, 2017). However, more recent studies have questioned the use of this information, and have found that the validity of using search volumes to predict suicidal activity and suicide rates is actually low (Tran, 2017). For example, online searches for suicide-related terms in Italy are more likely to be related to personal interest or bereavement than suicidality (Solano et al., 2016).
+Social media posts can also be used to identify individuals who are at risk of suicide or self-harm. Twitter may provide an opportunity for real time monitoring of suicide risk factors on a large scale (Jashinsky et al., 2014). While having a Twitter account alone is not associated with suicidal behavior, tweets containing the phrases “want to die” and “want to commit suicide” are significantly related to suicidal ideation and behavior (Sueki, 2015). Other studies have suggested that simply monitoring posts for suicide-related terms is not sufficient. Sophisticated machine learning algorithms (Braithwaite, Giraud-Carrier, West, Barnes, & Hanson, 2016) and linguistic analyses can be used to identify tweets with higher levels of risk (O'Dea, Larsen, Batterham, Calear, & Christensen, 2017), and can even account for cultural variability in markers of suicide risk and emotional distress (Cheng, Li, Kwok, Zhu, & Yip, 2017). One study analyzed social media profiles of military personnel who died by suicide, and was able to identify temporal sequences in different types of posts that were unique to suicidal individuals (Bryan et al., 2017).
+Recently, Facebook introduced updated suicide prevention tools, in addition to the teams they already have in place, who monitor reported posts for suicidal content. They now use artificial intelligence technology to recognize patterns in a person's posts and make reporting these concerning posts easier for anyone who sees them. They also provide options for real-time assistance to people streaming on Facebook Live, which is a feature that allows users to live stream videos to their friends through the website or mobile application. Additionally, they now offer live chat support provided by crisis intervention organizations through their Messenger application (Callison-Burch, Guadagno, & Davis, 2017).
+10	| PREVENTION AND INTERVENTION OPPORTUNITIES
+Adolescents are more likely to communicate distress on social media to their peers, meaning that teens are likely exposed to high risk posts on a regular basis without recognizing it. Despite the many dangers and risks the Internet has created in terms of suicide risk, it also has potential to provide many new methods of prevention and intervention. In fact, the Internet may provide the opportunity to be a part of a community for many people who are otherwise isolated in everyday life, as evidenced by what appears to be the protective influence of low levels of Internet use versus no Internet use at all. In addition, young people are increasingly using social media to communicate hardships and emotional pain to their peers. Internet forums have been found to be supportive in terms of helping suicidal individuals find communities they could relate to, share their experiences with, and find mental health resources; however, there were also pro-suicide forums that encouraged and normalized self-harm and suicidal behaviors. Whether or not these exposures lead to suicidal behavior is still unclear and likely varies with individual circumstances (Marchant et al., 2017).
+One major advantage of the Internet is that is has provided access to communities that are traditionally difficult to engage, such as lesbian, gay, bisexual, transgender, and questioning/queer (LGBTQ) individuals or those who are severely isolated (Marchant et al., 2017; Robinson et al., 2017). Research has shown that school psychoeducational prevention programs that teach appropriate responses to suicidal posts may mitigate some of the negative influences on the Internet (Marchant et al., 2017). Young people can safely participate in developing suicide prevention messages for their peers and help disseminate them on social media, allowing the messages to have more meaning and influence than if they came from adults. Educating young people on howto talk about suicide safely online has multiple benefits, and there is no evidence of it causing distress (Robinson et al., 2017).
+Some researchers have suggested using new media as a way to provide therapy to young people in a medium they are comfortable with. A randomized controlled trial is underway to study the effectiveness of a safety plan via mobile app in reducing suicidal ideation and behavior versus the traditional safety plan on paper that is used by many institutions before discharging suicidal patients from the emergency room or inpatient psychiatric units (Andreasson et al., 2017). More research is needed on the effectiveness of Internet-based cognitive behavioral therapy (CBT) in reducing suicidal behaviors among school students, but early results indicate that young people were engaged and experienced reduced suicidal ideations and other mood symptoms (Hetrick et al., 2017). Educational websites targeting young people in crisis have demonstrated suicide preventive effects and have been effective in increasing suicide-related knowledge (Till, Tran, Voracek, & Niederkrotenthaler, 2017).
+11	| RECOMMENDATIONS
+Overall, the literature has shown that there are various ways in which new media may facilitate, encourage, or prevent suicidal behavior.
+Several recommendations have been made on how to mitigate the potential negative influences posed by the Internet. One view focuses on increasing regulation of the material that young people are exposed to, via either parental controls, government regulation, or social media platform policies regarding suicide-related content. Some have suggested penalizing media for glamorizing tragic events, or reporting pro-suicide website users to police (Klein, 2012; Marchant et al., 2017).
+Other strategies attempt to exploit the positive influences that new media can have on suicidal behavior and contagion, including crisis support, reduction of social isolation, delivery of therapy, and outreach. There are multiple websites that can provide resources and help for these vulnerable populations, including those run by organizations such as the AFSP (2018), Suicide Awareness Voices of Education (SAVE, 2018), and the Suicide Prevention and Resource Center (2018). In addition, the Trevor Project is an organization that is especially helpful for LGBTQ youth. They provide crisis interventions via phone, online chat, or text messaging that connect LGBTQ youth with counselors who are trained to recognize the special challenges and vulnerabilities faced by this population. They also provide suicide prevention training and other community resources for anyone affected by suicide, or who may be interested in helping this specific community (Trevor Project, 2017).
+In addition, online forums can provide safe, anonymous ways for people to connect with mental health professionals or support groups. Furthermore, covering suicide carefully, even briefly, can change public misperceptions and correct myths, which can encourage those who are vulnerable or at risk to seek help (Klein, 2012). Given the popularity of self-harm videos, some have suggested developing videos that focus on health and recovery (Marchant et al., 2017). Improved identification of clinically significant suicidality on social media sites can make it relatively effortless to send users a private message with information and links to resources such as hotlines, support groups, or websites with reliable psychoeducation (Braithwaite et al., 2016).
+Finally, it is important for mental health professionals to remember that they too can play a role in mitigating the harmful effects of traditional and new media on suicidal behaviors and suicide contagion, while also maximizing the positive influences. The American Psychological Association (APA) released guidelines for mental health professionals in 2003 to try to help mitigate the effect of suicide contagion. They include encouraging responsible reporting of suicide if interviewed by media, referring reporters to responsible reporting guidelines, giving fact-oriented quotes, stressing that treatable mental illness may underlie many suicides, and to write to newspapers and television news outlets about misguided reporting on suicides (Smith Bailey, 2003). In addition, mental health professionals should be familiar with appropriate websites and resources to recommend to clients and families who may benefit from preventive information (Perry, Werner-Seidler, Calear, & Christensen, 2016). The Centers for Disease Control and Prevention (CDC) has a useful website that publishes web-based and printable materials with up-to-date suicide prevention strategies, effective programs, and the latest research on preventive interventions (CDC, 2017).
+Mental health professionals should also become more comfortable with asking patients about their social media use as part of their routine assessments (Marchant et al., 2017). When working with patients who use the new media extensively and preferentially, the mental health professionals may even wanttoexplore ways of integrating itinto their practice. However, if a mental health professional opts to use new media to correspond with patients, it would be prudent to obtain advice from their malpractice carrier about ways to minimize risk. In addition, they should also, at a minimum, seek out and regularly review their professional association's ethical guidelines regarding the use of new media, and engage patients (or their parents, if the patient is a minor) in an informed consent discussion about the risks and benefits of using new media for clinical purposes before doing so, as well as periodically throughout its use.
+12	I CONCLUSION
+Despite much variability, research largely supports the idea that traditional and new media reporting on suicide can, and does, have a significant impact on suicide rates. Questions remain regarding the impact of specific components of “the Werther effect”; the complexity of the numerous factors related to characteristics of the initial suicide victim, quality and type of reporting in different forms of media, and traits of subsequent suicide victims, including their individual risk factors for suicide, make research in this area especially challenging. Nonetheless, over the last few decades, reporting recommendations have been developed to guide news organizations and other media entities in tailoring their publications in a manner that may reduce any potential harmful effects to vulnerable populations.
+The development of the Internet, and the successive astonishingly rapid growth and popularity of websites, blogs, social media, and other forms of instantaneous world-wide communication, has presented new challenges in the fight to prevent suicide. Parents of young children, adolescents, and young adults need to be aware of the potential risks involved with exposure to suicide-related content, and be ready and willing to monitor their children's media consumption, as well as discuss any concerns that may arise from it. Public health organizations, mental health professionals, media groups, and possibly even lawmakers, need to prioritize the monitoring of media and communication technology innovations, in addition to tracking and containing any dangerous new trends in suicidal behaviors. Mental health professionals and educators should be well versed in the latest literature on the topic of reputable and effective suicide prevention resources and strategies,1 as they have a responsibility to share accurate and quality information with at-risk individuals and their loved ones. Efforts are also needed to advocate for additional research on suicide, and for providing sufficient and relevant mental health resources to vulnerable populations.
+Finally, in addition to the public health implications associated with the issue of media's influence on suicidal behavior and contagion, there are also implications relevant for the forensic clinician. Forthose performing psychological autopsies in suicide cases, in this technological golden age, consideration of the potential contagion of specific ideas about suicide through new media should not be overlooked. Familiarity with various aspects of this issue may become pertinent in investigating specific suicide cases, especially in more susceptible populations such as adolescents.

Brief Reports.txt

@@ -0,0 +1,28 @@
+(adjusted odds ratio [AOR]=3.04, 95% confidence interval [CI]=2.84— 3.26), a history of suicide attempts (AOR=2.77, CI=2.64-2.90), depressed mood (AOR=1.69, CI= 1.62-1.76), and nonalcoholic substance abuse or dependence (AOR= 1.13, CI= 1.07-1.19). Conclusions: For nearly a third of all suicide decedents, better mental health care might have prevented death. Efforts to reduce access to lethal doses of prescription medications seem warranted to prevent overdosing with commonly prescribed substances. (Psychiatric Services 65:387-390, 2014; doi: 10.1176/ appi.ps.201300124)
+In 2010, suicide accounted for approximately 38,000 deaths in the United States, corresponding to a suicide rate of 12.43 per 100,000 individuals (1). Although mental health treatment helps reduce suicidal behavior (2), each year an estimated 30% of suicide decedents will have received treatment within one month of their death (3). This fact suggests that providers may have opportunities to improve suicide prevention efforts. Compared with suicide decedents who did not receive mental health treatment, those who received treatment often had more severe symptoms (4). Research is currently scarce on the co-occurring health- and life-stress-related circumstances among suicide decedents who received treatment. Life events that are considered rele
+vant factors to suicidal behaviors (5) are routinely documented in the National Violent Death Reporting System (NVDRS) but have not yet been investigated in relation to mental health treatment before suicide. The objective of this explorative study was to assess associations between recent mental health treatment and circumstances of death among suicide decedents to better understand the unique qualities of individuals who had received mental health treatment and to help inform suicide prevention efforts.
+Methods
+We obtained 2005-2010 data from the NVDRS, which captures details on violent deaths among the deaths registered within each of 18 states (Alaska, Colorado, Georgia, Kentucky, Maryland, Massachusetts, Michigan, New Jersey, New Mexico, North Carolina, Ohio, Oklahoma, Oregon, Rhode Island, South Carolina, Utah, Virginia, and Wisconsin). Data for Michigan and Ohio were available for only 2010. Data sources for NVDRS include death certificates, law enforcement reports, and coroner and medical examiner reports; these sources are used to more comprehensively describe each violent incident. Suicide deaths are identified according to the manner of death recorded in the various data sources. State abstractors follow a strict coding manual to ensure consistent reporting and reconcile any differences across the data sources (6).
+Adult suicide decedents who received mental health treatment
+within two months before death were compared with suicide decedents who were not known to have received mental health treatment shortly before death. Because help-seeking behavior among adolescents differs from behavior among adults, only decedents over age 18 were considered (7). Treatment was defined as seeing a psychiatrist, psychologist, general medical doctor, therapist, or other counselor for a mental health or substance misuse problem; receiving a prescription for a psychiatric medication; attending anger management classes; or residing in an inpatient or halfway house facility for mental health problems (6).
+To qualify as suicide by poisoning, a substance had to be ingested and deemed coresponsible for the death. Drugs on the scene that were not ingested were not counted (6). Suicides by poisoning were coded as poisoning involving commonly prescribed substances if one or more of the substances used in the act was technically a controlled substance that would require a prescription (6).
+Suicide decedents were compared with respect to sociodemographic characteristics, health- and stress-related characteristics, and the suicide method involved. Logistic regression was used to calculate odds ratios for receiving treatment before suicide for all above characteristics. We adjusted comparisons for age, sex, raceethnicity, and history of suicide attempt. History of suicide attempt was adjusted only for age, sex, and raceethnicity. Because of multiple testing, we set the level of statistical significance to #.001. We performed analyses using PASW Statistics 18. This study was determined to be exempt from human subjects review by the Institutional Review Board of the Centers for Disease Control and Prevention.
+Results
+Of the 57,877 suicides among persons >18 years of age recorded in NVDRS between 2005 and 2010, 16,471 (28.5%) had received treatment within two months of suicide. Of those who did not receive treatment in the two months before suicide (N=41,406), 3,198 (7.7%) had received mental health treatment in the past. Being male (adjusted odds ratio [AOR]=.47),
+race-ethnicity other than non-Hispanic white (AORs=.61-.73), and being ages 19-49 (AOR=.69-.91) or $70 (AOR=.63) were all associated with lower odds of receiving treatment (Table 1). Among life events registered in the NVDRS, intimate partner problems were the most prevalent type of problem before suicide and affected 15,168 (30.3%) of all 50,024 decedents with known circumstances (Table 1).
+Compared with persons who died from hanging, those who died by drug poisoning involving a substance that commonly requires prescription (AOR=3.04), by sharp instruments (AOR=1.30), or by falling or jumping (AOR=1.44) had higher odds of recent mental health treatment. Suicides by firearms were associated with lower odds of receiving treatment (AOR=.88) (Table 1). Among 3,758 persons who received treatment and died by poisoning involving commonly prescribed substances, 3,060 (81.4%) were tested for use of antidepressants at the time of death, with 2,278 of them (74.4%) testing positive.
+Among the decedents, having recent mental health treatment was positively associated with having depressed mood at time of death (AOR=1.69), a history of suicide attempt (AOR=2.77), and substance use problems other than alcoholism (AOR=1.13) (Table 1). Receiving treatment was inversely associated with having intimate partner conflicts (AOR=.75), perpetrating interpersonal violence (AOR=.64), fi-nancialproblems (AOR=.87), criminal legal problems (AOR=.60), other legal problems (AOR=.82), and homelessness (AOR=.66).
+Discussion
+Nearly a third of suicide decedents received help from some type of mental health care provider before taking his or her own life. Earlier studies have found a similar proportion of service utilization, which was even higher when general health care services— particularly, primary care provider visits—were also taken into account (3). The demographic distribution among suicide decedents known to have received mental health care in the two months prior to death generally reflected patterns of mental health seeking in the general population, in that
+smaller proportions of males, persons of minority race-ethnicity, individuals #30 years, and older adults ($70 years) were known to access mental health services before suicide. Genderspecific help-seeking behavior, stigma, and socioeconomic factors often play a large role in these treatment disparities (8). However, when controlling for age, race-ethnicity, sex, and history of suicide attempt, we still found that some health- and life-stress-related circumstances were more common among decedents who had sought treatment, which indicates an area for improvement in the delivery of mental health services.
+Depressed mood and substance misuse were associated with receiving mental health treatment. Although the effect size for substance misuse was relatively small, this association is consistent with research showing that patients were treated more often for depression when comorbidities were present (9). However, we also found that many suicide decedents who killed themselves by drug poisoning had received mental health treatment before their suicide, and commonly prescribed substances were often involved in these deaths. There is common agreement that drugs should be prescribed only in small package sizes to at-risk individuals to prevent suicide (10). In Britain, reduced pack size of analgesics have been shown to be effective in reducing suicides with paracetamol (acetaminophen in the U.S.) (11). More research is needed to investigate substances used and their responsibility for fatal outcomes among decedents, mechanisms involved, and best prevention practices. In contrast to suicide decedents who used poison, those who used firearms were less likely to have received treatment before death. Firearms are considered one of the most violent and lethal methods of suicide, and the use of violent methods has been described as reflecting a further step in the suicidal process. Individuals choosing firearms as their method may be less inclined to seek or accept treatment (12).
+The higher odds of receiving mental health treatment observed among persons with a history of suicide attempts underscore that mental health treatment can provide an opportunity to address the needs of some previous
+suicide attempters. More follow-up treatment, therapies tailored to specifically reduce self-directed violence (including cognitive or other therapies and strategies intended to improve
+coping skills to better handle risk factors associated with suicide [13]), and monitoring of prescription medications might reduce the risk of subsequent attempts.
+Connecting mental health providers to other services relevant to the circumstances frequently seen among decedents may also help prevent some suicides. Some life events, particularly
+intimate partner problems, were prevalent for more than 15,000 of all suicide decedents, reflecting sociological concerns with intimacy, including marriage and the association of divorce with high suicide rates (14). However, decedents experiencing partner problems had lower odds of receiving treatment before suicide. Clinicians and other public health professionals may be able to collaborate with successful programs and strategies that involve friends or family of the at-risk individual in order to reach out to individuals affected by family problems (15).
+This study had several limitations. NVDRS data do not indicate which type of mental health service was received. Different quantities and treatment types were subsumed as mental health treatment, and findings for specific treatments may be different. The data are not nationally representative but representative of only the 18 states participating in the NVDRS. The information provided on the circumstances of deaths was from proxies and was subject to recall bias. Further, we could not assess whether substances that commonly require prescription were actually prescribed, because drugs were assigned to the prescription drug category only on the basis of the substance name or the name of the metabolite identified. Some of these substances might have been acquired on the street. Even in cases in which the drug was actually prescribed, we cannot rule out that the prescription may have been prescribed for a person other than the decedent.
+390
+PSYCHIATRIC SERVICES ♦ ps.psychiatryonline.org ♦ March 2014 Vol. 65 No. 3
+Conclusions
+The findings suggest that the substances used in suicides by poisoning and efforts to reduce access to lethal doses of prescription medications warrant further research. Further, better collaboration between mental health service providers and providers of other services, including outreach to individuals with intimate partner problems, may help reduce suicide deaths.

COVID-19 effect on mental health patients and workforce.txt

@@ -0,0 +1,9 @@
+Early career psychiatrists are crucial in the medical response to COVID-19. Although we are ready to provide help to those in need, we are made to count on insufficient access to WHO-standard personal protective equipment and training when trying to safely support others' mental health face-to-face. Furthermore, feelings of uneasiness or ill-preparedness arise when countries start redeploying mental health-care professionals to general medical care for patients with COVID-19 in overwhelmed health-care systems (table and appendix).
+Telepsychiatry (ie, providing mental health care remotely, using telecommunications such as telephone or video conferencing tools) in several settings is suddenly being introduced or massively expanded to serve patients with pre-existing disorders, health professionals on the frontline, and the general population, during a time of uncertainty, misinformation,
+and physical distancing.4 Still, telepsychiatry is scarce in several low-income and middle-income countries, posing challenges for health-care workers and patients where face-to-face care is not safe because of the risk of virus infection. We also perceive that attention given to the public's mental health during the outbreak came late, and overlooked vulnerable populations, such as refugees, people without secure housing, people living in overcrowded spaces, and patients with severe psychiatric disorders.
+Apart from disrupting usual mental health care, the COVID-19 pandemic could lead to further psychological trauma. The huge toll such trauma can take on medical professionals, which can include delusional episodes and suicidality, in countries as deeply struck by COVID-19 as Italy is of particular concern. Psychiatric sequelae could be reduced by the early involvement of mental health professionals in drawing up comprehensive public
+The coronavirus disease 2019 (COVID-19) outbreak has raised several concerns regarding its mental health effect on patients with psychiatric disorders and the health-care workforce.1,2 Worldwide, psychiatrists are navigating a fast, unpredictable tempest, in developing plans to respond to their own mental health needs and those of their country's population.
+We are a group of 16 early career psychiatrists connected by the Early Career Psychiatrists Section of the World Psychiatric Association,3 working across different WHO regions in countries (other than China) that have been severely affected by COVID-19. The pandemic led us towards a collective endeavour to share our country-specific experiences, plans, and concerns.
+health policies and in supporting the health-care workforce.
+Many early career psychiatrists are part of the millennial generation familiar with technology,5 and are channelling this strength to deliver far-reaching telepsychiatry, share online mental health-promotion resources, and connect with colleagues worldwide. Thanks to social media and the internet, international associations of early career psychiatrists are providing educational resources (eg, real-time news, journal clubs, and webinars), and group emotional support for peers. Colleagues in countries with a recent history of humanitarian and public health crises (eg, the epidemics of Zika virus disease in the Americas and Ebola virus disease in Africa), bring their experience of providing mental health care during and after such disasters, and those in countries with an earlier onset of the COVID-19 outbreak share the lessons already learned there. The spontaneity, resilience, and solidarity with which many colleagues have joined forces is inspiring.
+Early career psychiatrists are an essential resource in the mental health management of the COVID-19 pandemic and its aftermath. Mental health authorities are called to count upon early career psychiatrists, warranting the training and resources to enable us to safely and effectively work for our patients, colleagues, and communities. We express our gratitude to all early career psychiatrists taking risks to care for their patients, and we invite them to seek peer support and join forces both locally and across the world.

CT Meta-analysis Working Paper with Appendix 1.txt

@@ -0,0 +1,74 @@
+1 Introduction
+Cash transfers (CTs) - commonly understood as direct payments made to people in poverty - are among the most extensively studied and implemented interventions in low- and middle-income countries (LMICs) (Vivalt, 2015). Previous systematic reviews and meta-analyses of CTs found improvements on several outcomes. These outcomes include material poverty (Kabeer & Waddington, 2015), human capital (Baird et al., 2013b; Millan et al., 2019), social capital (Owusu-Addo et al., 2018), health (Lagarde et al., 2007; Behrman & Parker, 2010; Crea et al., 2015), intimate partner violence (Baranov et al., 2020; Buller et al., 2018), child labor (Kabeer & Waddington, 2015), the spread of HIV (Pettifor et al., 2013), spending on tobacco and alcohol (Evans & Poponova, 2014; Handa et al, 2018), and labor supply (Baird et al., 2018; Banerjee et al., 2017).
+Although these factors are relevant to wellbeing, measures of mental health (MH) and subjective wellbeing (SWB), which probe how individuals themselves assess the quality of their lives, are often thought to track wellbeing more accurately. Indeed, measures of SWB are increasingly considered to be essential components in applied policy analyses (Benjamin et al., 2020; Frijters et al., 2020). It therefore seems pertinent to evaluate the effectiveness of CTs with respect to these measures.
+Individual income and SWB are known to be positively associated (Powdthavee, 2010; Stevenson & Wolfers, 2013; Jebb et al., 2018), especially for those at low income levels (Clark, 2017; Deaton, 2008). A similar relationship is observed in the MH literature (Karimli et al., 2019; Tampubolon & Hanandita, 2014; Schilbach et al., 2016; Ridley et al., 2020). Moreover, mental health problems may engender and perpetuate poverty (Haushofer & Fehr, 2014). Unfortunately, the literature on the link between income and SWB and MH in LMICs has long lacked causal evidence, which the growing body of primary research on CTs may address.
+While CTs may improve the SWB and MH of recipients, these interventions could also have negative psychological consequences on non-recipients. Qualitative research suggests the presence of negative psychological spillovers (Fisher et al., 2017; MacAuslan & Riemenschneider, 2011), and some recent quantitative work echo this worry (Haushofer et al., 2019). For example, envy among non-recipients may be a concern (Ellis, 2012). Community disruptions and crime rates may also increase if CTs are mistargeting to formally ineligible recipients (Agbenyo et al., 2017; Fisher et al., 2017). However, there is also some evidence of positive spillovers. For example, CTs have been found to decrease the intergenerational transmission of depression (Eyal & Burns, 2019) and to lead to decreased suicide rates in the areas they are implemented (Alves et al, 2018).
+We know of no previous systematic reviews on this subject. A non-systematic meta-analysis by Ridley et al. (2020), which evaluates the impact of CTs on MH, is closest to our work.1 We build on their work in four directions. First, we conducted a full systematic review and search of the existing literature in accordance with the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidance (Moher, Liberati, Tetzlaff, & Altman, 2010). Second, we consider SWB measures alongside MH measures2. Third, we consider quasi-experimental designs (in addition to randomised controlled trials (RCTs)). Fourth, we evaluate the quality of included studies, assess publication bias, and perform a moderator analyses across (1) outcome type (MH and SWB), (2) CT value, and (3) duration of the transfer.
+Methods
+2.1	Eligibility criteria
+For a study to be included it must satisfy four criteria: First, the study must investigate the effect of an unbundled cash transfer (defined below). Second, the study must include a measure of self-reported affective mental health or subjective wellbeing, but these need not be the primary focus of the study. Third, the study context must not be a high-income country.3 Fourth, the study design must be experimental or quasi-experimental4 and afford standardizing the mean difference between treatment and control groups.
+Regarding our first criterion, we distinguish between unconditional cash transfers (UCTs) and conditional cash transfers (CCTs). Conditional cash transfers formally require adherence to certain actions, such as school enrollment or vaccination. The strictness of conditions varies widely, and conditions are sometimes left unmonitored due to high administrative costs (Davis et al., 2016). UCTs have no requirements, although they are often targeted to a vulnerable subset of the population, commonly defined by a combination of regional statistics, means tests and selection by prominent members of the community. We consider noncontributory social pensions and enterprise grants to be UCTs. CTs are typically paid out in lump-sums or streams (monthly installments). Some stream or multi-installment CTs have graduation mechanisms where individuals stop receiving transfers once they meet certain conditions (Villa & Nino-Zarazua, 2019). All included CTs must be “unbundled”, i.e. implemented and tested independently of other services such as asset transfers, training, or therapy.
+Concerning our second criterion, we note that SWB measures tend to assess overall wellbeing (Diener, 2009; Diener et al., 2018), which sometimes include separate measures of positive and negative mental states (Busseri & Sadava, 2011). By contrast, affective MH questionnaires tend (1) to only measure the negative components of SWB, i.e., how badly someone is doing and, (2) to also capture information on an individual's behaviors and habits (in addition to their thoughts and feelings). In our analyses, we include measures of valenced mental states, but no measures of behavior or habits. See the “Measures” column of Table A3 in the appendix for a list of all included measures.
+2.2	Data
+We searched studies using academic search engines and databases. These included: EBSCO: MEDLINE, PsycINFO, PubMed, Business Source Complete, EconLit, Social Sciences Full Text (H.W. Wilson), APA PsycARTICLES, Psychology and Behavioral Sciences Collection, Academic OneFile, Academic Search Premier, CINAHL, Open Dissertations, Web of Science, Science Direct, JSTOR, ECON PAPERS, 3ie, IDEAS/REPEC, and Google scholar. These efforts were complemented by a forward and backward citation search of eligible studies, contacting authors, and through Google Scholar notifications. Our search string can be found in Appendix A.
+We stored all retrieved records in the reference management system Zotero. Double-blind screening of the titles and abstracts was done using the software Rayyan by JM and CK. Any disagreements were discussed until consensus was reached. Studies that passed the double-screening were reviewed in full text by JM.
+We extracted study details such as author name, CT program, number of participants, MH and SWB outcomes, and effect sizes. We also collected information on the size of the cash transfer, time between start of intervention and follow-up, and whether it was a CCT or UCT, paid out in a stream or lump sum, or directed towards adolescents, prime age adults or elders. All data were extracted by one author (JM) and the full extraction results were checked for accuracy by CK and ABM.
+2.3	Quality
+To assess the quality of included research, we evaluated the following domains: causal identification strategy, pre-registration, balance between treatment and control groups, attrition, sample size, contamination, treatment compliance, and whether intention-to-treat (as opposed to a complete case) analyses were performed.
+2.4	Statistical Methods
+We used the statistical programming language R for data analysis. Since most RCTs and quasi-experimental designs are based on mean differences,5 we standardized these using Cohen’s d. We used the independent t-statistic from a test of the mean difference to calculate Cohen’s d in nearly all cases. We use d = t'1/n! + 1/nc where n! - treatment sample size and nc - control sample size (Goulet-Pelletier & Cousineau, 2018). If the effect size of a study was expressed via odds ratios (n = 2), we converted from odds ratios to Cohen’s d using d = ln(01)V3/5.6
+If a study contained multiple outcome measures, we coded each as MH or SWB. To achieve a single effect size for each study-follow-up combination, we combined outcomes using the method of Borenstein et al., (2009), specifying a correlation of 0.7 for within construct aggregations, 0.5 for between constructs and 0.6 for both within and between aggregations. Specifying different correlations changes only the aggregate standard error, not the mean of effect sizes.
+We used random effects (RE) models for our meta-analysis, which assume that true effects of each included study are drawn from a distribution of true effects (Borenstein et al, 2010). Each study in our model was weighted by the inverse of the standard error of the study’s estimated effect size. Since there are sometimes multiple follow-ups in a study and multiple studies in a sample or program, we clustered standard errors at the level of the study and program. We assessed evidence of publication bias and p-hacking by using a funnel plot, the Egger regression test (Borenstein et al, 2011), and a “p-curve” (Simonsohn et al., 2014).
+We conducted meta-regressions to test if certain study characteristics moderated estimated effect sizes. We focused on three potential moderating variables: years since CT began, size of CT, and whether CTs had conditionality requirements.
+Concerning size of CT, we considered both the absolute and relative CT size. We operationalized absolute size as the average monthly value of a CT in purchasing power parity (PPP) adjusted US 2010 dollars, with lump sum CTs (comprising about 25% of our sample) divided by 24 months, which is the mean follow-up time.7 For relative size, we used monthly CT value as a proportion of previous
+household monthly income. This was either directly reported or easily derived in many studies (21 out of 37 studies). If a study did not report sample information on income, we used consumption (10 studies) or expenditure (3 studies) information as a proxy. To convert between individual income and household income (8 studies) we assumed that household income = individual income * ^household size (see Chanfreau & Burchardt, 2008). If there was insufficient information to impute average household income (4 studies), we used regional statistics. Finally, as a robustness test, we also computed yearly CT value as a proportion of annual gross domestic product per capita (GDPpc).
+3	Results
+3.1	Description of Studies and Quality
+We retrieved 1,870 records from implementing our search string. After removing duplicates, we were left with 1,147 records. After an initial round of double screening titles and abstracts by JM and CK, 143 met the eligibility requirements (see Figure 1 for a diagram of selection flow). After JM performed the final round of screening, there were 32 unique studies drawn from the initial search and five from Google Scholar alerts and citation searches. We thus included a total of 37 studies8 reporting on 100 outcomes. Table A3 in the appendix summarizes the key characteristics of the included studies. Of the outcomes, 46 measured depression or general psychological distress, 21 measured happiness or positive feelings, 18 measured life satisfaction and two measured anxiety. The remaining 13 were summary indices of MH, SWB, or both.
+Most of the studies were conducted in Africa (23), followed by Latin America (10) and Asia (4). The most commonly investigated CT type was UCT (26; 19 plain, 6 pensions and 1 enterprise grant) followed by CCTs (10) and one study that contained both a CT and UCT (Baird et al., 2013a). Country context
+was relatively evenly divided into low, low-middle, and upper-middle income countries (see Figure A2 in the appendix). Over half of the included studies included random assignment (22), while the rest were quasi-experimental (15).9 The average time from the start of the CT to follow-up was two years. The average monthly payment was $38 PPP. A quarter of the studies were implemented as predominantly lump sum (10). All other studies (27) were paid out on a monthly basis.
+In Table 1, we list the results of our quality assessments. While blinding of participants is impossible for CTs, blinding personnel and outcome assessment was mentioned (but not performed) in only one study (McIntosh & Zeitlin, 2020). Overall, few studies (9/37) referred to pre-registered protocols. The adherence to pre-specified statistical procedures and outcomes was generally unclear, thus making it
+impossible to assess whether outcomes were ‘cherry-picked’ post treatment. Moreover, about half of the included studies (17/37) did not assess treatment compliance. Therefore, aspects relating to implementation (e.g. intervention fidelity and adaptation) could not be assessed (Moore et al., 2015). Furthermore, contamination by the CT on control groups was rarely discussed or addressed. Only 13 out of 37 studies were geographically-clustered RCTs (cRCTs), which are more robust to possible contamination effects. Of the 15 quasi-experimental studies, one used a natural experiment (Powell-Jackson et al., 2016), two used instrumental variables (Ohrnberger et al., 2020a; Chen et al., 2019), and four used a regression discontinuity approach (based on a means test). The eight remaining studies used a propensity score matching approach. Of those using propensity score matching, six also employed a difference-in-difference estimator.
+Despite the aforementioned concerns, we assess the synthesized evidence to be fairly reliable. Importantly, most studies clearly explained their causal identification strategy, were well balanced, performed intention-to-treat analyses, and controlled for differential attrition when present. Sample sizes were generally large compared to common sample sizes in clinical or psychological studies (n<500; Billingham et al., 2013; Kuhberger et al., 2014; Sassenberg & Ditrich 2019).
+3.2	Baseline results
+For our baseline results, we aggregated effect sizes across studies using a random effects model. Throughout our analyses, we omitted measures of stress, optimism, and hope, and one outcome reported from Galama et al. (2017), which was a clear outlier.10 The average overall effect size, as indicated by a black diamond at the bottom of Figure 2, is 0.10 SDs in the composite of SWB & MH measures (95% CI: 0.08, 0.12; given by the width of the diamond). The overall effect size does not
+Cohen's d
+Note: Forest plot of the 37 included studies. Subjective wellbeing (SWB) and mental health (MH) outcomes in each study are aggregated with equal weight. Mo. after start is the average number of months since the cash transfer began. $PPP Monthly is the average monthly value of a CT in purchasing power parity adjusted US 2010 dollars. Lump sum cash transfers were converted to monthly value by dividing by 24 months, the mean follow-up time.
+change substantially when accounting for dependency between multiple follow-ups, and multiple studies in a program in a multilevel model (ES: 0.095, 95% CI: 0.071, 0.118, or if we combine all the outcomes, without first averaging at the study-follow-up level (ES: 0.091, 95% CI: 0.066, 0.116.
+Heterogeneity, as calculated by the D2index, is substantial; 63.7% of the total variation in outcomes is due to variation between studies.11 In other words, 63.7% of total variability can be explained by variability between studies instead of sampling error. To account for the impact of this substantial heterogeneity, we calculate a 95% predicted interval.12 The estimated 95% prediction interval, given by the dashed line bisecting the black diamond in Figure 2, suggests that 95% of similar future studies would be expected to fall between 0.001 and 0.201 SDs in our composite of MH and SWB.
+Figure 3 displays the risk of publication bias and “p-hacking” (researchers testing a high number of outcomes and cherry-picking the coefficients that fall below a threshold p-value). In Figure 3a, we show a funnel plot, with standard error plotted against effect size, and the mean effect shown as a black vertical line.13 If there are significantly more studies to the right than the left of the mean effect size, this would suggest that studies on the left may be missing, possibly indicating publication bias. This is known as asymmetry. Figure 3a shows little asymmetry, indicating that studies with more positive effects appear no more likely to be published. We use Egger’s regression test to check this quantitatively by regressing the standard error on the effect size. The test does not reject the null of funnel plot symmetry (p=0.549), supporting our reading of the plot.
+Figure 3b shows the percentage of results with different p-values. If “p-hacking” were an issue, we would expect that the distribution of p-values is left-skewed (an upward slope in the figure). The p-curve is downwardly sloped, which suggests no widespread p-hacking. However, it is possible that regression specifications with insignificant dependent variables were not reported at all. P-curves are unable to address such scenarios (Bishop & Thompson, 2016).
+3.3	Meta Regression and Moderator Analysis
+We focus on three types of variables that we expect to moderate the observed effects: (1) Whether a CT had conditionality requirements or not. (2) Value of CT (in absolute terms and relative to previous income). (3) Years since the transfer began, allowing us to assess whether effects dissipate over time. Throughout, we use multi-level models that account for multiple outcomes in a follow-up, multiple follow-ups in a study and multiple studies in a sample or program. Standard errors are clustered at the study and program level.14 In every specification presented, the dependent variables are the study’s estimated effect on MH or SWB. We standardized the effect sizes into Cohen’s d.
+In Figure 4, we present six plots that illustrate the bivariate moderating relationship of our variables of interest. Panel (a) shows the distribution and average effect size for UCTs and CCTS. Panels (b) through (f) show effect size on the y-axis and the time or size on the x-axis. Plots (b) through (f) are simple scatter plots meant to illustrate the raw correlation between two variables.
+In Table 2, we present our main results. All models include a measure of CT size and years since the CT began. Model 1 includes a dummy indicating whether the CT had conditionality requirements. Models 1, 2 and 3 estimate the effect of relative CT size. Models 4 and 5 estimate the effect of absolute CT size (using $PPP monthly value). Models 3 and 4 include an interaction term between payment mechanism and “years since CT began” to identify the effect of decay conditional on whether a CT was paid out in a lump sum or stream.
+In Model 1 we find that conditionality requirements reduce estimated effect sizes by almost 50%. In so far as UCTs are less costly to administer than CCTs, this suggests that UCTs are likely to be more efficient in promoting recipients’ wellbeing.
+In Model 2 we omit the indicator of whether CTs where CCTs or UCTs. Based on this specification, one can expect that doubling a recipient’s consumption (by receiving a CT 100% of previous consumption) to roughly lead to a 0.10 SD increase in MH/SWB at the average follow-up time. Results in Models 1 and 3 are similar. See panels (e) and (f) of Figure 4 for the correlational relationship between relative size of a CT and magnitude of effect.
+Models 4 and 5 shows our results for absolute CT value, yielding a significant and positive coefficient in both specifications. These results indicate that a CT with a monthly value of $100 PPP leads to an approximately 0.07 to 0.08 SD increase in SWB and MH outcomes. See Figure 4, panel (c) for the bivariate relationship. Increases in income are typically assumed to yield diminishing gains in wellbeing. To test if that is the case in our sample of studies, we log transformed our measures of relative and absolute CT size. We find a significant effect for log-relative value but no significant effect of logabsolute value (see Table A2 in the appendix).15
+Taken together, models 1, 2 and 4 provide evidence that the effect of CTs on wellbeing decays over time. Using the coefficient from Model 2, each year the effect is estimated to decline by 0.015 SDs. With that estimate, a CT which doubles household income would take almost two decades to decay.16 However, the effects of “years since CT began” could differ depending on whether the recipient was given the CT in a lump sum or still receives monthly transfers. Our bivariate plot (Figure 4, panel (b)) suggests a difference in decay between the two payment mechanisms. Lump CTs appear to decay over time while stream CTs (which are nearly all ongoing at the time of the last follow-up) show a flat trend. In Models 3 and 4 we formally test for differences in decay between lump and stream CTs. The interaction, “years since * CT is lump sum” gives the difference in decay between lump and stream CTs. Since stream CTs are ongoing, we expected lump CTs to exhibit a larger decay in effect size than streams. Surprisingly, this is not the case in models 3 and 4. These display a positive, albeit insignificant interaction term. Thus, although there is a significant overall decay in effect size (as indicated by Models 1, 2, and 5), we are unable to precisely estimate the effect over time for a specific payment type.
+Finally, we note that seven studies in our study include multiple follow-ups. As shown in Figure A1 in the appendix, six of these show a decline in effects size across follow-ups. A repeated t-test of whether mean effect size is different between first and second follow-up yields a p-value of 0.007, indicating that this decline is statistically significant.
+The relatively large and significant intercepts in Table 2 suggest that CTs could have an effect independent of the size of the cash transfer (i.e., an effect from being enrolled). An enrolment effect, however unintuitive, is not implausible. Being awarded an amount of cash might boost someone’s sense of good fortune, which could explain the intercept. Another explanation for the intercepts is that they are an artifact of a concave relationship between CT size and effect. A linear model will generally overestimate the intercept on data that contains a true concave relationship. However, the insignificance of the log-transformed absolute CT value is evidence against a clear concave relationship (see appendix Table A2, Model 2).
+In addition to these analyses, we also tested whether RCT design, type of measure, or the study context moderated the effect size (see Table A1 in the appendix). Whether a study uses a RCT design does not affect the magnitudes of the estimated effects of CTs. This suggests that studies which rely on natural experiments or other causal identification strategies are reasonably robust. However, we do find that, compared to pure MH measures, effects of CTs on measures of SWB are significantly larger. Moreover, the largest effect sizes occur for studies in which a compound index of both MH and SWB was used.17 Notably, CTs conducted in Latin America have a near zero estimated effect. This appears to be primarily driven by the fact that many CTs in Latin America have conditionality requirements. When including both a dummy for conditionality and for the CT being conducted in Latin America, we find that the coefficient on Latin America is roughly halved and significant at the 10% level only.
+As discussed in section 2, we ran alternative specifications of our size variables (see appendix Table A2). In particular, we checked if using CT value relative to GDP per capita changes our results. Although the coefficient is somewhat larger compared to results presented in Table 2 (with p<0.05), our conclusions remain unaffected.
+Finally, in appendix D we consider how our type of results could potentially be used in policy analyses to study cost-effectiveness. Specifically, we calculate how many “wellbeing-adjusted life years” (see De Neve et al. 2020, Frijters et al. 2020), a given type of cash-transfer could buy for a given transfer size. We find that 1000$ lump-sum payment may be expected to buy roughly 0.330 “wellbeing-adjusted life years”.
+3.4	Spillovers
+Four RCTs (two with multiple follow-ups) in our sample enabled assessment of spillover effects on non-recipients of CTs by including two control groups in a geographically-clustered RCT design: a spillover control made up of non-recipients living near recipients, and a “pure” control comprising non-recipients living spatially separate from the treatment locations.18
+This design allowed comparison of wellbeing across (a) non-recipients who are “treated” to a spillover effect by living near recipients to (b) recipients living further away (who form the “pure” control). To ascertain the average effect of spillovers we performed a meta-analysis of the observed effects, using a multilevel random effects model, inverse-weighted by study standard error, and errors clustered at the level of the sample. Our results are illustrated in Figure 5.
+The average effect of CTs on non-recipients’ MH and SWB (represented by the diamond), is close to zero and is not significant at the 95% level, suggesting no significant spillover effects on average.
+4	Discussion
+Our results represent a systematic synthesis and meta-analysis of all the available causal evidence of the impact of CTs on mental health and subjective wellbeing in low- and middle-income contexts. In sum, we find that CTs, on average, have a positive effect on MH and SWB indicators among recipients. More precisely, we find an average impact of about 0.10 SDs. Additionally, we observe that the effects
+RE Model	♦	-0.01 [-0.06, 0.03]
+I I I I I I
+-0.3 -0.1	0.1
+Note: A forest plot of the studies in our sample that include MH and SWB spillovers. A random effects multilevel model (with levels for study and sample) with robust standard errors (clustered at the level of the program) shows an effect of -0.01. The 95% confidence interval overlaps with zero. All of the CTs except Baird et al., (2013a) were implemented by GiveDirectly, an NGO.
+of CTs appear to only dissipate slowly over time. The estimated effects were substantially larger for unconditional CTs. Our results were consistent across a battery of robustness tests and the observed effects did not vary according to study design (RCT and quasi-experimental). Notably, our results indicate that CTs are less efficacious in Latin America, which may be explained by the prevalence of CCTs (as opposed to UCTs) in that region. We find no significant evidence of negative spillover effects on non-recipients. However, spillover effects were rarely reported upon (n=4). We therefore encourage more research on this aspect going forward.19
+4.1 Limitations
+Like most meta-analyses, using study averages for moderator variables means that we do not capture within-study variation, which limits the precision of our estimates. Some of our insignificant results may be due to low power. This could be remedied if we had access to the data at the level of the individual. Some of the studies we include have open access data policies (Haushofer et al., 2016; Paxson & Schady, 2010; Ohrnberger et al., 2020a). An individual level analysis may therefore be possible but was outside the scope of this paper. Another limitation arises from the paucity of longitudinal follow-ups. There was only one study in our sample that followed up more than five years after the cash transfer began (Blattman et al., 2020). This limits what we can say about the long run effects of CTs on SWB and MH. There is also only one study that discusses effects of CTs on the SWB and MH of individuals who share a household with recipients.20 Unfortunately, our evidence was limited to spillovers relating to non-recipients in the geographic proximity of recipients.
+An important feature of this meta-analysis is that it does not offer evidence on the mechanisms by which CTs improve SWB and MH. One possible mechanism worth investigating is whether the effect on
+SWB or MH stems from increased consumption relative to one’s peers or from previous levels of consumption. Indeed, there is a rich set of possible mediators and moderators, and we have only analyzed a small subset of them.
+Finally, we know of no other systematic review and meta-analysis which estimates the total effect of an intervention on SWB and MH. This limits our capacity to compare the cost-effectiveness of CTs to other poverty alleviation or health interventions.
+4.2 Implications and suggestions for future research
+Although there is some preliminary evidence that CTs are cost-effective interventions in LMICs compared to a USAID workforce readiness program (McIntosh & Zeitlin, 2020) and psychotherapy (Haushofer, Shapiro & Mudida, 2020), the work done to compare the cost-effectiveness of interventions in terms of SWB and MH is scarce, especially in LMICs. Our meta-analysis contributes to this literature by providing a comprehensive empirical foundation to compare the cost-effectiveness of cash transfers to interventions aimed at improving MH or SWB. Although limited, the practical implications of our meta-analysis are clear: direct cash transfers improve the wellbeing of poor recipients in LMICs.
+There are several research questions to be pursued in future work on subjective wellbeing and mental health. What are the long run (5+ years) effects of CTs? What are the effects on a recipient’s household and community? Relevant spillover data should be collected in RCTs or evaluated in quasiexperiments. The costs of CTs and other poverty alleviation interventions should be published. For instance, since a UCT requires less administration (as there are no conditions to monitor), it seems likely that UCTs are cheaper and, based on our results, more effective than CCTs. However, there appears to be no available evidence to answer this question. More broadly, we recommend a greater inclusion of SWB and MH data in intervention evidence collection efforts such as Aid Grade.21
+5 Conclusion
+Cash transfers have a small22 (d<0.2) but significant and lasting effect on wellbeing with only mild adaptation effects. Although modest in size, if SWB and MH measure wellbeing more directly than other indicators, these reported improvements are an indicator of genuine success. How important CTs are as a means of improving wellbeing depends on their cost-effectiveness relative to the alternatives. Even if effect sizes are small, CTs may nevertheless be among the most efficient ways of improving lives. There is no evidence that CTs have, on average, significant negative spillover effects within the community they are implemented in. However, the evidence on this is scarce, meriting further research on the topic.

Child - 2018 - Thompson - Associations of adverse childhood experiences and suicidal behaviors in adulthood in a U S .txt

@@ -0,0 +1,65 @@
+1 | INTRODUCTION
+Suicide is the second leading cause of death among youth (Centers for Disease Control and Prevention, 2018). Suicide attempts are a significant predictor of suicide deaths, and nonfatal suicide attempts are 25-60 times
+more prevalent than fatal ones. Approximately half a million people were treated in emergency rooms following suicide attempts in 2015 (Centers for Disease Control and Prevention, 2018). Suicide has far-reaching impacts, affecting family members and friends of those who attempt suicide or die by suicide (Feigelman, Ceral, McIntosh, Brent, & Gutin, 2018).
+122^—Wl LEY--------------------------------------------------
+Recent research has highlighted the increased risk for suicidal behavior among those who have experienced certain adverse childhood experiences (ACEs; Choi, DiNitto, Marti, & Segal, 2017; Dube et al., 2001). Below, we review the literature on links between different types of ACEs and suicidality.
+1.1	| Child abuse and neglect and suicide
+Increased risk for suicidality among those who experience child abuse or neglect has been reported in meta-reviews (Devries et al., 2014; Evans, Hawton, & Rodham, 2005; Miller, Esposito-Smythers, Weismoore, & Renshaw, 2013) and has been replicated in numerous studies comprising different types of samples (Afifi, Boman, Fleisher, & Sareena, 2009; Sachs-Ericsson, Stanley, Sheffler, Selby, & Joiner, 2017). In a population sample, 78% of those who had attempted suicide had experienced childhood sexual abuse compared with 16% of those who had never attempted suicide. Approximately three quarters of those who had attempted suicide had experienced childhood physical abuse compared with 30% of those who had never attempted suicide. Further, those who had attempted suicide reported twice as many experiences of childhood emotional abuse than nonattempters (Briere, Madni, & Godbout, 2016). In a large retrospective cohort study using data from adult participants in a health maintenance organization, those who reported having experienced emotional, physical, or sexual abuse were three to five times more likely to have attempted suicide at some point in their lives (Dube et al., 2001).
+1.2	| Parent alcoholism and parent incarceration and suicide
+Data from a large representative sample of adult participants in the National Epidemiological Survey on Alcohol and Related Conditions indicated that those with a family history for paternal or maternal alcoholism were more likely to attempt suicide than those without a history of parental alcoholism (Thompson, Alonzo, Hu, & Hasin, 2017). The aforementioned study with adult health maintenance organization members also found that those who reported having a household member incarcerated were more than twice as likely to have attempted suicide than their counterparts (Dube et al., 2001).
+1.3	| Parental death and suicide
+In a Scandinavian population-based study, children whose parents had died when they were less than 18 years of age were twice as likely to have died by suicide during a 25-year follow-up compared with children matched on age and sex but who had not lost a parent in childhood (Guldin et al., 2015). Data from a Swedish national cohort showed that parental loss during childhood was associated with an increased likelihood of hospital admission following a suicide attempt in young adulthood (Rostila, Berg, Arat, Vinnerljung, & Hjern, 2016).
+1.4	| Family history of suicidality and suicide
+Having a family member attempt suicide or die by suicide is a significant risk factor for suicidal behavior (Guldin et al., 2015; Qin, Agerbo,
+Key messages
+•	Suicide remains a significant public health problem and is the second leading cause of death for adolescents and young adults.
+•	Adverse childhood experiences (ACEs) have been shown to be associated with an increased risk for suicidal ideation and suicide attempts.
+•	Our results, based on a U.S. nationally representative sample, indicated that physical, sexual, and emotional abuse, parental incarceration, and family history of suicidality increased the risk for suicidal ideation and suicide attempts in adulthood.
+•	An accumulation of ACEs was associated with increased odds of both suicide ideation and attempts.
+•	Intervention strategies need to prevent ACEs from occurring and, if they do occur, they should take into account the impact of cumulative ACEs on suicide risk.
+& Mortensen, 2002). Among children of parents with mood disorders, those whose parents had a history of a suicide attempts were five times more likely to attempt suicide than a comparison group whose parents had a mood disorder but no suicide attempt history (Brent et al., 2015). Another study found that males who died by suicide were significantly more likely to have experienced the suicide of another family member than were their male counterparts who were still living (Feigelman, Joiner, Rosen, & Silva, 2016).
+1.5	| Cumulative trauma
+Central to the study of ACEs is the notion that the higher the number of adverse experiences, the greater the likelihood of poor outcomes. The previously cited study of a large sample of health maintenance organization members in San Diego found that for each additional adverse event, the risk for making a nonfatal suicide attempt increased by 60% (Dube et al., 2001). In a prospective study of children in South Africa, an accumulation of nine types of ACEs predicted suicide attempts 1 year later, such that the risk for suicide attempts increased as the number of ACEs increased (Cluver, Orkin, Boyes, & Sherr, 2015). Among youth referred for delinquency to the Florida Department of Juvenile Justice, the more ACEs experienced, the greater the likelihood of having made a suicide attempt at some point in their lives (Perez, Jennings, Piquero, & Baglivio, 2016). Data from the National Comorbidity Survey indicated that even after controlling for psychological and demographic variables, the more ACEs experienced, the greater the risk for making a suicide attempt in adulthood (Afifi et al., 2008). Data from the National Epidemiological Survey on Alcohol and Related Conditions also indicated that a higher number of ACEs was associated with an increased risk for lifetime suicide attempts after controlling for demographic variables (Choi et al., 2017).
+THOMPSON et al.
+In sum, several studies have found that an accumulation of ACEs was associated with an increased risk for suicidal ideation and/or suicide attempts. Some of these studies have used national data (Afifi et al., 2008; Choi et al., 2017), many have measured a wide range of ACEs (Choi et al., 2017; Cluver et al., 2015; Dube et al., 2001; Merrick et al., 2017; Perez et al., 2016), some have assessed for suicidality in adulthood (Afifi et al., 2008; Dube et al., 2001) or prospectively (Cluver et al., 2015), and some have controlled for other suicidal behavior risk factors (Afifi et al., 2008; Choi et al., 2017; Perez et al., 2016). However, none has incorporated all of these methodological strengths into the same study. The purpose of the current study was to add to the body of research by testing the unique and cumulative associations of eight different ACEs with suicidal ideation and suicide attempts in adulthood using a nationally representative sample, after controlling for several established risk factors for suicide.
+------------------------------Wiley 1 123
+2.2	| Measures
+2.2.1 | Outcomes—Suicide attempts and ideation
+We created two dichotomous outcome variables that reflected if respondents had engaged in suicide ideation and if they had attempted suicide in adulthood. Because we were interested in ensuring that the ACEs temporally preceded suicidal ideation and suicide attempts, we used outcome measures assessed at Waves 3 and 4. Respondents were asked “During the past 12 months, did you ever seriously think about committing suicide?” and “During the past 12 months, how many times did you actually attempt suicide?” Respondents who reported having considered suicide at least once at either Wave 3 or 4 were classified as having ideated, and respondents who reported having attempted suicide one or more times at Wave 3 or 4 were classified as having attempted suicide.
+2	| METHODS
+2.1	| Sampling procedures and sample
+The sample was derived from the National Longitudinal Study on Adolescent Health (Add Health; Harris, 2009). A multistage-stratified cluster design was used to sample public and private high schools across the United States (Harris, 2013); 79% (n = 132 schools) of the recruited schools participated, and all students attending these schools were invited to complete in-home surveys. The Add Health study was approved by the University of North Carolina School of Public Health Institutional Review Board using guidelines based on the Code of Federal Regulations on the Protection of Human Subjects 45CFR46. Local IRB approval for the secondary analysis of the data also was obtained. Add Health participants provided written informed consent for participation in all aspects of the study. For less sensitive questions, interviewers read the questions and entered respondents' answers. For more sensitive questions, respondents listened to prerecorded questions through earphones and entered the answers directly via audio computer-assisted self-interviewing. In 1995, Wave 1 surveys were completed by 20,745 seventh to 12th graders. Among respondents eligible for follow-up, 14,738 (89%) completed a Wave 2 in-home survey 1 year later, 15,197 (77%) completed a Wave 3 in-home survey approximately 7 years after the Wave 1 survey, and 15,701 (80%) completed a Wave 4 in-home survey approximately 13 years after the Wave 1 survey. We restricted our sample to those who completed surveys at all four waves and had a valid sampling weight. The weight variable we used was recommended for longitudinal data with Wave 1 in-home survey participants who also completed the surveys at Waves 2, 3, and 4 (Chen & Chantala, 2014). This resulted in an analytic sample of 9,421 participants. Participants' mean age was 15.03 years (SE = 0.11) at Wave 1. The sample was evenly divided by gender (50% male and 50% female) and urbanicity (49% lived in rural or partly rural areas, and 51% lived in urban areas). Two thirds (66%) of the sample participants were non-Hispanic white, 12% were Hispanic, 15% were non-Hispanic black, and 7% were of another or mixed race. For more detailed information on survey design, see Harris (2013) and Chen and Chantala (2014).
+2.2.2 | Predictors—ACEs
+We assessed for eight types of ACEs. Although some ACEs were assessed at Wave 3 or 4, the questions used to assess for them were worded to capture respondents' experiences when they were in childhood or adolescence.
+Physical abuse was assessed at Wave 4 and reflected if respondents had been hit with a fist, kicked, or thrown on the floor, into a wall, or downstairs by a parent or adult caregiver more than once before they were 18 years of age (Brumley, Jaffee, & Brumley, 2017; Smith, Smith, Oberleitner, Gerkin, & McKee, 2018). Sexual abuse was assessed at Wave 4 by asking respondents if before the age of 18, a parent or other adult caregiver had touched them in a sexual way, forced them to touch him or her in a sexual way, or forced them to have sexual relations. Emotional abuse was assessed at Wave 4 based on how often a parent or other adult caregiver had said things to the respondents before their 18th birthday that hurt their feelings or made them feel like they were not wanted or loved. We used a cut point of six or more times to increase the measure's specificity in reflecting moderate-to-severe emotional abuse (Scheidell et al., 2017; Smith et al., 2018). Neglect was assessed at Wave 3 by asking respondents if a parent or other adult caregiver had not taken care of their basic needs, such as keeping them clean or providing food or clothing more than once before sixth grade (when a child is typically 11-12 years of age; Brumley et al., 2017). Parental death, assessed at Waves 1 and 2, measured if respondents had experienced the death of either parent in childhood. Parental incarceration was assessed at Wave 4 and measured if a respondent's biological mother or father had spent time in jail or prison before the youth was 18 years of age. Parental alcoholism was measured in the Wave 1 parent interview by asking if the child's biological mother and/or biological father currently had alcoholism. Family history of suicidal behavior assessed if a family member had tried to kill themselves during the past 12 months, assessed at Waves 1 and 2.
+2.2.3 | ACEs score
+A cumulative score was created by summing the eight types of adverse experiences.
+124-1—Wl LEY----------------------------------------------
+2.2.4 I Covariates
+Demographic covariates included gender, age in years, race, and urbanicity assessed at Wave 1. In order to determine the unique role of ACEs in predicting suicidal ideations and attempts, we also controlled for other known risk factors for suicidality that were available in the Add Health dataset. These included depressive symptoms (Brent, Baugher, Bridge, Chen, & Chiapptta, 1999; Goldsmith, Pellmar, Kleinman, & Bunney, 2002), problem alcohol use (Norstrom & Rossow, 2016; Parks, Johnson, McDaniel, & Gladden, 2014), drug use (Parks et al., 2014), delinquent behaviors (Thompson, Ho, & Kingree, 2007), and impulsivity (Liu, Trout, Hernandez, Cheek, & Gerlus, 2017). Depressive symptoms were assessed with a modified version of the Center for Epidemiologic Studies Depression Scale (Radloff, 1977) and were computed as the sum of 20 items answered on a 0-3 scale (a = 0.86; M = 6.90; SD = 0.12). Delinquency was assessed with the mean score of 15 items answered on a 0-3 scale (e.g., took something from a store without paying for it; a = 0.84; M = 0.28; SD = 0.01). Respondents were considered as having an alcohol problem if they reported that they had been drunk at least three to 12 times or had experienced negative consequences of alcohol use at least twice in each of three or more life domains (e.g., family and school), in the past year (15%). The mean of three items assessed on a 1-5 scale was used to measure impulsivity (e.g., “when making decisions, you generally use a systematic method for judging and comparing alternatives;” a = 0.70; M = 2.23; SD = 0.01). Other drug use was a dichotomous variable that reflected if a respondent had used marijuana, cocaine, or another illegal drug in their lifetimes (28%). These covariates have been used in previous research with Add Health data (Swahn & Donovan, 2004; Thompson et al., 2007; Thompson & Light, 2011).
+2.3	I Data analytic strategy
+Variance estimates were adjusted to account for the complex sample design. Analyses were conducted using SPSS 24. We used multivariate logistic regression to test associations between each of the eight ACEs with suicidal ideation and suicide attempts in adulthood, controlling for depression, delinquency, alcohol problems, drug use, impulsivity, gender, age, race, and urbanicity. We next tested the cumulative association of ACEs with risk for suicidal ideation and attempts while controlling for the same covariates. We first examined the percentage of respondents within each ACE category who had engaged in suicide ideation and attempted suicide. We next conducted multivariate logistic regression analyses controlling for the covariates and using an ordinal ACE variable, with no ACE experiences being the reference category. Because only 2.5% of the sample had experienced more than four ACEs, participants with four or more ACEs were combined into a category with those who had three ACEs, such that 0 = no ACEs, 1 = one ACE, 2 = two ACEs, and 3 = three or more ACEs. We then conducted another set of multivariate logistic regression analyses to examine the association of an ACE continuous variable, ranging from 0 to 8, with suicide ideation and attempts.
+THOMPSON et al.
+3	I RESULTS
+3.1	I Descriptive data
+Among the 9,421 survey participants in the analytic sample, the percentage that experienced a specific ACE ranged from a low of 4.7% for parental death to a high of 16.2% for emotional abuse Approximately half the sample (54%) did not report any ACE, 26% had experienced one, 12% had experienced two, and 8% had experienced three or more. In terms of suicidal behaviors, 12.5% reported having seriously considered suicide and 3.3% reported having attempted suicide (see Table 1).
+3.2	I Associations of ACEs with suicidal ideation in adulthood
+The odds of suicidal ideation in adulthood increased twofold to threefold among those who had experienced sexual abuse (p < 0.001), physical abuse (p < 0.001), or emotional abuse (p < 0.001) in childhood. The odds of having suicidal ideation in adulthood increased approximately 1.5 times among those who had a family history of suicidality (p < 0.05) or had a parent incarcerated in childhood (p < 0.001). Neglect, parental death, and parental alcoholism did not increase the odds of suicidal ideation in adulthood (see Table 2).
+In terms of cumulative ACEs, 8.1% of those with no ACEs reported ideation, 14.1% of those with one ACE reported ideation, 18.7% of those with two ACES reported ideation, and 26.2% of those with three or more ACEs reported ideation in adulthood. The odds of suicidal ideation in adulthood increased 1.69 times (95% CI [1.33, 2.15]) among those with one ACE, 2.31 times (95% CI [1.71, 3.13]) among those with two ACEs, and 3.13 times (95% CI [2.34, 4.20]) among those with three or more ACEs when compared with those who had not experienced any ACEs (see Table 3). Further, the continuous variable representing cumulative ACEs predicted suicide ideation is adulthood (AOR = 1.38; 95% CI [1.27, 1.49]).
+3.3	I Associations of ACEs with suicide attempts in adulthood
+The odds of making a suicide attempt in adulthood increased twofold to threefold among those who had experienced sexual abuse (p < 0.001), physical abuse (p < 0.001), or emotional abuse
+TABLE 1 Descriptive data for adverse childhood experiences and suicidal ideation and attempts (n = 9,421)
+Note. Both models controlled for depression, delinquency, alcohol problems, drug use, impulsivity, gender, age, race, and urbanicity. ACE: adverse childhood experience; AOR: adjusted odds ratio; CI: confidence interval.
+*95% CI does not include 1; p < 0.05.
+(p < 0.001) in childhood or had a family member attempt or complete suicide during their childhood (p < 0.01). The odds of a suicide attempt in adulthood increased 1.5 times among those who had a parent incarcerated in childhood (p < 0.05). As with suicidal ideation, neglect, parental death, and parental alcoholism did not increase the odds of suicide attempts in adulthood (see Table 2).
+In terms of cumulative ACEs, 2.0% of those with no ACEs attempted suicide in adulthood, 3.5% of those with one ACE attempted suicide, 4.2% of those with two ACES attempted suicide, and 7.9% of those with three or more ACEs reported a suicide attempt in adulthood. The odds of attempting suicide in adulthood increased 1.57 times (95% CI [1.01, 2.44]) among those with one ACE, 1.99 times (95% CI [1.22, 3.23]) among those with two ACEs, and 3.53 times (95% CI [2.20, 5.66]) among those with three or more ACEs when compared with those who had not experienced any ACEs (see Table 3). Further, the continuous variable representing cumulative ACEs predicted suicide ideation is adulthood (AOR = 1.38; 95% CI [1.22,1.55]).
+4	| DISCUSSION
+We found that for suicidal ideation, the most significant ACEs were sexual and emotional abuse, followed by physical abuse, parental incarceration, and family history of suicidality. For suicide attempts, the most significant ACE was sexual abuse, followed by emotional abuse, family history of suicidality, physical abuse, and parental
+incarceration. Neglect, parental death, and parental alcoholism did not increase the risk for suicidal ideations or attempts in adulthood. Our study also documented the increased suicidality risk conferred by an accumulation of ACEs. Taken as a whole, our results were consistent with previous research demonstrating the ACEs-suicidality link (Afifi et al., 2008; Choi et al., 2017; Cluver et al., 2015; Dube et al., 2001; Merrick et al., 2017; Perez et al., 2016).
+Despite this study's strengths, there were some limitations that should be noted. First, the reporting period for the outcome measures was limited to 1 year and thus likely did not capture all incidences of suicide ideation and attempts the respondents may have experienced in adulthood. Second, suicide attempts were only assessed if respondents acknowledged engaging in suicidal ideation. Thus, a respondent who attempted suicide without first considering suicide would be misclassified as a false negative. Third, we focused on suicide attempts rather than deaths. Nonetheless, focusing on suicide attempts is important because a nonfatal suicide attempt has been found to be the strongest predictor of death by suicide (Bostwick, Parbati, Geske, Alastair, & McKean, 2016), and risk factors for attempts and deaths are similar (Gould & Kramer, 2001). Fourth, although the reporting timeframe for the ACE variables was prior to adulthood, ACEs were assessed retrospectively. This may have resulted in an underestimate of the ACEs, especially if the respondent was very young when experiencing the ACE. Unfortunately, most ACE studies also have relied on retrospective accounts, with the one exception being the study with South African youth by Cluver et al. (2015) discussed earlier. Future research on ACEs should strive to collect data on ACEs prospectively throughout childhood. Fifth, data were based primarily on self-report only, which raises the concern of social desirability bias. However, relying on parent reports also would be problematic, as parents may underreport events in which they played a role or were implicated. It is possible that this was the case for the parent-reported parental alcoholism measure; if so, then this could explain why we did not find a significant association between parental alcoholism and suicidality in our study whereas another study that relied on selfreport of both suicidality and parental alcoholism did (Thompson et al., 2017). Future research should ideally use multiple informants when measuring ACEs.
+12^Wl LEY------------------------------------------
+Our findings highlight the importance of adopting strategies to prevent exposure to ACEs as part of a comprehensive approach to suicide prevention (Ports et al., 2017). Promising research indicates that building community capacity and enhancing social networks can reduce community-wide ACE prevalence (Hall, Porter, Longhi, Becker-Green, & Dreyfus, 2012). Another type of ACE prevention approach is parent education programs. In a meta-analysis of randomized controlled trials of parenting education programs, several programs showed significant reductions in child maltreatment. Further, these positive reductions were found across low-, middle-, and high-income countries and for primary (i.e., families randomly selected from community), secondary (i.e., families at risk for child maltreatment), and tertiary prevention programs (i.e., families in which child maltreatment had already happened; Chen & Chan, 2016). Home visitation programs and prenatal programs also have been found to reduce the likelihood of child maltreatment (Eckenrode et al., 2000). Further, suicide prevention programs that promote connectedness and healthy relationships may be expanded to address childhood adversities (Ports et al., 2017).
+Future research also should focus on identifying mediating mechanisms for the ACEs-suicidality association. It is important to note that even though a higher number of ACEs was associated with an increased risk for suicide ideation and attempts, most who experienced three or more ACEs did not have suicide ideation or make a suicide attempt in adulthood. Among those who experienced three or more ACEs, approximately 74% did not report suicide ideation, and 92% had not made a suicide attempt. This suggests the need for research to identify factors that promote resiliency among youth who experience many ACEs and protect them from an increased risk for suicidality. Although our study showed that ACEs were associated with an increased risk for suicide ideation and attempts, it did not test for variables that might explain this link. Fortunately, some research has already begun to examine mediating mechanisms. For example, in a cross-sectional, population-based survey of Canadian adults, depression, anxiety, substance use, and chronic pain helped to account for a portion of the association between different ACEs and lifetime suicide attempts (Fuller-Thomson, Baird, Dhrodia, & Brennenstuhl, 2016). Another study with youth referred to the Florida Department of Juvenile Justice found that aggression and impulsivity helped explain why an accumulation of ACEs was associated with an increased risk for suicide attempts (Perez et al., 2016). In the only prospective study on the link between ACEs and subsequent suicide ideation and attempts, poor mental health symptoms, but not drug and alcohol misuse, were found to mediate this association (Cluver et al., 2015). This finding suggests that among youth exposed to ACEs, effective mental health services could potentially reduce the likelihood of suicidality. Using data from the National Comorbidity Study, researchers found that the number of psychiatric disorders accounted for the association between verbal child abuse and suicide attempts but not for violent child abuse and suicide attempts (Sachs-Ericsson et al., 2017). These findings led the authors to speculate that different mediating mechanisms may account for the associations of different types of ACEs and suicide attempts. Future research should test potential mediating pathways that account for the association of an accumulation of ACEs and subsequent suicidality and also determine
+if these mediating pathways account for associations between individual ACEs and suicide risk. Studies that use prospective designs with more than two timepoints provide the strongest methodology for testing mediation, as mediation based on cross-sectional designs can produce biased estimates and hence faulty conclusions regarding mediating mechanisms (Cole & Maxwell, 2003). Longitudinal studies with two timepoints are better than cross-sectional data, but these “half-longitudinal designs” with mediators assessed at the same time as the predictor or the outcome can still result in biased estimates. Thus, research to investigate mediators of the association between ACEs and suicidal behaviors should rely on data collected from at least three timepoints and ensure constructs occur and are measured in the correct temporal order. That is, mediators should be assessed at a subsequent timepoint than ACEs, suicidal behaviors assessed at a subsequent timepoint than mediators, and prior levels of mediators and outcomes controlled. With this type of design, research can shed light on mechanisms by which ACEs lead to increased risk for suicidality.
+Moreover, future research should determine what ACEs are most important to include in a cumulative ACE measure. Consistent with some other studies, our research indicates that sexual, physical, and emotional abuse, having a family member who attempted suicide, and having an incarcerated parent are important ACEs to include in a cumulative ACE measure. Although some studies have included mental illness in household member as an ACE variable (Dube et al., 2001), no study, to our knowledge, has included family history for suicide as an ACE variable. It is possible that a family history of suicide may best be conceptualized as an inherited genetic risk rather than an ACE.
+In sum, our study replicated research showing that certain individual ACEs, as well as an accumulation of ACEs, significantly increased the risk for suicide ideation and attempts in adulthood. It added to the literature by testing the unique and cumulative associations of a wide range of adverse childhood events with suicide ideation and attempts occurring in adulthood using a nationally representative sample, after controlling for several established risk factors for suicide. These findings underscore the importance of preventing ACEs and implementing suicide prevention strategies among those with a high number of ACEs.

Child Psychology Psychiatry - 2019 - King - Predicting 3‐month risk for adolescent suicide attempts among pediatric.txt

@@ -0,0 +1,63 @@
+n> Check for updates
+Introduction
+Suicide rates among adolescents in the United States continue to rise (Centers for Disease Control and Prevention, 2019), despite a downturn in the incidence worldwide (World Health Organization, 2017). Moreover, 5.1% of male and 9.3% of female high school students in the United States report a suicide attempt (SA) in the past year (Kann et al., 2018).
+Risk factors for adolescent SAs span demographic, clinical, and social domains, meaning that the risk profiles for suicidal adolescents are multidimensional and heterogeneous. Female adolescents and
+adolescents who self-identify as LGBTQ are at increased risk (Kann et al., 2018; O’Brien, Putney, Hebert, Falk, & Aguinaldo, 2016). Previous history of SA and suicidal ideation (SI) (Nock et al., 2013), presence, persistence, and severity of SI (Czyz & King, 2015), and nonsuicidal self-injury (NSSI) (e.g. Asar-now et al., 2011) have all been reported to be predictors of suicide attempts. Similarly, psychiatric symptoms, such as depression and hopelessness, are consistent correlates and predictors of SA (King, Ewell Foster, & Rogalski, 2013), and symptoms of distress (e.g. anxiety and agitation) and impulse control (e.g. aggression, substance abuse) have emerged as the strongest predictors of attempts among adolescents who report ideation (Nock et al.,
+1056 Cheryl A. King et al.
+2013). Sleep disturbance has been reported as an imminent risk factor for SA and death by suicide (e.g. Koyawala, Stevens, McBee-Strayer, Cannon, & Bridge, 2015).
+Interpersonal factors such as low social connectedness also have been related to the likelihood of suicidal ideation and behavior (Czyz, Liu, & King, 2012; Gunn, Goldstein, & Gager, 2018). Bully victims and perpetrators have reported an increased incidence of SAs (Borowsky, Taliaferro, & McMorris, 2013), and physical and sexual abuse have been prospectively associated with SAs (Castellvi et al., 2017). Interpersonal conflicts and losses, and legal/ disciplinary problems are acute stressors associated with SAs and suicide (e.g. Gould, Fisher, Parides, Flory, & Shaffer, 1996).
+Given this heterogeneity of suicide risk factors, it is challenging for healthcare providers to assess level of risk and for intervention and prevention specialists to identify potent and potentially modifiable targets for risk reduction. Moreover, extant research has focused on single risk factors (Franklin et al., 2017), despite the growing recognition of the multidimensional nature of suicidal risk and current clinical practice, which attempts to integrate available information about multiple risk factors. Consequently, further research that takes into account multiple risk factors is sorely needed.
+The challenge of suicide risk assessment and identification of potent prevention targets is exacerbated for males and for adolescents who conceal or deny their suicidal thoughts. Adolescent females are more likely than males to report SI and behavior (Kann et al., 2018) and to obtain mental health services (Rhodes et al., 2012), yet the rate of suicide is much higher among adolescent males than females (Centers for Disease Control and Prevention, 2019). An improved understanding of the short-term risk factors for SAs among males may enable us to improve risk recognition and prevention. Similarly, although many of the most commonly used screening tools assess SI (e.g. Horowitz et al., 2012), recent SI is not a significant predictor of SAs for all subgroups of adolescents (e.g. King, Jiang, Czyz, & Kerr, 2014).
+Our objective was to examine predictors of SAs during the 3-months following adolescents’ ED visits in the Study One dataset of the Emergency Department Screening for Teens at Risk for Suicide (ED-STARS) Study. This large-scale study was implemented in collaboration with the Pediatric Emergency Care Applied Research Network (PECARN). Its primary aim was to develop the Computerized Adaptive Screen for Suicidal Youth (CASSY), a relatively brief suicide risk screen with the potential for widespread implementation in emergency departments (King et al., under review). Because our baseline assessment included a broad array of previously identified risk factors for SAs, this study also enabled us to examine predictors of SAs following ED visits using multivariable models.
+J Child Psychol Psychiatr 2019; 60(10): 1055-64
+We examined predictors in the total follow-up sample and in subsamples defined by sex and the presence of recent SI. We hypothesized that predictors of SAs would include indicators of SI and behavior (e.g. past week suicidal ideation, lifetime history of suicidal behavior) and, reflecting a different domain, one or more interpersonal risk factors (e.g. peer victimization, low social or school connectedness). We expect interpersonal factors to be important in light of longitudinal studies (e.g. Gunn et al., 2018) and theoretical formulations about the salience of interpersonal processes to suicidal risk (e.g. Durkheim, 1897; Joiner, 2005).
+Methods
+Participants
+Adolescents (ages 12-17) were recruited from 13 EDs in PECARN (June 2015-July 2016) and the Whiteriver Indian Health Service (IHS) Hospital, which serves the White Mountain Apache Tribe (November 2015-April 2017). Among 10,664 approached adolescents, 6,448 (60.5%) completed a suicide risk survey. A subset of patients (n = 2,897 (43.6%) enriched for suicide risk (Figure 1 and Appendix S1) was randomly assigned to a 3-month telephone follow-up; 2,104 participants completed this follow-up (72% retention). The sample included 1,327 females (63.1%) and 777 males (36.9%) with a mean age of 15.1 years (SD = 1.6). Additional demographic information is in Appendix S2.
+Procedure
+At PECARN sites, adolescents were recruited during screening shifts that were randomly selected for each site from time periods when research coordinators were on site (primarily afternoons and evenings due to higher volume of adolescent patients). At the IHS Hospital, recruitment was ED-linked with a daily admission review and IRB permission to contact at home for recruitment. Exclusion criteria were as follows: previous study enrollment, ward of State, non-English speaking adolescents (non-English speaking parents enrolled), medically unstable, and severe cognitive impairment.
+Adolescents completed a self-report survey assessing demographics and suicide risk factors in the ED (except for IHS site). Participants were included if adolescent and parent (n = 1,799, 85.5%),adolescentonly (n = 183,8.7%),orparentonly(n = 122, 5.8%) follow-upinterviewswereconducted. Follow-upinformant (parent oryouth vs. both) was unrelated to participants’ lifetime histories of suicidal ideation and behavior, and to the suicide attempt outcome. Participants with only youth or only parent follow-up interviews were, however, older than those with both interviews. (p < .001, Kruskal-Wallis test.). Written-informed parent/guardian consent and adolescent assent were obtained, in addition to IRB approval from all sites. Adolescentswho turned 18 prior to follow-up were reconsented.
+Measures
+This study incorporated adolescent data from the baseline selfreport survey (92 primary, 27 follow-up questions; details in Appendix S3). Due to ED space and time constraints, a concern for respondent burden, and a need to assess a wide range of risk factors to develop CASSY algorithms, brief, adapted versions of standardized scales were used for many risk factors, all of which had been previously associated with adolescent SAs.
+© 2019 Association for Child and Adolescent Mental Health
+An adapted Columbia-Suicide Severity Rating Scale (C-SSRS; Posner et al., 2008) was used to assess history of SAs at baseline and SAs between baseline and 3-month follow-up. SA was defined as a positive response to either of two questions: “In the past 3 months, have you made a suicide attempt?’ “In the past 3 months, have you tried to harm yourself because you were at least partly trying to end your life?’ Past week SI was assessed with question #3 from the Ask Suicide-Screening Questions (ASQ; Horowitz et al., 2012): “In the past week, have you been having thoughts about killing yourself?’ In defining subgroups of adolescents who did and did not report recent SI, we removed participants who selected “unknown’ or did not respond to the question.
+Additional suicide risk factors assessed at baseline included lifetime severity of SI and suicidal behavior, suicidal rumination, NSSI, depression, hopelessness, homicidal ideation, anxiety, agitation, sleep disturbance, adaptive functioning, alcohol and drug use, impulsivity, aggression, connectedness (family, school, social), peer victimization, physical and sexual abuse, negative life events, and identification as a sexual or gender minority.
+demographics and variables pertaining to suicidal thoughts, suicidal behaviors, and NSSI were added to the model in a stepwise fashion; the model with the lowest Akaike Information Criterion (AIC) was carried forward. Remaining candidates, including all other clinical risk factors examined (see Table 1), were considered using forward stepwise selection. In the final stage, variables were dropped using backward selection (p > .05), such that all variables were statistically significant in the final model.
+To account for the oversampling of higher risk groups for follow-up, a weight equal to the inverse of the sampling probability of each of the three risk groups was applied in analyses. For categorical variables, the reference level was “No’, “None’, or equivalent, when possible. White and non-Hispanic were used as reference populations. When model separation became an issue due to low counts, categories of predictor variables were combined. For each final model, we calculated the predictive performance of the model as the area under the curve (AUC), with a 95% confidence interval (CI). As a sensitivity analysis, we conducted a 10-fold cross-validation of the final model for the full sample. Statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc, 2013).
+Statistical analysis
+Univariable associations between baseline demographic and clinical risk factors and SAs at 3-months were determined, and predictors with significant associations (p < .1) were candidates for inclusion in multivariable logistic regression models (Hosmer, Lemeshow, & Sturdivant, 2013). In stage one,
+© 2019 Association for Child and Adolescent Mental Health
+Results
+Retention
+Retention was greater for males than females (76.0% vs. 70.8%; p = .003) and varied by race (p < .001)
+and ethnicity (p < .001), with higher retention rates for Whites (75.1%) and multiracial youth (79.5%) than other races (range from 61.3-72.6%), and for non-Latinx than Latinx ethnicity (75.6% vs. 68.5%). Higher parental education was also associated with greater retention (p’s < .001). The retention rates for mothers and fathers, respectively, were as follows: high school or less (67.9%, 68.1%), some college/ technical (73.8%, 73.2%), college graduate (77.6%, 81.7%), unknown/not applicable (65.0%, 70.0%).
+Descriptive statistics: suicidal thoughts, suicide attempts, and NSSI
+At baseline, 1,090 adolescents (51.9%) reported a lifetime history of SI and 815 adolescents (39.4%) reported a lifetime history of suicidal behavior, including actual, aborted, and interrupted attempts. The mean number of lifetime SAs reported was 1.67 (SD = 6.91; Median = 0). Regarding number of pastyear NSSI incidents, 1378 adolescents (65.7%) reported none, 339 adolescents (16.2%) reported 12, 121 (5.8%) reported 3-4, and 261 (12.4%)
+© 2019 Association for Child and Adolescent Mental Health
+reported 5 or more (data missing, n = 5). A SA between ED visit and 3-month follow-up was reported for 104 adolescents (4.9%; 84 females, 6.3%; 20 males, 2.6%). There was one suicide death, which was included as a SA in analyses.
+Spearman’s correlations among risk factors are reported in Tables S1-S4. As examples of the strength of correlations, lifetime severity of SI was highly positively correlated with lifetime history of suicidal behavior (.70, p < .001) and moderately positively correlated with number of NSSI incidents during the past 12 months (.53, p < .001). Social and school connectedness were moderately positively correlated (.47, p < .001).
+Site differences were identified in suicide risk predictors and outcomes. This information is provided in Tables S5-S9.
+Predictors of suicide attempt during 3 months following ED visit
+Univariable associations with suicide attempts. -Sex, sexual, and gender minority status, and all of
+the examined psychosocial and clinical characteristics predicted SAs at 3-month follow-up (see Table 1).
+Multivariable regression models. The final multivariable model for the total sample included past week SI (yes/no), lifetime severity of SI, history of suicidal behavior, and school connectedness (AUC = 0.86, 95% CI: 0.82-0.89; Table 2). In the sensitivity analysis, the ORs, (CIs), and AUCs fitted from each of the 10 subsamples (each approximately 90% of full cohort) were similar, with a median AUC of 0.87 and IQR 0.84-0.90.
+To examine replicability of this model across sites, we examined a model including site and the interaction between site and the final model risk score (fitted logit values for each patient). The interaction was nonsignificant (p = .55), suggesting that the relationship between the predicted risk and SA outcome does not differ by site. Site was also unrelated to SA risk (p = .70) after taking into account risk factors.
+For adolescents without past week SI at baseline, the final model included lifetime SI severity and social connectedness (AUC = 0.84, 95% CI: 0.780.90; Table 3). For adolescents with recent SI at baseline, the final model included family public assistance, suicidal rumination (repetitive thoughts), and social connectedness (AUC = 0.69, 95% CI: 0.62-0.76; Table 3).
+For male adolescents, the final model included past week SI and lifetime SI severity (AUC = 0.89, 95% CI: 0.85-0.94; Table 4). For female adolescents, the model included past week SI, number of NSSI incidents during the past 12 months, and social connectedness (AUC = 0.84, 95% CI: 0.81-0.87).
+Discussion
+In this prospective study of adolescent ED patients, we identified baseline predictors of SAs across a 3month period of follow-up using multivariable models for the entire sample, and for subsamples defined
+by sex and the presence or absence of recent suicidal thoughts. These subgroups included two particularly vulnerable groups: adolescent males who receive fewer mental health services (Rhodes et al., 2012) and have a much higher rate of suicide than adolescent females (Centers for Disease Control and Prevention, 2019), and adolescents who do not report recent suicidal thoughts, which challenges risk recognition.
+Study results replicate the importance of previously identified suicide risk factors. Every clinical risk factor included in our baseline suicide risk survey was associated significantly with the likelihood of a SA between the baseline ED visit and 3month follow-up. Concordant with hypotheses, past week SI, lifetime severity of SI, lifetime history of suicidal behavior, and an interpersonal factor, school connectedness, emerged as the key predictors of attempts for the total sample. Moreover, emphasizing the importance of connectedness to our understanding of risk, either school or social connectedness emerged as a key predictor for three of the four subgroups of adolescents studied. Contrary to hypotheses, however, the model for males included only two factors: recent SI and lifetime severity of SI.
+Lifetime severity of SI was found to be a key predictor for the overall sample, and three of the four subgroups of adolescents examined. This finding is consistent with previous studies indicating that adolescents who develop a suicide plan are more likely to make an attempt than ideators without a plan (Nock et al., 2013), that intensity of SI predicts SAs (Peters, Mereish, Solomon, Spirito, & Yen, 2018), and that “worst ever’ SI is as strong a predictor of suicide risk as current ideation (Beck, Brown, Steer, Dahlsgaard, & Grisham, 1999). Similarly, the importance of lifetime history of suicidal behavior is consistent with studies showing that increased risk for subsequent self-harm and death by suicide persists for years after initially seeking health care for self-harm (Finkelstein et al., 2015).
+© 2019 Association for Child and Adolescent Mental Health
+School or social connectedness emerged as a key predictor for several subgroups of adolescents, which is consistent with a growing body of research (Gunn et al., 2018) indicating that higher levels of school connectedness were associated with less suicidal behavior in general school samples, high-risk adolescents, and sexual minority adolescents (Marraccini & Brier, 2017). Social connections may have long-term consequences for mortality as well as morbidity. A 14-year follow-up of adolescent hospitalized for SI and behavior found that those assigned to an intervention to mobilize social support from adults had reduced self-injury mortality (King et al., 2019). Therefore, social and school connectedness are likely to be an important target for risk assessment and preventive intervention.
+Adolescents who do not report recent SI, who comprised nearly one-third of the youth who made SAs in this study, can be challenging to identify in EDs and other settings where the focus is on current risk. In this subgroup, lifetime severity of SI and
+© 2019 Association for Child and Adolescent Mental Health
+social connectedness were the primary risk indicators. The accuracy of prediction in this ‘hidden’ subgroup provides particularly strong support for the need for suicide risk screening in the pediatric ED. Surprisingly, the accuracy of prediction for this subgroup (AUC = 0.84) was higher than the accuracy of prediction for the subgroup of adolescents who reported recent suicidal ideation (AUC = 0.69). This may be due to the inconsistency of adolescents’ reports of SI across study measures, which will be the focus of a future study.
+NSSI only emerged as a primary risk factor for females. It is unknown whether or not this relates to the different types of NSSI reported by females (Sornberger, Heath, Toste, & McLouth, 2012), social influences, and interpersonal challenges associated with engagement in NSSI (Victor & Klonsky, 2018), or females’ higher likelihood of experiencing suicidal thoughts and engaging in suicidal behavior (Kann et al., 2018). The more limited statistical power for adolescent males, due to fewer SA outcomes, may
+1062 Cheryl A. King et al.
+also be important as NSSI was a predictor of SAs among males in univariable analyses.
+The prediction model AUCs for the full sample, the sample of adolescents who did not report recent SI at baseline, and the subsamples of males and females each ranged between 0.84 and 0.89, which can be considered excellent classification accuracy (Hosmer et al., 2013), and contrasts with the disappointing performance of previous single risk factor approaches to suicide risk prediction (Franklin et al., 2017). Although the heterogeneity of suicide risk factors and the low base rates of SAs and suicide are challenges to risk stratification (Belsher et al., 2019), findings suggest that a multivariable prediction model can be useful for the short-term prediction of adolescent SAs. However, of equal or greater importance, these models identify potentially important targets for clinical risk evaluation and prevention. Screening tools for risk recognition can be developed using prediction algorithms developed from large data sources (Belsher et al., 2019). We used this strategy in developing the CASSY, which is being validated in a new sample.
+Results should be considered within the context of study limitations. We used brief and adapted scales to assess most suicide risk factors to reduce respondent burden and facilitate patient flow in EDs. Although each of the baseline clinical risk factors we assessed was found to be a significant univariable predictor of SAs, the use of brief scales may have reduced the reliability of measurement and our ability to fully capture each construct. Furthermore, this study was conducted primarily in pediatric EDs of academic health systems, which are not representative of the range of EDs in the United States. In addition, we had lower levels of retention for adolescents from racial and ethnic minority groups, females, and adolescents whose parents had less education. Although we considered weighting the sample for nonresponse, we chose to prioritize adjusting for the oversampling of higher risk groups because we had specific information pertinent to the oversampling and did not want to apply multiple weights to relatively small subgroups. Moreover, for the most part, these variables were not predictive of SA, and therefore our predictive models are most likely not biased due to nonresponse. Finally, despite the relatively large size of this study, the relatively low number of youth with SAs limited our statistical power for identifying multiple predictors, especially within critical subgroups such as males, for whom the number of attempts was smaller than for females. While in this study, our focus was on identifying key risk factors, in future reports we will describe how we also used study data to develop and validate an adaptive screening tool.
+In summary, in this short-term prospective study of predictors of SAs in a large and diverse sample of adolescents recruited from pediatrics EDs, we found that past week SI, lifetime severity of SI, lifetime
+history of suicidal behavior, and connectedness were critical risk and protective factors. We also documented variation in key risk factors across important subgroups, including adolescent males and adolescents who did not report recent SI. The risk and protective factors identified may be important to assess clinical risk evaluations and can serve as important targets for intervention and prevention strategies.

Collaboration-with-People-with-Lived-Experience-of-Mental-Illness-to-Reduce-Stigma-and-Improve-Primary-Care-Services-A-Pilot-Cluster-Randomized-Clinical-TrialJAMA-Network-Open.txt

@@ -0,0 +1,55 @@
+Introduction
+Collaboration with people with lived experience of mental illness (PWLE), also referred to as service users, is increasingly recognized as an integral strategy to improve mental health care.1-3 The commitment to collaboration has been endorsed by the World Health Organization (WHO),4 national governments in their mental health policies,5-8 and mental health professional organizations.9 However, there is a scarcity of evidence-based approaches with demonstrated safety for PWLE and effectiveness in improving care.10 This gap in evidence for collaboration with PWLE is especially pronounced in low- and middle-income countries (LMIC) where efforts are under way to rapidly expand access to mental health services in primary care.11
+Collaboration with PWLE can be particularly important to reduce stigma among primary care practitioners (PCPs), which is a barrier to effective integration of mental health care.12-14 Stigma among PCPs is one contributor to low rates of detection of mental illness,15-20 which is a common shortcoming in primary care mental health programs in LMICs.15,21,22 One avenue to reduce stigma is through social contact interventions between PWLE and stigmatizing groups, such as PCPs. In social contact interventions, stigmatized and stigmatizing groups interact through sharing personal stories, engaging in collaborative activities, and having structured and unstructured social interactions.23-25 Unfortunately, most research on social contact and mental illness stigma is limited to high-income countries, and few studies in any setting have demonstrated long-term attitudinal change (eg, only 1 study in LMICs included a 12-month follow-up12); nor have they routinely evaluated behavioral changes, such as the association between stigma reduction and improved clinical care.12,25-28 There has recently been a call for more methodological rigor in social contact intervention trails.25 Moreover, despite expanding research29 on the WHO Mental Health Gap Action ProgrammeIntervention Guide (mhGAP-IG)30—the global training initiative for primary care-based mental health services in low-resource settings—there is a lack of studies on structured involvement of PWLE and potential benefits of social contact during these trainings.
+Therefore, we conducted a pilot cluster randomized clinical trial (cRCT) of Reducing Stigma Among Healthcare Providers (RESHAPE) in Nepal. RESHAPE is a stigma-reduction intervention conducted in collaboration with PWLE to change attitudes of PCPs participating in mhGAP-IG training.31 This pilot cRCT was deemed necessary before proceedingto a full-scale trial to address the specific objectives of assuring that PWLE could safely participate, that PCPs would be willing to attend trainings with PWLE, and to establish parameters for recruitment, randomization, and retention. We also sought preliminary estimates of benefit among PCPs measured as reduction in stigma and improvements in clinical competency, operationalized as accuracy of mental illness diagnoses. A cluster design, with primary care facilities being the unit ofclustering, was used because of the shared mental health care responsibilities among PCPs working at the same facility.
+Methods
+Design
+The study protocol is available in Supplement 1 and has been published.32 This report follows the Consolidated Standards of Reporting Trials Extension (CONSORT Extension) reporting guideline for randomized studies,33 including extensions for pilot and feasibility trials34 and for cluster trials.35 This pilot cRCT was conducted in Chitwan, Nepal, using a 1:1 allocation ratioof primary care facilities (the unit of clustering). Nepal was selected because it exemplifies low-resource conditions in LMICs, and there was an existing research infrastructure evaluating mental health care integration into primary care through the Programme for Improving Mental Health Care (PRIME).36-38 No methodological changes were made after trial commencement.
+Ethical Review of the Study
+The study was granted ethical approval by the Nepal Health Research Council, Duke University institutional review board, and George Washington University institutional review board. All participants completed a signed consent form in Nepali. Before the start of PhotoVoice training, PWLE were evaluated by psychiatrists to appraise ability to safely participate in the program. The psychiatrist was available if PWLE had symptom relapse duringthe weeks of the PhotoVoice trainings. A psychosocial counselor was present to support PWLE for all PhotoVoice sessions and the PCP trainings. For the diagnostic accuracy component of the study, any patients found to have an incorrect diagnosis had their medical records corrected by the study psychiatrist, and they were started on the appropriate treatment for the corrected diagnosis.
+Participants and Setting
+All primary care facilities in Chitwan district in which mental health services had not yet been integrated were eligible for inclusion as clusters. All PCPs who had prescribing privileges at the primary care facilities were eligible. Primary care facilities typically have only 1 or 2 PCPs. Therefore, although each cluster included all eligible staff, there were few PCPs per facility. Approximately 1 year after PCPs were trained and supervised, patients whom they newly diagnosed with depression, psychosis, or alcohol use disorder were evaluated by a psychiatrist for accuracy of the PCP diagnosis. Caste and ethnicity data were recorded for all PCPs and patients participating in the study. Caste and ethnicity data were documented by research assistants based on participants’ last names, which indicated caste and ethnic background. In cases of last names that could be categorized into multiple groups, research assistants asked the participant to clarify the caste and ethnic identification. Caste and ethnicity were recorded for this study because the social categorizations have been associated with stigmatization and discrimination, mental illness risk factors, and differential treatment within the health system.39-45
+Interventions
+The RESHAPE intervention is designed based on social neuroscience, social psychology, and medical anthropology theories to create a what-matters-most approach to stigma reduction.46,47 Figure 1 depicts the components of RESHAPE. Full details on the content development and proof of concept testing have been published.31 The design of the RESHAPE intervention and implementation of this trial were conducted in collaboration with PWLE.
+The RESHAPE intervention engages PWLE to participate as cofacilitators in a 40-hour mhGAP-IG training30 adapted for Nepal.37 Within the mhGAP-IG training, PWLE present recovery testimonials through photographic narratives, using a technique known as PhotoVoice.48 PhotoVoice is a commonly used participatory methodology in global health.49 PWLE are eligible for the PhotoVoice skill-building if they have been treated at primary care settings for 1 of 4 priority disorders (depression, psychosis, alcohol use disorder, and epilepsy) and are now in a state of recovery, based on evaluations conducted by a Nepali psychiatrist. Selected PWLE were trained
+through 12 PhotoVoice sessions over 3 months in which they were taught to develop a photographic recovery narrative. PWLE constructed narratives that were approximately 7 minutes in duration and included 3 components: life before treatment, the experience of treatment, and life after treatment. On days 2 through 8 of the 9-day mhGAP-IG training, PLWE participated by providing their narratives followed by question-and-answer sessions, totaling approximately 45 minutes of direct facilitation per day. PWLE also participated in structured and unstructured social activities with PCPs (eg, ice-breakers, energizers, and meals).
+In addition, the RESHAPE model included presentations from aspirational figures. Aspirational figures were PCPs who had previously been trained to provide mental health services and who were recognized by supervisors as enthusiastic about treating patients with mental illness. These PCPs were referred to as aspirational figures because of the hope that PCP trainees will aspire to similar commitment to caring for patients with mental health concerns. Aspirational figures presented a myth-busting session and a recovery story from the perspective of a health care clinician. PWLE and aspirational figures participated over approximately 3 months of PCP trainings.
+The key themes addressed in RESHAPE were identified through the what-matters-most framework for understanding origins of stigma.31 The 3 stigma domains were survival threats, social threats, and professional threats. Survival threats included beliefs that people with mental illness are violent, including lay ethnopsychology understandings that the brain-mind (Nepali: dimaag) controls social behavior and inhibits violence, but a damaged brain-mind in mental illness leads to loss of behavioral control.20 Social threats referred to beliefs that interacting with people with mental illness could cause mental illness in health care clinicians and result in loss of social status,50 as captured in the saying “the doctor of mad patients is also mad” (Nepali: “paagal ko daktar pani paagal ho'').20 Professional threats included beliefs that providing health care for people with mental illness is
+ineffective, burdensome, and ultimately would jeopardize other patient care responsibilities.20,50,51 In addition, there was intersectional stigma resulting from the dual-burden of discrimination experienced by low-caste and ethnic minorities and women, who are disenfranchised in society and also considered more likely to experience mental illnesses and alcohol use disorders.39-42 The PhotoVoice narratives of PWLE and the discussion led by aspirational figures were structured to address these 3 stigma domains of survival, social, and professional threats.31
+The control condition-training as usual (TAU)—was the Nepali adaptation of mhGAP-IG without the structured participation of PWLE. mhGAP-IG included flow-charts on clinical decision-making with basic information on diagnosis and treatment.30 The mhGAP-IG was adapted for Nepal as part of PRIME.36 Through PRIME in Nepal, a 9-day PCP training was developed including approximately 40 hours of learning with 24 hours dedicated to mhGAP modules, 12 hours on psychosocial basics, and 4 hours on logistical implementation processes.37
+The RESHAPE version of the PCP training and the TAU control were time matched at 9 days of training such that some sections in the TAU group that would be covered by a psychiatrist were covered by PWLE in the RESHAPE approach. Group supervision was held separately by groups for approximately 4 to 6 hours in a 1-day session conducted once every 3 months following the training. The training and supervision were conducted by Nepali psychiatristsand psychosocial counselors.
+Outcomes
+The prespecified outcomes for determiningfeasibility and acceptability and progression to a full trial were identification of qualitative themes related to recovery; 75% fidelity of PWLE and aspirational figures to the items on the RESHAPE fidelity checklist; comparable PCP baseline characteristics for the groups; retention of 50% of service users trained in PhotoVoice; retention of 66% of PCPs at the end point; fewer than 15% missing items on outcome measures; and fewer than 10% adverse events. The current analysis focused on the quantitative outcomes. Qualitative outcomes have been previously presented.31,52,53 Fidelity on the RESHAPE fidelity checklist was recorded by a research assistant who observed all of the trainings and noted what activities were done for each section of the training related to RESHAPE components. For example, (1) did PhotoVoice narratives include the 3 components of pretreatment experience of mental illness, experience of treatment, and life after starting treatment?; (2) was there a question-and-answer session after the PhotoVoice presentation in which PWLE responded to PCP trainees?; and (3) did the myth-busting section by aspirational figures include all 8 myths? For PWLE, adverse events were measured at each PhotoVoice training session and after each PCP training by a psychosocial counselor asking about adverse experiences, including both specific concerns (eg, suicidality, symptom relapse) as well as giving PWLE an opportunity to raise any other concerns.52 Family members of PLWE were also given an opportunity to discuss any adverse events.53 Adverse events among PCPs were recorded at supervision sessions, and as well as ad hoc documentation of events raised by PCPs contacting the research team or clinical supervisors.
+Quantitative outcomes at the level of individual PCPs were included to evaluate within-group trends over time. The main assessment periods for PCPs in both groups were (1) baseline, which was thefirst day of the training; (2) midline, which was 4 months after training; and (3) end line, which was the primary end point occurring 16 months after training.
+The primary quantitative outcome measure was PCPs’ level of stigma as measured with the Social Distance Scale (SDS).54,55 The SDS consists of 12 questions about willingness to participate in different activities with PWLE (eg, how willing would you be to spend time with, work with, or have a meal with a person with mental illness). The SDS was previously used in Nepal56 and adapted from sections of the Stigma in Global Context-Mental Health Study.57,58 A number of secondary PCP outcomes were also included. The mhGAP Attitudes Assessment examines stigmatizing beliefs and stereotypes (eg, people with mental illness are violent, contagious, or to blame for their illness). The Implicit Association Test (IAT)59 is a computer-based implicit measure of stigma adapted for use with stimuli appropriate for Nepali health care clinicians.60 The mhGAP Knowledge Assessment is a
+26-item true-false and multiple-choice test.61 Clinical competency in common factors of mental health care was evaluated with the Enhancing Assessment of Common Therapeutic Factors (ENACT) tool.62 The ENACT tool, developed in Nepal,63 is used by raters observing standardized role-plays of PCP trainees. Actors presented 1 of 3 vignettes (depression, psychosis, or alcohol use disorder). At the end of the role-play, PCPs were asked to provide a provisional diagnosis.
+The evaluation of diagnostic accuracy of actual patients was done approximately 14 to 22 months after training to give PCPs at least 1 year of supervised practice to establish their diagnostic skills. After PCPs made a new diagnosis of mental illness for a primary care patient, the accuracy of diagnosis was determined by a Nepali psychiatrist administering the Nepali-validated version of the Composite International Diagnostic Interview (CIDI)64 to the patient. The psychiatrist was blinded to the PCP’s diagnosis of the patient. The diagnoses of interest were the priority disorders from the mhGAP-IG training modules, which included depression, psychosis, alcohol use disorder, and epilepsy. For the purposes of the current analyses, we excluded epilepsy because of its nature as neurological disorder, and we focused onaccuracy of depression, psychosis, and alcohol use disorder diagnoses. No changes were made to the assessment tools after the trial commenced.
+Sample Size
+Following recommendations for the design of pilot studies which discourage between group hypothesis testing with small samples,65-67 we did not use inference testing (ie, measuring between-group effect sizes) as the criteria for determining the number of clusters and participants. Instead, we used all eligible primary care facilities in Chitwan district. Based on results of this pilot study, we will be able to make inferences to inform estimation for the coefficient of intracluster correlation (k) for a fully powered trial using primary care facilities as the unit of randomization with a comparable population of PCPs and patients. No interim analyses or stopping guidelines were planned.
+Randomization and Masking
+Randomization of primary care facility clusters was performed by the study statistician using a random number generator in Stata statistical software version 14 (StataCorp),68 with no restrictions or stratification. PCPs were included in the study group to which their health facility was randomized. To address demand characteristics typical in social contact interventions,25 PCPs were informed that the study was an overall mental health training evaluation, rather than specific to stigma reduction. Research assistants, psychosocial counselors who performed ENACT, and psychiatrists who performed the CIDI were blinded to the study group. No a priori unblinding specifications were established. Potential sources of contamination across groups were the movement of PCPs from a facility in one group to a facility in another group (eg, moving from a primary care clinic in the RESHAPE group to the control group).
+Statistical Analysis
+The quantitative outcomes of interest for PCPs were summarized descriptively using appropriate summary statistics (mean and standard deviation for continuous outcomes and numbers and proportions for categorical outcomes). Using the baseline measurement for PCPs (ie, prior to beginning the training), preliminary estimates of within- and between-cluster variances and intracluster correlation coefficients were estimated using 1-way random effects analysis of variance. Within-group changes over time were estimated using separate linear mixed models for each group, with a random intercept for health facility. Regarding missing data, only participants with data available at follow-up time points were included in analyses and no imputation was conducted for missing participants. No between-group comparisons were estimated owing to the pilot nature of the study.65-67 Analyses were conducted using Stata statistical software version 16 (StataCorp) from February 2020 to February 2021.69
+JAMA Network Open | Global Health	Collaboration With People With Lived Experience of Mental Illness to Improve Primary Care Services
+Results
+Participants
+Thirty-four facilities were eligible for randomization (Figure 2). For the 17 facilities allocated to mhGAP-IG trainingas usual, 45 PCPs were eligible. For the 17 facilities allocated to RESHAPE, 43 PCPs were eligible. PCP demographics are shown in Table 1. Among the overall sample of 88 PCPs, 75 (85.2%) were men and 67 (76.1%) were upper caste Hindus; the mean (SD) age was 36.2 (8.8) years (range, 21-56 years). Nine of the PCPs (10.2%) were physicians, whereas the remaining 79 PCPs (89.8%) were health assistants or auxiliary health workers.
+This cluster randomized clinical trial study was conducted as planned from February 7, 2016, to August 10, 2018. Training of PCPs took place from February 7, 2016, through May 18, 2016. PCP assessments took place from February 7, 2016, through July 4, 2017. Patient enrollment occurred from July 16, 2017, through December 31, 2017.
+area, and 6 did not attend the end line evaluation session. In the RESHAPE group, 6 PCPs were reassigned, 3 did not attend the end line session, and 1 retired. Table 1 includes the baseline demographics of PCPs who participated in the end line vs those who were lost to follow-up. The participants lost to follow-up were more likely to be younger (aged <30 years), have a medical degree (MBBS), be stationed ata primary health care center, and have fewer than 5 years of experience in health care services. Only 4 physicians with MBBS (44.4%) enrolled were retained, compared with 54 auxiliary health workers (81.8%) and 17 health assistants (70.8%). Because many health facilities had only 1 or 2 PCPs at baseline, the PCP dropouts led to a loss of 3 clusters in the control group and 2 clusters in the RESHAPE group (ie, a loss of 14.7% of the clusters). See eTable 1 in Supplement 2 for information on missingness of data.
+We also tracked reassignment of PCPs across facilities within the study. The government reassigned 2 control PCPs to RESHAPE facilities between baseline and midline (both PCPs participated in the midline assessment and 1 participated in the end line assessment). One RESHAPE PCP was reassigned to a control TAU facility between midline and end line; this facility had 3 control PCPs in the study at the time. Taking all of these transfers into account, at end line there were 4 control PCPs working alongside RESHAPE-trained PCPs, which suggests that 12% (n = 4) ofall control PCPs assessed at end line potentially experienced contamination of attitudinal and/or behavioral changes.
+Regarding PWLE who cofacilitated the RESHAPE trainings, 15 PWLE were trained in PhotoVoice, of whom 11 (73.3%) went onto cofacilitate RESHAPE trainings for PCPs. Among the 4 PWLE who dropped out during the PhotoVoice training process, reasonsfor dropout were family refusal (n = 1), time constraints (n = 1), symptom relapse (n = 1), and concerns for additional stigma by speaking in public (n = 1).53 Of the 2 RESHAPE-based trainings conducted, 8 PWLE participated in each training (approximately 2 PWLE for each priority disorder: depression, psychosis, alcohol use disorder, and epilepsy). Both trainings were above the 75% fidelity benchmark for antistigma components conducted by the PWLE and aspirational figures.
+Outcomes
+For the primary stigma outcome, SDS, there was a mean change from baseline to end line of -10.6 points (95% CI, -14.5 to -6.74 points) for PCPs in the RESHAPE group compared with -2.8 points (95% CI, -8.29 to 2.70 points) in the control group, where decreases in scores corresponded to decreases in reported social distance, ie, lower stigma (Figure 3 and Table 2). For mhGAP Knowledge, mhGAP Attitudes, and ENACT competencies, there was within-group improvement for both RESHAPE and control. For IAT, neither the control nor RESHAPE showed within-group improvement.
+At 4 months after training, diagnostic accuracy in standardized role-plays was 78.4% (29 of 37) in the RESHAPE group and 57.5% (23 of 40) in the control group (Figure 3 and Table 2). At 16 months after training, accuracy was 78.1% (25 of 32) in RESHAPE and 66.7% (22 of 33) in the control group. Patients newly diagnosed during the period of 14 to 22 months after training were enrolled in the study for assessment of diagnostic accuracy (patient demographics are in eTable 2 in Supplement 2). For actual patient diagnoses confirmed with the psychiatrist-administered CIDI, 72.5% (29 of 40) of patients in the RESHAPE group were correctly diagnosed and 34.5% (10 of 29) in the control group. The incorrectly diagnosed patients were false positives (ie, the PCP diagnosed them with a particular disorder and the psychiatrist did not confirm the diagnosis). Notably, 13 of 14 patients diagnosed with psychosis by the control PCPs were false positives, and 8 of 14 patients diagnosed with psychosis by RESHAPE-trained PCPs were false positives. Among the disorders, the disorder with the largest absolute difference between groups in diagnostic accuracy was depression in both standardized role-plays and actual patient evaluations. No serious adverse events were reported for PWLE, PCPs, or patients in either group.
+Discussion
+This pilot cRCT of a stigma reduction intervention for PCPs was conducted in collaboration with PLWE. The goal was to determine the feasibility and acceptability of study procedures in preparation for a full trial. All a priori benchmarks for progression to a full trial were met, including retention rates ofparticipants, limited missingness ofdata, high intervention fidelity, and a lack ofseriousadverse events. These quantitative findings for feasibility and acceptability support our previously published qualitative findings for RESHAPE.31,52,53 The preliminary findings suggest that RESHAPE may have the potential to reduce stigma among PCPs without introducing substantial risk of harm to PWLE collaboratingintrainings. Regardinggeneralizabilitytothe broaderfield ofstigma interventions, the
+potential trend of greater stigma reduction in the RESHAPE group is consistent with other findings for social contact.14,26,27
+Exposure to structured recovery testimonials from PWLE may help to increase accuracy of diagnosis by PCPs. This is important because the study revealed high rates of incorrectly diagnosing patients with psychosis (ie, false positives). Misdiagnosis, especially of psychotic disorders, increases exposure to medications with adverse effects, and misdiagnosis is costly and stigmatizing for patients and families.70-72 Although some misdiagnoses may be mitigated by improving attitudes of PCPs, it also draws attention to the need for greater supervision of diagnostic practices after mhGAP-IG training.
+Strengths and Limitations
+This study had some strengths and limitations. The pilot design addressed a number of the limitations raised about social contact intervention research25: use of a control group, trial registration, reducing demand characteristics by including a range of outcomes beyond stigma, and evaluating behavior change in the form of clinical skills. Examining long-term outcomes, specifically a 16-month follow-up, was also considerably longer than the 6-month follow-up of most antistigma interventions.14,27 The pilot findings suggest that RESHAPE may be beneficial across outcomes, except IAT, a few months after training, but the longer-term benefit compared with standard training (ie, at 16 months) may be limited to fewer domains (social distance and diagnostic accuracy).
+The long duration of our pilot was important to estimate actual retention rates of participants and clusters. By having a long duration and a large number of clusters, we found that 14.7% of the clusters did not have enough PCPs working on site for participation in the end point. If we had fewer clusters or a shorter duration in the pilot, we may not have been able to establish a reliable cluster dropout rate. We also found that PCPs who were physicians, younger (aged <30 years), and had fewer years of experience in health care (<5 years) were the most likely to drop out. This is likely due to government programs that place young physicians in rural areas after they complete training for brief assignments of only 1 to 2 years. In addition, physicians assigned to rural areas have been criticized for absenteeism in which they are working at the government health facilities.73-75 More than half of the physicians enrolled in the study dropped out, compared with 82% of auxiliary health workers and 71% of health assistants who were actively engaged in health services to their communities. Our pilot also revealed PCPs being transferred across study group health facilities (eg, RESHAPE to control and vice versa). Regarding contamination, 12% of control PCPs were working alongside RESHAPE-trained PCPs at the end line assessment, which may have influenced their attitudes and clinical practices. In a full trial, alternative strategies are needed for defining and retaining clusters, as well as preventing contamination.
+A study limitation was that mental health specialists conducting the trainings could not be blinded to the participation of PWLE in the trainings. The presence of PWLE may have impacted the psychiatrist trainer’s behavior in some manner. Because of this, a full trial should record the psychiatrist trainer’s fidelity to mhGAP-IG components to see if this differs between groups. Based on the lessons learned regarding strengths and limitations of the pilot trial, a full cRCT of RESHAPE is now under way in Nepal.76
+Another limitation of this study was that it focused on in-service health training of certified PCPs, which is the equivalent of continuing medical education courses. However, strategies are also needed to reduce stigma and improve mental health diagnostic skills of during preservice training (eg, in medical schools and auxiliary health worker vocational training programs). Evidence-based preservice stigma reduction programs are also lacking in LMICs.28 To effectively reduce the mental health care treatment gap in Nepal, preservice and in-service stigma reduction and mental health training are especially important for auxiliary health workers and health assistants who provide the majority of care in primary care settings.
+JAMA Network Open | Global Health	Collaboration With People With Lived Experience of Mental Illness to Improve Primary Care Services
+Conclusions
+This pilot cRCT met its prespecified feasibility and acceptability measures, and a larger cRCT is ongoing. Ultimately, the potential to collaborate with PWLE to reduce stigma and improve diagnosis is encouragingfor enhancing the success of mhGAP-IG implementation and more broadly for successful integration of mental health services into primary care settings around the world.

Colorectal, cervical and prostate cancer screening in.txt

@@ -0,0 +1,55 @@
+Introduction
+People with severe mental illness (SMI), such as schizophrenia and bipolar affective disorder, have higher rates of morbidity and mortality (Lawrence et al., 2013; Liu et al., 2017). As a result, their life-expectancy is 10-20 years lower than that of the general population (Lawrence et al., 2013; Liu et al., 2017), even though incidence rates of many cancers, including colorectal and prostate cancer, are similar between people with and without SMI (Kisely et al., 2008). Given the similar incidence rates, differences in risk factor prevalence (smoking, alcohol consumption, obesity) are less likely to be the cause of higher cancer mortality in those with SMI. One explanation might be that people with
+Population Health Department, QIMR Berghofer Medical Research
+Institute, Herston, QLD, Australia
+2School of Public Health, The University of Queensland, Herston, QLD, Australia
+3School of Medicine, The University of Queensland, Brisbane, QLD, Australia
+4Metro South Addiction and Mental Health Service, Brisbane, Metro South
+Health, QLD, Australia
+5Department of Gastroenterology and Hepatology, Princess Alexandra
+Hospital, Brisbane, QLD, Australia
+6Departments of Psychiatry and Community Health and Epidemiology, Dalhousie University, Halifax, NS, Canada
+Corresponding author:
+Karen M Tuesley, QIMR Berghofer Medical Research Institute, 300
+Herston Road, Herston, QLD 4006, Australia.
+Email: Karen.Tuesley@qimrberghofer.edu.au
+SMI present with more advanced cancer at diagnosis or receive less cancer-directed treatment, which could be due to either diagnostic delays, poorer access to cancer services or lower participation in cancer screening programmes (Kisely et al., 2013).
+Prior studies have investigated particular types of cancer screening in people with SMI, but these were mostly conducted in specific populations and results may not be broadly generalisable (Howard et al., 2010). Furthermore, some studies used self-reported participation in screening, which may not be optimal in this population (Fujiwara et al., 2017; Howard et al., 2010; Mo et al., 2014; Siantz et al., 2017), while others had no comparison group (Howard et al., 2010; James et al., 2017). The aim of this study was to investigate the frequency of colorectal, prostate and cervical cancer screening among people with and without SMI, throughout Australia, using a large, nationally representative administrative data set of 10% of the Australian population.
+Methods
+We conducted a retrospective cohort study using de-identi-fied administrative data from a sample of 10% of all Australians registered for Medicare. Medicare is Australia’s universal health care scheme that provides access to government-subsidised medical services (via the Medicare Benefits Scheme [MBS]) and prescriptions (via the Pharmaceutical Benefits Scheme [PBS]) for all citizens and permanent residents. The 10% data set included both MBS and PBS information. Within the MBS data, each specific medical service subsidised by the Australian government is denoted by an item number. The PBS data included details of all medicines dispensed to patients at government-subsidised prices. In accordance with the 2014 National Health and Medical Research Council Statement on ethical conduct in human research, ethics approval was not required to use these de-identified data. We adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for reporting observational studies (Von Elm et al., 2007).
+Defining exposure
+We used PBS data to define people with SMI (schizophrenia or bipolar affective disorder). In Australia, the most commonly prescribed medications for these conditions are lithium and second-generation antipsychotic agents. Lithium is specific to bipolar affective disorder and rarely prescribed for other conditions, while second-generation antipsychotics require an indication-specific authority code for subsidy through the PBS. These are almost solely for treatment of either schizophrenia or bipolar affective disorder (Supplementary Table 1). During the study period, the only other PBS indication
+for second-generation antipsychotics for adults was behavioural disturbance secondary to either dementia or autism, and this was restricted to risperidone. Our study only included people aged 18-69 years (see in the following). As the onset of dementia before 70 years is uncommon (<4% prevalence; Anstey et al., 2010), and the prevalence of autism in adulthood is also very low (<1% in 2012; Australian Institute of Health and Welfare [AIHW], 2017), the few people receiving risperidone for behaviour disturbance in dementia or autism would have been minimal.
+We classified a person as having an SMI once they had two prescriptions for one of these medicines dispensed within a 12-month period (Supplementary Table 1). PBS data from the year prior to cohort entry were used to determine exposure status at cohort entry. Participants who had not met the criteria for SMI prior to study commencement (January 2004) were considered unexposed to SMI until they received a second prescription for an SMI medication.
+Prior to 2012, lithium prescriptions were only recorded in the PBS for people who held a means-tested concession card, as before that, pharmacies only recorded subsidised prescriptions. As lithium was relatively cheap, people without a concession card would have paid the full cost without subsidy. To minimise any bias as a result of this differential inclusion of people of lower socioeconomic status, we treated people who used lithium, and no other SMI medication, as unexposed to SMI. We conducted sensitivity analyses to (a) treat lithium-only users as exposed from 2012 onwards, (b) include lithium-only users as exposed for the whole study period and (c) exclude lithium-only users entirely. All authority prescriptions for second-generation antipsychotics were recorded for the entire study period for both concessional and non-concessional patients because of the higher costs of these medicines. First-generation antipsychotics are below co-payment, do not require an authority prescription recording indication for use, and make up only a small proportion of prescribed antipsychotics (Hollingworth et al., 2010). They were therefore not used for analysis.
+Data were available to 31 December 2014, which was the study end date. We created study cohorts to examine screening for three cancers according to age-specific screening recommendations. In each case, screening is billed to Medicare and therefore captured by MBS records. We did not examine breast cancer screening as the Australian Breast Screen programme is not funded through Medicare. Although there is no formal prostate cancer screening programme in Australia and population-based screening is generally not recommended in clinical practice guidelines, prostate-specific antigen (PSA) testing is both commonly requested by male patients and recommended by clinicians (Pickles et al., 2016). Thus, differences in PSA testing between people with, and without, SMI may serve as a marker of access to preventive care.
+The cervical cancer screening cohort included all women aged from 18 to 69 years, the eligible age group for screening under the National Cervical Screening Programme. During the study period, most pap smears were performed by a general practitioner (GP) and billed to Medicare (Lew et al., 2012). The colorectal cancer screening cohort included all men and women aged 50-69 years, while that for prostate cancer was restricted to men aged 50-69 years. The MBS data did not include screening performed through the Australian National Bowel Cancer Screening Programme (NBCSP) that was progressively rolled out over the period covered by our study. This national programme commenced in 2006, and initially, only people turning 55 and 65 years old each year were invited to participate and sent a kit for faecal occult blood testing (FOBT). In 2008, this programme was extended to people in the year they turned 50 years and in 2013 to people turning 60 years (Jenkins, 2016). From 2015 (after the end of follow-up in this study), the programme was extended to people aged from 50 years up to 74 years. Over the transition period from 2006, GPs were still encouraged to screen the many people who fell outside of eligibility for the NBCSP by requesting FOBT via the MBS (Foreman, 2009).
+Participants entered our study on 1 January 2004 or the date they reached the minimum age for entry into each cohort. Participants left the study either on 31 December 2014 or when they turned 70 years. Person years were adjusted to include only 12 months before the first MBS record and 12 months after the last MBS record, given we did not have death records or records of participants arriving or leaving Australia. We removed participants with no MBS or PBS records within 12 months of study entry and exit (Figure 1). We performed sensitivity analyses to include
+person years for both 2 and 3 years from an individual’s first and last MBS record.
+Variables
+FOBT was used to define colorectal cancer screening, pap smears for cervical cancer screening, and PSA testing for prostate cancer screening using MBS item codes (Supplementary Table 2). It was possible for an individual to have multiple MBS items relating to a potential cancer screening test, therefore only one test per calendar year per participant was recorded as an incidence of cancer screening so as to exclude tests repeated for follow-up of an abnormality. For colorectal cancer screening, we recorded the incidence of each of the three specific FOBT item codes, which differed according to whether one, two or three samples were collected during a 28-day period. As noted previously, colorectal cancer screening through the Australian NBCSP was not recorded in the MBS records; we therefore only included FOBT organised by a medical practitioner outside of this programme. We only included the PSA item codes relating to probable cancer screening and not for tests performed as follow-up for previously diagnosed prostate disease.
+Covariates included age at study entry, gender, state of residence and average annual number of GP visits. As only the year of birth was included in the de-identified data, we used the year’s midpoint (30 June) to estimate age at entry into the cohort. The MBS data contained five states of residence categories (New South Wales/Australian Capital Territory, Victoria/Tasmania, Queensland, South Australia/ Northern Territory and Western Australia), and we defined state of residence as the last category provided in the MBS records for each participant. Participants with missing state data were excluded from the cohort (Figure 1).
+We estimated the average number of GP visits per year across the study period using the total number of GP visits for all calendar years, divided by the total number of calendar years that the participant was included in the cohort. If the participant was prescribed an SMI-defining medication, average GP visits for the exposed time included the year they were defined as having an SMI. Average GP visits were also used as a categorical variable, split by less than five and five or more visits per year.
+Statistical methods
+We used Poisson regression to estimate incidence rate ratios (IRR) and 95% confidence intervals (CIs) for the association between SMI and rates of FOBT, pap smears and PSA testing. We performed multivariable analyses adjusting for age at entry, state of residence and gender (for colorectal cancer screening), with and without average GP visits in the models, as prior studies have shown that the number of GP visits may be a mediator for the association between SMI and cancer screening. As a sensitivity analysis, we split the cohorts into the two categorical groups for average GP visits to explore the association between SMI and cancer screening for people with similar GP contact. We also stratified the data by age categories (50-59 and 60-69 years for FOBT and PSA, and 18-29, 30-30, 40-49, 50-59 and 60-69 years for pap smear screening), to investigate whether associations between SMI and cancer screening varied by age group.
+Additional sensitivity analyses were performed for the colorectal cancer screening cohort, by splitting the data into two study periods (2004-2006 and 2007-2014) given the introduction of the NBCSP in 2006, as well as each of the three FOBT MBS codes (66764, 66767 and 66770).
+Results
+There were 760,058 people in the colorectal cohort, 918,140 in the cervical screening and 380,238 in the prostate cancer screening cohorts (Figure 1). Approximately 2% of each cohort had a diagnosis of SMI (Table 1). The maximum follow-up was 11 years, with a median of 7.5 years for the colorectal and prostate cancer screening cohorts and median of 11 years for cervical cancer screening. Table 1 shows the characteristics of the cohorts by SMI status. For all cohorts, the average number of GP visits was higher for people with SMI than those without SMI (Table 1). Age at entry and distribution by state were similar between the groups.
+The associations between SMI and cancer screening are shown in Table 2. Adjusting for age at entry, state and gender (FOBT only) did not materially change the results, and IRRs are adjusted for these factors unless otherwise stated. Having SMI was associated with lower rates of pap smears (IRR=0.83, 95% CI = [0.82, 0.84]) and PSA testing (IRR=0.83, 95% CI = [0.81, 0.85]) compared to people
+without SMI. When the average number of GP visits was included in the model, the IRRs declined further for pap smears (IRR = 0.74, 95% CI = [0.73, 0.75]) and PSA testing (IRR = 0.72, 95% CI = [0.70, 0.74]).
+Having SMI was associated with slightly higher rates of FOBT compared to not having SMI (IRR = 1.15, 95% CI = [1.10, 1.20]) although overall, FOBT rates were low for both people with and without SMI (2.6 and 2.2 per 100 person years, respectively). However, after adjusting for average number of GP visits, people with SMI had lower rates of FOBT (IRR = 0.90, 95% CI = [0.86, 0.94]). To investigate this further, we dichotomised the cohort into those with an average of less than five GP visits per year and those with an average of five or more GP visits per year. In those with an average of less than five GP visit per year, SMI was associated with lower rates of FOBT (IRR=0.83, 95% CI = [0.73, 0.94]). By contrast, in those who visited their GP an average of five or more times per year, SMI was associated with slightly higher rates of FOBT (IRR = 1.04, 95% CI = [1.00, 1.09]). Incidence rates for FOBT during 2007-2014 were close to double the rates in 2004-2006, but this did not substantially alter our unadjusted or adjusted IRRs (Supplementary Table 3).
+Our analyses stratified by age showed that having SMI was consistently associated with reduced rates of pap smear screening in the 30-39, 40-49, 50-59 and 60-69 year age groups but the association was weaker in women aged 1829 years (IRR=0.96, 95% CI = [0.93, 0.99]; Supplementary Table 5). When we adjusted for average GP visits, the association between having SMI and pap smear testing was similar across the different age groups (Supplementary Table 5) and was consistent with our main analysis. SMI was associated with slightly lower rates of PSA screening in men aged 60-69 years (IRR=0.78) than those aged 5059 years (IRR=0.84), although rates in both groups were still statistically significantly lower than in those from the general population (Supplementary Table 5). Adjusting for GP visits did not have an effect on the association between SMI and PSA testing in either age group. With respect to FOBT, rates were only statistically significantly higher in those with SMI among those aged 50-59 years (IRR = 1.20, 95% CI = [1.13, 1.28] compared with IRR = 1.06, 95% CI = [0.99, 1.13] for those aged 60-69 years, Supplementary Table 5). When we adjusted for GP visits, results for both age groups were consistent with our main analysis.
+People with SMI were more likely to have one or two sample FOBT tests (rather than three) compared to people without SMI. However, there was not a significant difference in screening rates for FOBT screens with three samples taken, which was also the most commonly performed procedure (Supplementary Table 4).
+Finally, other sensitivity analyses extending person years to 2 and 3 years from the first and last MBS date, to treat lithium-only users as exposed from 2012 onwards, to include lithium-only users as exposed for the whole study
+period, and to exclude lithium-only users entirely did not materially alter the results (see Supplementary Tables 6 and 7). Across the three cohorts, 65-75% of identified lithium users were also prescribed another SMI-defining medication.
+Discussion
+Main findings
+These results from a large, national longitudinal study showed that people with a SMI had significantly lower rates of pap smears and PSA testing. Overall, we did not see lower rates of FOBT, although rates were significantly lower among those with SMI who visited a GP on average less than five times per year.
+Context and implications
+Our results indicate that cervical and prostate cancer screening rates in people with SMI are lower than those from the general population, as has been suggested by some but not all of the smaller studies in less representative populations (Happell et al., 2012; Howard et al., 2010). The findings are also similar to those of a cohort study of pap smears in women with schizophrenia that was restricted to one Canadian province (Martens et al., 2009). Our results may partly explain why the cancer mortality to incidence ratio in these cancers is higher for people with SMI compared to the general population (Kisely et al., 2008) and that those with SMI are more likely to have metastases at presentation than those without SMI (Kisely et al., 2013). Even in the case of PSA testing where the value of screening is contested (Catalona, 2018), reduced uptake in people with SMI may serve as a marker of their access to preventive care in general. People with SMI may have a number of comorbidities or competing needs, and medical practitioners may therefore be less likely to consider preventive screening during consultations. There also may be cognitive and behavioural challenges with patients with SMI making it more difficult to gather medical histories and make treatment plans (Druss et al., 2002). In addition, clinicians may attribute emerging somatic symptoms to the underlying psychiatric disorder resulting in missed diagnoses, sometimes termed ‘diagnostic overshadowing’(Pelletier et al., 2015). It is also possible that people with SMI are treated differently by medical professionals with negative attitudes towards SMI patients leading to disparities in care (Walker et al., 2015).
+By contrast, FOBT rates were only lower in people with SMI compared to people without SMI among those who visited GPs less than five times per year. These findings are similar to a study restricted to US veterans that found that frequency of GP contact had an effect on colorectal cancer screening rates (Kodl et al., 2010). Poor access to health care may therefore have a greater impact on cancer prevention for people with SMI than the general population. The higher rates of FOBTs for people with SMI who visited doctors more
+frequently may also reflect the use of FOBT as a diagnostic tool, rather than screening test. GPs may opt for less invasive approaches to colorectal symptoms because of potential concerns that somatic symptoms reported by people with SMI might actually be a manifestation of their disease (psychosis), as well as the greater challenges of facilitating colonoscopy in those with SMI. We found some evidence of this in that those with SMI had higher rates (compared to those without SMI) of one or two FOBTs in an episode rather than the three recommended for screening. We did not have information on colonoscopy rates after FOBT but FOBT will not lead to a reduction in colorectal cancer mortality if not followed up with further diagnostic testing and treatment (Liss et al., 2016). This needs investigation in future studies.
+Strengths and limitations
+Our study had a number of limitations. We used medication to define SMI rather than medical records. While this may have created some misclassification of SMI status, it allowed us to investigate screening in a very large nationwide population. Importantly, the PBS item codes we used to define our exposed population are largely restricted to SMI. The one exception is risperidone, which is also indicated for behavioural disturbance in people with dementia or autism. However, these disorders are uncommon in the age groups included in our study (AIHW, 2017; Anstey et al., 2010). We were also unable to capture first-generation antipsychotics, so there is the potential that our control group included people with SMI taking first-generation antipsychotics only, thus making our estimates more conservative. However, clinical practice guidelines in Australia advise the use of second-generation antipsychotics in the first instance (Galletly et al., 2016), and the use of first-generation antipsychotics (for any indication) is decreasing, with a reduction from 39% to 23% of all antipsychotics prescribed in the period covered by the study (Hollingworth et al., 2010). Lithium use was also inconsistently recorded in the PBS before 2012, but our sensitivity analyses exploring this issue were not materially different from the main findings suggesting the effect was minimal. We were unable to adjust for sociodemographic measures other than age, gender and state of residence. Given that people with SMI are a generally disadvantaged group (Lawrence et al., 2013), other socioeconomic factors could have been mediators on the pathway between SMI and cancer screening. Finally, we did not have data on FOBT performed as part of the NBCSP. Nevertheless, we saw an increase in FOBTs performed after the commencement of NBCSP in 2006, and this may be due to an increase in testing for age groups not covered in the staggered roll out of the NBCSP, as well as additional guidance for GPs recommending FOBT screening during this time (Foreman, 2009). It is unknown whether FOBT rates through the NBCSP would be similar between those with and without SMI, although studies have
+found that other disadvantaged groups are less likely to participate in NBCSPs (He et al., 2017).
+This study has a number of strengths. It is the first nation-wide large-scale cohort study of cancer screening for people with SMI and had a follow-up of 11 years. The 2% prevalence of SMI is in keeping with rates reported in the World Health Organization (WHO) 10-country study (Jablensky et al., 1992). Use of linked administrative health records reduced the very real potential for selection bias that can occur when conducting research with people with SMI and allowed for accurate capture of screening tests performed. This study also provided cancer screening rates for a cohort with access to a universal health care system.

Combined association of central obesity and depressive symptoms with risk of heart disease A prospective cohort study.txt

@@ -0,0 +1,36 @@
+1.	Introduction
+Depressive symptoms are common mental disorders, with more than 264 million people suffered, leading to disability, suicide and death. Meanwhile, the prevalence of obesity is increasing and obesity is also becoming an emerging clinical and public health burden (Disease et al., 2018). Worse, obesity and depressive symptoms often co-occur at the individual level, and there has been shown a bidirectional relationship between obesity and depressive symptoms. Some data on the prevalence of comorbidity suggested that nearly 43% of adults with depressive symptoms have obesity (Pratt and Brody, 2014) and the highest prevalence of depressive symptoms was observed among obese adults (24%) (Carey et al., 2014). A meta-analysis of 19 studies showed that adults with depression had a 37% increased risk of being obese, and those who
+were obese had an 18% increased risk of depression (Mannan et al., 2016). Due to their co-occurrence, randomized clinical trials have gradually been conducted to explore integrated interventions and treatment for obesity and depressive symptoms (Linde et al., 2011; Ma et al., 2019; Pagoto et al., 2013).
+The coexistence of obesity and depressive symptoms was associated with increased health care use and costs (Nigatu et al., 2017), as well as many adverse health-related outcomes (Haregu et al., 2020; Licinio and Wong, 2003; Nigatu et al., 2016). These health outcomes are often worsened when obesity and depressive symptoms co-occur (Haregu et al., 2020; Licinio and Wong, 2003; Nigatu et al., 2016). Therefore, examining the interaction or combined effect of obesity and depressive symptoms may have important implications for alleviating disease burdens.
+Heart disease is a major concern of premature mortality and increased health care costs. The burden of heart disease, in number of disability-adjusted life years and deaths, continues to rise globally (Roth et al., 2020). Although obesity and depressive symptoms are known risk factors for heart disease (Yusuf et al., 2020), evidence on their possible synergistic effect on heart disease is scarce. A prior study suggested that the higher body mass index (BMI) was a stronger predictor of incident CVD in adults with depression than lower BMI (Polanka et al., 2018). Besides, one Germany cohort study showed that the combination of depressive symptoms and obesity had a significant higher risk of incident coronary heart disease (CHD) in men but not in women (Ladwig et al., 2006). The other cohort study conducted in Korea showed depression comorbid with overweight amplified the risk of heart diseases (Park et al., 2020). Both BMI and waist circumference (WC) are strongly and continuously associated with the risk of heart disease (Wormser et al., 2011). However, no investigation has examined central obesity, where the accumulation of abdominal fat has detrimental metabolic consequences and serves as an independent risk factor for cardiovascular events (Cornier et al., 2011). We aimed to explore the combined association of central obesity and depressive symptoms with the risk of heart disease in a national prospective cohort study of the Chinese population.
+2.	Methods
+2.1.	Study population
+Participants enrolled in our study were from the China Health and Retirement Longitudinal Study (CHARLS). The details of the CHARLS
+have been previously described (Zhao et al., 2014). In brief, CHARLS is a nationally representative longitudinal study by recruiting residents (>45 years) from 150 county-level units across 28 provinces in China. The baseline survey was conducted on 17,708 participants in 2011-2012 (wave 1), with a response rate of 80.5%, and involved self-administered questionnaires and physical examinations. Participants were followed up every two-three years until 2018 (wave2: 2013, wave3: 2015, and wave4: 2018). The follow-up rate is 76.6% in wave 4, 2018. (Zhao et al., 2020). People with a history of heart disease, stroke, or cancer at baseline were excluded. We further excluded people who had no data on age, sex, waist circumference, or depressive symptoms. In addition, people with an abnormal distribution of BMI < 14 or > 40 in the general population were further excluded. Finally, 10,722 Chinese men and women were eligible for this analysis (Fig. 1).
+Ethics approval for CHARLS was obtained from the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015). All the participants have completed a written informed consent since the survey began.
+2.2.	Assessment of central obesity
+The measurement of WC, height, and weight were conducted by well-trained staff in the physical examinations. The WC was measured with a soft tape that circled the subject horizontally at the navel level. Participants breathed calmly at the standing pose, and when holding the breath at the end of the expiration, the staff took a reading. According to the International Diabetes Federation consensus statement, men with a WC of > 90 cm and women with a WC of > 80 cm were considered as having central obesity in the Chinese population (Alberti et al., 2006).
+2.3.	Assessment of depressive symptoms
+Depressive symptoms were identified using the 10-item Center for Epidemiologic Studies Depression Scale (CESD-10). The CESD-10 contains 10 items and belongs to a 4-point rating scale: none or rarely, some days (1-2 days), occasionally (3-4 days), and most or all of the time (5-7 days). The total score ranges from 0 to 30, with higher scores indicating a higher level of depressive symptoms. Previous studies showed the internal consistency of the CESD-10 was good with the Cronbach alpha of 0.79 (Lian et al., 2021). Previous studies have confirmed that a cut-off value of 10 provides the optimal threshold to define clinically significant depressive symptoms (Cheng and Chan, 2005; Jing et al., 2020; Lian et al., 2021).
+2.4.	Ascertainment of heart disease
+Ascertainment of heart disease incidence was based on self-reported questionnaires. In the baseline survey, participants were asked by the question of “have you been diagnosed with heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems by a doctor?” Those who answered “yes” were further asked when the condition was first diagnosed and which treatment was taken (Chinese traditional medicine, Western modern medicine, or other treatments). In the following surveys, the same questions were used to identify heart disease cases.
+2.5.	Assessment of covariates
+Data on sociodemographic characteristics, disease histories, lifestyle and behavior risk factors were based on interviewer-administered questionnaires, including age, sex, living area, marital status, education level, physical activity, smoking, drinking status, and history of hypertension, dyslipidemia, and diabetes that were accessed by selfreport of doctor diagnosis.
+2.6.	Statistical analysis
+Baseline characteristics of participants according to central obesity and depressive symptom status, presented as mean values (SDs) for continuous variables or percentages for category variables, were calculated using generalized linear regression (SAS GLM procedure) with adjustment for age at study entry. The differences of obtained values were tested using the same SAS procedure. The duration of follow-up for each participant was from the date of baseline survey (2011-2012) to the date of heart disease incidence (based on selfreported first diagnosed date) or the last survey (2018), whichever came first. Participants who did not develop heart disease during follow up and lost to follow-up were censored. All participants were divided into 4 groups: without central obesity and depressive symptoms, with central obesity alone, with depressive symptoms alone, or with both the two conditions. Person-years were calculated as the sum of follow-up duration of single participant for each group. Cox proportional hazard regression (SAS PHREG) was used to obtain the hazard ratios (HRs) with 95% confidence intervals (CIs) of heart disease in relation to central obesity and depressive symptom status, with the group having no central obesity or depressive symptoms being treated as the reference group. We adjusted age only in model 1. In model 2, we further adjusted for established cardiovascular risk factors to prevent distorting the real effect of central obesity and depressive symptoms on the incidence of heart disease risk, including sex, living area (urban or rural), level of highest education (primary school or lower, middle school, high school, college or higher), marital status (married, divorced, widow, or unmarried), history of hypertension (yes or no), history of dyslipidemia (yes or no), history of diabetes (yes or no), smoking status (never, former, or current), drinking status (never, former, less than once per month, once or more per month), and level of vigorous activity (0, 1-3,
+4-6, or 7/per week). The combined association of central obesity and depressive symptoms with incidence of heart disease was calculated using the relative excess risk due to interaction (RERI), with an additive model being employed: RERI = HRAB-HRA-HRB+1 (Andersson et al., 2005). We also conducted pre-defined stratified analyses according to sex and age. To test the robustness of the associations, we further performed sensitivity analyses by further adjusting for use of anti-depressant medication, by using a cut-off of 12 points in CESD-10 for assessment of depressive symptoms, by excluding individuals with a short length of follow-up (< 3 years) and by excluding individuals with a previous history of diabetes, hypertension and dyslipidemia, respectively. All analyses were carried out using SAS software version 9.4 and STATA version 16. P values < 0.05 were considered statistically significant.
+3.	Results
+Table 1 presents the baseline characteristics of 10,722 men and women according to the status of central obesity and depressive symptoms. Among all the participants, 5183 people (48.4%) had central obesity at baseline. Compared with people without central obesity, those with central obesity tended to be slightly younger and were more likely to be women, urban residents, and have a history of hypertension, dyslipidemia, and diabetes, but were less likely to smoke, drink, and do vigorous activity. The prevalence of depressive symptoms (25.9% vs 26.7%) did not substantially differ among people with and without central obesity. Among 10,722 participants, 2819 people (26.3%) had depressive symptoms. Compared with people without depressive symptoms, those with depressive symptoms were slightly older and were more likely to be men, rural residents, less educated, and have a lower BMI and waist circumference, but were less likely to smoke and drink.
+In the cohort, 3853 (35.9%) had central obesity alone, 1489 (13.9%) had depressive symptoms alone, and 1330 (12.4%) had both the two conditions. During 7 years of follow-up, we identified 1080 cases of heart diseases. Compared with people without central obesity and depressive symptoms, the age-adjusted HRs (95% CIs) of heart disease were 1.77 (1.52, 2.06) for those who had central obesity alone, 1.45 (1.19, 1.77) for those who had depressive symptoms alone, and 2.42 (2.02, 2.89) for those who had both central obesity and depressive symptoms (Table 2). After adjustment for covariates including demographic factors, disease histories, and lifestyles, the associations were attenuated but remained significant; the corresponding multivariable-adjusted HRs (95% CIs) were 1.39 (1.18, 1.64), 1.44 (1.18, 1.77), and 1.88 (1.55, 2.30), respectively (Table 2 and Fig. 2). However, the relative excess risk for heart disease due to interaction between central obesity and depressive symptoms was not significant (RERI = 0.05 [-0.33, 0.44]).
+We next performed stratified analysis by sex and age. The multivariable-adjusted HRs of heart disease for the group with both central obesity and depressive symptoms were somewhat greater in men than that in women (2.32 [1.60, 3.37] vs. 1.75 [1.35, 2.27]) and in middle-aged people than that in elderly people (2.11 [1.59, 2.80] vs. 1.66 [1.25, 2.20]). Tests for additive interaction between the two conditions yielded non-significant RERIs for either group.
+Sensitivity analyses were performed by further adjusting for use of anti-depressant medication and by using a cut-off of 12 points in CESD-10 for assessment of depressive symptoms. Other sensitivity analyses that excluded participants with a follow-up < 3 years, and people with a history of diabetes, hypertension and dyslipidemia, respectively, were also performed. Overall, the sensitivity analyses yielded very similar results (Supplementary Figure).
+4.	Discussion
+In this study, we examined the combined association of central obesity and depressive symptoms with the risk of heart disease in a large
+cohort of Chinese adults. We found that there was a heart disease risk gradient with the coexistence of central obesity and depressive symptoms, and the coexistence of the two conditions were associated with an 88% increased risk of heart disease than the absence of either condition. The combined association in men was more evident than that in women. However, no significant interaction on an additive scale between the two conditions was detected.
+Our study extended the evidence that the coexistence of obesity and depressive symptoms can potentially increase the risk of heart disease. In line with our results, a nationwide Korean cohort study reported the combination of overweight (BMI > 23 kg/m2) and depression had a higher risk (HR, 1.63; 99% CI, 1.29-2.07) of incident ischemic heart disease compared to those without either condition, which was much higher than depression (HR, 1.16; 99% CI, 0.81-1.67) or overweight (HR, 1.28; 99% CI, 1.19-1.36) alone (Park et al., 2020). A prior Germany cohort study demonstrated that only men with obesity (BMI > 30 kg/m2) and depressive symptoms reach the statistical significance for CHD incidence, where the HR was 2.32 (95%CI, 1.45-3.72) compared to those with normal BMI and no depressive symptoms, but not statistically significant in women (1.84, 95% CI 0.79-4.26) probably due to lack of power associated with low event rates (Ladwig et al., 2006). These results had important implications for the prevention and public health measures of CVD, where weight management, including general and abdominal obesity combined with depression treatment should be targeted to the high-risk groups. A recent randomized clinical trial tested a collaborative care intervention for coexisting obesity and depression by integrating behavioral weight loss treatment (dietary changes and physical activity), problem-solving therapy, and antidepressant medications for a duration of 12 months, resulting in improvements of both conditions (Ma et al., 2019). However, further studies are needed to verify whether the improvements will translate to important health outcomes, such as CVD over longer periods.
+Obesity and depressive symptoms may share common biological pathways for the development of CVD (Milaneschi et al., 2019), which might contribute to the observed risk gradient of heart disease. Previous studies linked central obesity with CVD via the inflammatory process, where adults with central obesity exhibit elevated proinflammatory cytokine levels that play important roles in endothelial dysfunction, hypertension, and atherosclerosis (Ellulu et al., 2017). The inflammatory markers, including leptin, tumor necrosis factor (TNF)-a, and interleukin-6 (IL-6) are mainly secreted from white adipose tissue in the abdomen (Shelton and Miller, 2011). Additionally, depressive symptoms are associated with decreased parasympathetic activity in the autonomic nervous system that could trigger inflammation response (Kop and Gottdiener, 2005). Thus, the inflammation response caused by obesity might be amplified or prolonged. Moreover, the high concentrations of proinflammatory cytokines caused by inflammation response
+may in turn enhance hypothalamic-pituitary-adrenal axis activity (Penninx et al., 2003), which is engaged in activating the release of glucocorticoids, with consequent increases in heart rate, blood pressure, and lipid metabolism abnormity (Mello et al., 2003), which are directly linked with the progression of CVD.
+On the other hand, behavioral pathways might be another potential mechanism. Obesity, especially the central or visceral type, is a predisposing factor for the development of several chronic diseases, such as hyperlipidemia, hypertension, and diabetes (Sowers, 2003). Meanwhile, depressive symptoms might interfere with adherence to self-care behaviors and treatments for these chronic diseases, including weight management (Berntson et al., 2015), poor adherence to recommended diet and physical activity changes (Berntson et al., 2015; Sumlin et al., 2014), and medication (Grenard et al., 2011), which may augment the progression of CVD among patients with obesity. Overall, the overlap of obesity and depressive symptoms are synergistic in terms of deterioration of cardiovascular function.
+To our knowledge, this is the first study to evaluate the combined association of obesity and depressive symptoms with incident heart disease among the Chinese population. The strengths of the present study included a large nationally representative sample of Chinese adults with up to a 7-year follow-up. Moreover, the assessment of obesity was firstly based on WC that reflected visceral fat accumulation, which could better predict obesity-related outcomes than subcutaneous fat measured by BMI (Janssen et al., 2004). However, several limitations to this study need to be acknowledged. First, the assessments of heart diseases and other medical history were defined by self-reported of doctor diagnosis, probably resulting in measurement error to our results. However, it has been proved to be with relatively good specificity and positive predictive values of self-reported illness in cohort studies (St Sauver et al., 2005; Yuan et al., 2015). Moreover, potential misclassification of heart disease could have occurred since the specific subtypes, such as CHD, heart failure, valvular heart disease, and other subtypes were not defined. Second, the duration or detailed kinds of medicine of depressive symptoms and obesity were not considered in these analyses, which might interfere with the risk of heart disease. Third, the details on depressive symptomology were lacking, though depression subtypes are differently associated with cardio-metabolic risk factors (Lasserre et al., 2017).
+5.	Conclusion
+In conclusion, our study provided evidence that the coexistence of central obesity and depressive symptoms were associated with a substantially increased risk of heart disease compared to those without these two conditions. This finding suggested the possibility that the integrated screening, monitoring, and treatment of obesity and depressive

Community-facility-and-individuallevel-outcomes-of-a-district-mental-healthcare-plan-in-a-lowresource-setting-in-Nepal-A-populationbased-evaluationPLoS-Medicine.txt

@@ -0,0 +1,78 @@
+Introduction
+Mental health is part of the Sustainable Development Goals, which set an agenda for improved treatment coverage by 2030 [1]. Treatment contact coverage is defined by the ratio of people who have contacted the service to the total target population in need of that service [2]. Increasing treatment coverage addresses the vast gap between availability of, and needs for, mental healthcare, especially in low- and middle-income countries (LMICs) [3,4]. The question is how to go about increasing coverage at a population level, especially in rural areas where there is little to no mental healthcare infrastructure. In keeping with the framework established by Tanahashi, which presents different levels of coverage related to the different stages of service provision [2], the fundamental issues underlying this question are (1) the allocation of resources in order to serve the maximum number of people, (2) the extent to which services are reaching the people they are intended for, and (3) the extent to which the services meet the people’s needs [2].
+The integration of mental healthcare in community and primary healthcare settings has been advocated as a strategy to reduce the treatment gap in LMICs. The call for decentralised mental healthcare integrated into general health service settings has been made since the early 1970s, and this strategy was implemented through the WHO Collaborative Study on Strategies for Extending Mental Health Care [5]. Although there was limited success in implementing this strategy in LMICs during the following decades, renewed efforts have been made more recently. The World Health Organization (WHO) has developed the Mental Health Gap Action Programme (mhGAP) intervention guide, providing evidence-based clinical guidance for health workers to detect and diagnose mental illness [6]. Furthermore, recent reviews demonstrate promising results for psychological treatments by non-specialists in LMICs [7,8]. Task-sharing strategies are currently being adapted and implemented in many LMICs [9]. Yet, to date, there are few evaluations of coverage of mental health programmes [10], and to our
+knowledge none that combines evaluation methods at the community, facility, and individual levels to assess the impact of district mental healthcare plans (MHCPs). The aim of this report is to evaluate contact coverage, detection, and treatment outcomes as a result of a complex multi-component district-level mental healthcare programme for adults in Nepal.
+Methods
+Setting
+The Programme for Improving Mental Health Care (PRIME) is a multi-country research programme that implements and evaluates district-level MHCPs in Ethiopia, India, Nepal, South Africa, and Uganda [11]. In Nepal, PRIME was implemented in Chitwan, a district in the south of the country with a total population of 579,984. During the evaluation phase the programme covered 10 primary healthcare facilities. Before the implementation of PRIME, mental health services were restricted to the district-level hospital. The Nepal health system consists of (1) district hospitals for specialised care, (2) primary healthcare centres for general medical care and first referral from health posts, and (3) the village-level health posts for basic health services. The major challenges in the existing health system ahead of implementing the MHCP were the lack of a formal government focal point for mental healthcare, the lack of basic psychotropic medicines in the essential medicines list, and the frequent transfer of primary health workers [12].
+Interventions
+The MHCP that was developed and implemented in Nepal, in partnership with the Ministry of Health, has been described in detail elsewhere [13]. In summary, the MHCP comprised interventions at the community, health facility, and health service organisation levels—see Table 1. The community-level packages included community sensitisation, proactive case detection [14], and adherence support through home-based care. In addition, community counsellors were trained to provide the Healthy Activity Programme [15] for depression and Counselling for Alcohol Problems [16] for AUD. The facility-level packages included training and supervision for health workers to detect, diagnose, and initiate treatment (i.e., emotional support, psycho-education and psychotropic medication) for individuals with a diagnosis of a priority disorder (i.e., depression, psychosis, AUD, and epilepsy) following the mhGAP intervention guide [6]. In most LMICs, epilepsy is considered a psychiatric condition and is treated by mental health specialists. Because of this, the WHO mhGAP includes epilepsy in its priority mental health conditions in the intervention guidelines. Based on our priority-setting activity in Nepal [ 17], we determined that epilepsy should also be considered to be a priority mental health condition that could be treated in primary care settings. Finally, health-service-organisation-level packages included ensuring reliable supply of psychotropic medication, referrals to specialised care, and mechanisms for monitoring, capacity building, and resource mobilisation. For all services, regular ongoing supervision was part of the MHCP. Different types of service providers were involved in implementing the interventions. At the health facility, medical officers (5 to 6 years of training), health assistants (3 years of training), and auxiliary health workers (15 months of training) were involved in assessment, diagnosis, and management of priority mental health conditions. The staff nurse and auxiliary nurse mid-wife (18 months to 3 years of training) were responsible for providing brief psychosocial support in the health facilities. At the community level, counsellors are a new cadre of psychosocial workers trained by non-governmental organisations, responsible for providing psychological treatment to those referred by primary health workers. Female community health volunteers were responsible for proactive case detection and home-based care.
+Adapted from [13].
+CAP, Counselling for Alcohol Problems; CIDT, Community Informant Detection Tool; FCHV, female community health volunteers; HAP, Healthy Activity Programme; mhGAP, Mental Health Gap Action Programme; n/a, not applicable.
+https://doi.org/10.1371/journal.pmed.1002748.t001
+Study designs
+This paper presents the primary results of a collection of study designs, in order to present findings for each component in the process of evaluating the above-mentioned district MHCP: a community study, routine service utilisation data, a facility study, and cohort studies —described below (see Fig 1 and Table 2). The study designs and analysis plans have been described in detail elsewhere [13,18-21]; summaries are presented below.
+We will structure the presentation of methods according to the 4 components of the service delivery pathway: (1) contact coverage of primary care mental health services, (2) detection of mental illness among participants presenting in primary care facilities, (3) initiation of minimally adequate treatment after diagnosis, and (4) the outcomes of patients receiving primarycare-based mental health treatment.
+Evaluating changes in contact coverage. We conducted a community study to determine whether adults affected by depression or alcohol use disorder (AUD) were more likely to contact a health worker for help coinciding with the PRIME implementation period. A detailed description of the aims, design, recruitment, and questionnaire are available [19]. Briefly, 2
+AUDIT, Alcohol Use Disorders Identification Test; HMIS, health information management system; MNS, mental, neurological, and substance abuse; n/a, not applicable; OPD, outpatient department; PANSS, Positive and Negative Syndrome Scale; PHQ-9, Patient Health Questionnaire-9 item; SIP-2R, Short Inventory of Problems-Revised; WHODAS, WHO Disability Assessment Schedule.
+https://doi.org/10.1371/journal.pmed.1002748.t002
+population-based cross-sectional surveys with independent samples were conducted, one before and one 30 months after implementation started. With 2,000 participants per round, the study had 80% power to detect a change in contact coverage from 5% to 25% among probable cases for each disorder, which we estimated would be 10% of the sample. Of the randomly selected adults (16 years and older, following Nepal legal classification) from randomly selected households in the implementation area, 99% provided informed written consent. The field workers orally administered a structured questionnaire that contained sections on demographic characteristics, food security, depression screening, depression symptoms in the past 12 months, and AUD screening. A probable case of depression had a PHQ-9 screening score of 10 or more or had depression-associated symptoms for at least 2 weeks in the past year [22]. A probable case of AUD had an AUDIT screening score of 9 or more [23]. Probable cases were asked whether they had contact with different health workers in the past 12 months, including non-specialist providers (e.g., medical officer, health assistant, auxiliary health worker) in government clinics. The timing of the data collection was as follows: baseline between May and July 2013 and endline between December 2016 and February 2017.
+In addition, we used 1-year routine service utilisation data to assess change in contact coverage for all 4 priority disorders (depression, AUD, epilepsy, and psychosis). Change in contact coverage was calculated as the number of cases diagnosed with mental illness in 10 health facilities for a period of 12 months before the start of the MHCP and for 12 months during the implementation of the MHCP (baseline: 1 January-31 December 2013; endline: 25 August 2014-24 August 2015). The reasons for using both methods for assessing changes in contact coverage are that (1) service utilisation data were available for all 4 disorders, whereas the community survey only focused on depression and AUD, and (2) we were aware of the risk of being underpowered in the community survey, due to limited financial resources to conduct the survey.
+Evaluation of changes in health workers’ detection of mental illness and initiation of adequate treatment. We conducted a facility study to determine whether adult attendees of primary healthcare facilities who were affected by depression or by AUD were more likely to be detected and adequately treated by clinicians during the PRIME implementation period [21]. Three cross-sectional surveys with independent sampling were conducted: before MHCP
+implementation and approximately 6 months and 24 months after initiating the MHCP. In the 10 health facilities, research staff recruited adults seeking outpatient services. All adult outpatients who were capable of providing informed written consent and who did not have an emergency medical problem were eligible for study recruitment. Among the eligible adult outpatients, 95% provided informed consent. In a private area adjacent to the waiting room, field workers verbally administered a structured questionnaire that contained sections on demographic characteristics and screening for depression and AUD. All participants who screened positive and a 10% random selection of screen-negative participants were given a consultation form for their clinician to complete and return to the participant. The form contained open-ended entries for diagnoses, treatments, advice, and referrals. A field worker made a copy of the form immediately after the consultation. A psychiatrist on the research team used the copy to determine whether each participant had been clinically diagnosed with depression or with AUD, and if so, whether there was evidence of minimally adequate treatment provision following mhGAP treatment guidelines. The timing of the data collection was as follows: baseline between September 2013 and February 2014, midline between August 2014 and August 2015, and endline between May and December 2016.
+Evaluation of changes in treatment outcomes. We conducted 4 cohort studies to assess whether patients diagnosed with depression, AUD, psychosis, or epilepsy benefitted from receiving treatment under the MHCP. Patients were followed up for 1 year, to assess change in symptom severity and functional impairment, using a before-and-after comparison without control groups. A detailed overview of the methods has been previously published [20]. Briefly, individuals were eligible for inclusion in the treatment cohorts if they were diagnosed with 1 of the 4 priority conditions by a primary health worker in the health facilities implementing the MHCP. In addition, participants needed to be adults, living in the study district Chitwan, and willing to provide informed consent. For participants with psychosis, a caregiver was also recruited into the study to participate in a caregiver component of the study interview. Sample size was calculated based on a 20% reduction in symptom severity at the 12-month follow-up, with a 90% power and 2-sided alpha of 0.05, as well as an attrition rate of 15% to 20%. To allow the analysis of equity of treatment effects, the sample size was set at 200 for the depression and AUD cohorts, and at 150 for the psychosis and epilepsy cohorts. Patients were screened with the PHQ-9 and AUDIT by PRIME field workers before their consultation with the medical officer. They were then again followed up after their consultation to assess whether a diagnosis was made. If diagnosed, they were recruited into the respective treatment cohort. In case of patients with multiple diagnoses, priority was given to the more severe disorder. Participants were allocated to the psychosis or epilepsy cohort, in case of comorbidity with depression or AUD. If an individual was diagnosed with both AUD and depression, the participant was recruited into the AUD cohort. Baseline assessments were initiated at the clinic on the day of recruitment, and completed in the participants’ home, on average 1 day after recruitment. The follow-up assessments were conducted in the participants’ homes 3 months (depression and AUD) or 6 months after recruitment (psychosis and epilepsy), and again 12 months after recruitment (all cohorts). Data were collected using Android devices linked to an online application (Mobenzi, https://www.mobenzi.com). Participants were considered lost to follow-up if data for the 12-month assessment could not be collected. The timing of the data collection was as follows: baseline between September 2014 and August 2015, midline between December 2014 and November 2015, and endline between August 2015 and July 2016.
+We mobilised the same field workers for all study components. Two months of extensive training was provided covering qualitative and quantitative research, interviewing skills, rapport building, informed consent, and inclusion/exclusion criteria. Additionally, 2 weeks of trainings were organised for each study component covering recruitment strategy and content
+of the questionnaire, including field practice. Field workers visited each sampled household or health facility, assessed eligibility criteria, performed the procedure for selecting participants, and obtained written consent among the selected participants for interviews. Interviews were conducted in a confidential place, and field workers used tablets for data collection.
+Study measures
+For marital status, participants were grouped into 2 categories based on whether they had ever been married or not. All participants who were either working in an occupation sector or studying were put into a single ‘employed’ category. Participants were classified as being food insecure if anyone in their household had gone hungry in the past month due to lack of resources.
+In the PHQ-9, participants reported the frequency with which they had experienced 9 symptoms over the past 2 weeks on a Likert scale ranging from 0 (‘not at all’) to 3 (‘nearly every day’), and scores from the 9 items were summed [24]. From a validation study in primary care settings in Nepal, a cutoff score of 10 or more had 94% sensitivity and 80% specificity, and internal consistency of a = 0.84 [22]. The 10-item AUDIT was developed by WHO and is used widely in LMICs [25]. With the sum of 10 items, a score of 8 or more is indicative of hazardous, harmful, or dependent drinking behaviours in the pastyear. Internal consistency of the AUDIT in Nepal has been shown to be a = 0.82 [23].
+The 12-item interviewer-administered WHODAS 2.0 has items relating to difficulties engaging in daily activities due to health problems in the past 30 days, and items are scored on a Likert scale from 1 (‘none’) to 5 (‘extreme/cannot do’) [26]. The WHODAS has been validated in a range of settings [26], and has been used in previous research in Nepal [27]. Internal consistency for the WHODAS based on baseline data is a = 0.84 (depression cohort) and a = 0.85 (AUD cohort). Item response theory-based weights were used for the total scoring, to allow comparisons across populations. The assessment of accuracy of diagnosis and minimally adequate treatment for depression and AUD was done by a mhGAP-trained psychiatrist following predetermined decision rules based on the mhGAP intervention guide, stipulating inclusion and exclusion criteria for diagnosis and treatment (see S1 Table for criteria). In case of doubt, another psychiatrist (BAK) was consulted, to come to a consensus decision.
+The 15-item SIP-2R is the short form of the Drinker Inventory of Consequences [28]. Each item is scored on a Likert scale from 0 (‘never’) to 3 (‘daily or almost daily’) with regard to the effects of drinking in the past 3 months. The 14-item PANSS is a symptom-based checklist for severity of psychosis symptoms [29]. The PANSS has not been validated in Nepal but has been culturally adapted for administration to patients and family members, with strong internal consistency for positive items (a = 0.82), negative items (a = 0.88), and combined (a = 0.89) in a rural sample in Nepal [30]. Internal consistency of the PANSS using the cohort baseline data was a = 0.84. The main outcome for patients with epilepsy was number of seizures in the past month. Participants in all studies were also assessed on a range of measures, including demographic and socio-economic, healthcare use, stigma, and discrimination measures (results not reported here).
+Statistical analyses
+We collected data on demographic and health-related characteristics for participants who were recruited into the baseline rounds of the community, facility, and cohort studies. We summarised data using medians and interquartile ranges for continuous measures and counts and proportions for categorical measures.
+For the community study participants with probable depression and with probable AUD, we reported the proportions who contacted any health worker or non-specialist health
+provider at each survey round. We used binomial regression to estimate the change in contact between rounds, and Cohen’s h for the effect size (ES). The regression estimates account for the complex survey design, i.e., strata and probability sampling weights. The analysis for participants with probable AUD was limited to men only, as previous analysis revealed that relatively few women had AUD [19]. These analyses were adjusted to account for the populationbased survey design.
+For calculating the change in contact coverage based on actual service utilisation data, we used the following equation:
+Number clinically diagnosed with a mental illness Contact coverage =  ------------—— ---------—n—... r .....
+Prevalence x Catchment population of health facilities
+The number of cases is based on all cases registered in health facility records over a period of 12 months. Baseline includes the total number of cases 12 months prior to PRIME, endline includes the total number of cases during 12 months when the PRIME cohort studies were implemented. The catchment population is the total adult population of the 10 Village Development Committees from the 2011 census (last available census data) (N = 63,189). Prevalence figures for depression and AUD are based on representative population-level prevalence rates from neighbouring India for (current) depression (2.7%), AUD (4.7%), and psychoses (0.4%) [31], and from a community study in Nepal for epilepsy (0.73%) [32].
+For the facility study, for the participants who screened positive for depression or for AUD, we reported the proportions who returned their clinical consultation forms. Among those who returned their forms, we reported the proportions who had been diagnosed with depression or with AUD, and among those with a diagnosis, the proportions who had evidence of minimally adequate treatment provision. We used binomial regression to estimate the change in diagnosis at each round in comparison to the baseline round, and Cohen’s h for the ES. As it was not possible to use binomial regression to estimate the change in treatment at each follow-up round due to 0 counts in the baseline round, we used Fisher’s exact test to compare the proportions against the baseline round, and used a 1-sample test of proportions to calculate the 95% confidence intervals at each follow-up round.
+For the cohort studies, differences in baseline demographic and clinical characteristics between participants with 12-month data and those lost to follow-up were assessed using nonparametric tests (Fisher’s exact test for categorical variables and Mann-Whitney U test for continuous variables). Because none of the continuous outcomes (WHODAS, PHQ-9, AUDIT, SIP-2R, and PANSS) were normally distributed, univariate negative binomial regression was used to assess change in total score on each outcome from baseline to midline and from baseline to endline in each cohort. Change in number of seizures in the past month in the epilepsy cohort was assessed using Poisson regression. Effect sizes (Cohen’s d) for paired sample analyses were calculated for each outcome. Equity of treatment effect by sex and caste was assessed using negative binomial regression, this time including sex or caste as an interaction term in the model. This was followed by a Wald chi-squared test.
+All community, facility, and cohort data were analysed using Stata (StataCorp, College Station, Texas, US) version 14.
+Ethics
+Ethical approval for the different study components was obtained from the Nepal Health Research Council; the Faculty of Health Sciences, University of Cape Town, South Africa; and the World Health Organization, Geneva, Switzerland.
+Results
+Change in contact coverage
+In the baseline community survey round, 1,983 participants were screened, of whom 60% were male and 46% were between 30 and 50 years of age (see Table 3). Over 1 in 10 (11%) were probable cases of depression. Of the probable cases of depression identified at baseline, 8.5% had contacted a health worker in the past 12 months, in comparison to 11.8% of probable cases at endline; this change of +3.3% (95% CI -5.1%, 11.7%) was not significant. Contact with a non-specialist provider showed a non-significant increase of 2.3% (95% CI -2.8%, 7.5%). For probable cases of AUD among men, non-significant changes were observed for contact with any health worker (+6.3%, 95% CI -3.3%, 15.9%) and for contact with a non-specialist provider (+3.0%, 95% CI -1.6%, 7.6%) (Table 4). Based on actual service utilisation data over 12 months, we observed significant increases in contact coverage for all disorders. As shown in Table 5, the increases ranged from 7.5% for AUD to 50.2% for psychoses.
+Detection of persons with mental illness
+In the baseline round, 1,252 participants were screened, of whom 65% were male and 46% were between 30 and 50 years of age (Table 3). There were 186 participants (15%) who screened positive for depression, of whom 179 returned their outpatient consultation forms. Using outpatient form data, 16/179 (8.9%) were judged to have received a diagnosis of depression. The proportion receiving a diagnosis increased from baseline by 15.7% (95% CI 7.3%, 24.0%), with an ES of 0.432, at the midline round and by 10.2% (95% CI 1.2%, 19.2%; ES 0.301) at the endline round. There were 92 participants (7.4%) who screened positive for AUD. Diagnosis increased from baseline by 58.9% (95% CI 42.0%, 75.7%; ES 1.562) at midline and by 11.0% (95% CI 0.7%, 21.3%; ES 0.500) at endline (Table 6).
+Initiation of adequate treatment
+Among facility survey participants who received a depression diagnosis at baseline, none received adequate treatment. At midline, among those diagnosed, 93.9% (95% CI 77.9%, 98.6%) received adequate treatment, as did 66.7% (95% CI 41.7%, 84.8%) at endline. Among those diagnosed with AUD, 95.1% (95% CI 88.6%, 98.0%) had adequate treatment at midline and 75.0% (95% CI 17.6%, 97.7%) at endline, up from 0% at baseline (see Table 6).
+Clinical and functional treatment outcomes
+A total of 2,139 patients were eligible and consented to take part in the cohort studies. Of these, 137 received a primary diagnosis of depression, 175 were diagnosed with AUD, and 42 were diagnosed with epilepsy—all were recruited into the respective cohort. A total of 95 caregivers of patients diagnosed with psychosis were also recruited into the psychosis cohort. Participants’ demographic characteristics are presented in Table 3. Attrition at the 12-month follow-up was 20.0%, 18.9%, 9.5%, and 9.5% for the depression, AUD, psychosis, and epilepsy cohorts, respectively. Participants lost to follow-up differed from active participants only in the epilepsy cohort: they were all single, and had greater baseline WHODAS and PHQ-9 scores (see S2 Table for reasons for loss to follow-up).
+Results of the negative binomial regressions and Poisson regression are presented in Table 7. Participants in the depression cohort showed significant improvement from baseline to endline, with a significant reduction in WHODAS (P = -15.89; 95% CI -23.03, -8.74; d = -0.41) and PHQ-9 scores (P = -7.22; 95% CI -9.54, -4.89; d = -0.58). In addition, 68.2% (95% CI 58.8%, 76.3%) of participants showed a 50% reduction in PHQ-9 (response) at endline. In
+the AUD cohort, change in score from baseline to endline was significant for the WHODAS (P = -8.57; 95% CI -12.64, -4.49; d = -0.35), AUDIT (p = -9.68; 95% CI -14.35, -5.00; d = -0.34), and SIP-2R (p = -9.13; 95% CI -12.73, -5.54; d = -0.42). Change in WHODAS score from baseline to endline among the psychosis cohort was also significant (p = -13.56; 95% CI -20.78, -6.34; d = -0.40), and so was change in PANSS score (P = -6.42; 95% CI -9.55, -3.28; d = -0.43). Change in WHODAS or symptom severity score was also significant at midline in the depression, AUD, and psychosis cohorts. However, change in WHODAS score or number of seizures in the epilepsy cohort was not significant, neither at midline nor endline. Moreover, among participants who scored above the validated PHQ-9 cutoff for depression at baseline, 86.7% (95% CI 77.8%, 92.3%) of the depression cohort scored below the cutoff at endline. For AUD, 31.9% (95% CI 24.7%, 40.1%) of participants scoring above the validated AUDIT cutoff at baseline scored below cutoff at endline.
+Equity analyses suggest that change in the primary outcomes for the depression cohort (PHQ-9), AUD cohort (AUDIT and SIP-2R), and psychosis cohort (PANSS) did not differ according to the sex or caste of the participants. In the epilepsy cohort, however, the decrease in number of seizures in the past month from baseline to endline was significantly greater among men compared to women (%2 = 10.4, p < 0.001). The decrease in number of seizures from baseline to endline was also greater among the ‘upper’ caste groups (Brahman/Chhetri) (X2 = 47.35, p < 0.001). This was due to 1 outlier (> 100 seizures reported by a participant of the Brahman caste). When outliers were excluded, change in number of seizures over time was no longer different by sex, but was significantly lower from baseline to midline among the ‘lower’ and ethnic minority castes (Janajati, Dalit, or other) (%2 = 61.4, p < 0.001).
+Discussion
+These combined outcomes demonstrate promising results of a district-level MHCP in a low-resource community and primary care setting. We see improvements in actual contact coverage, detection of mental illness by trained health workers, the initiation of minimally adequate treatment, and treatment outcomes. Together these results show the potential of a district MHCP to increase effective coverage for MNS disorders. However, there are also important areas that require further attention, such as preventing attrition in AUD detection rates over time, improving detection rates for depression, maintaining adequacy of treatment over time, and achieving better treatment outcomes for some disorders.
+This research programme is, to our knowledge, unique in that it aims to evaluate each of the steps in the process of integrating mental healthcare in community and primary healthcare platforms in a low-income setting. Through a combination of studies, it provides a population-level perspective on the impact of a district-wide MHCP, covering the extent to which (1) people are seeking care at health facilities, (2) disorders in people attending health facilities are being detected, (3) people being diagnosed are starting adequate treatment, and (4) people are benefiting from treatment.
+People seeking treatment (contact coverage)
+Based on a representative community study, we see modest, non-significant increases in contact coverage as a result of introducing the district-level MHCP; our measure included any treatment contact and contact with a primary health worker. The endline rate of 4%,
+however, does not come close to the targets that were set at the onset of the programme [33]. One possible explanation for this is that the community surveys, although representative, were underpowered to detect changes at the population level. True coverage change may be estimated more efficiently by combining routine clinic data with population prevalence estimates from national surveys [10]. For example, the change in contact coverage using actual service utilisation data is especially promising for psychosis, for which we achieved the target of 50% coverage. In interpreting the contact coverage rates using routine clinic data, it is important to keep in mind that these are based on contact with primary healthcare services. Contact with specialised services is excluded from the calculation and may explain why the baseline rates are low, especially for epilepsy and psychosis, compared to coverage rates in other studies and settings.
+When interpreting the different estimates of contact coverage, it is important to note that routine clinic data provide a more accurate measure of the numbers actually taking up services while the community survey is a more accurate measure of the proportion of the population at large seeking treatment. The former is limited in that it is difficult to ascertain the characteristics of people who need but do not seek care, and the latter is limited by underestimating numbers who actually take up care or by problems of requiring large samples in order to be adequately powered.
+The changed rates in treatment coverage reported in this study are similar to rates seen in high-income settings [3]. One of the elements that may have contributed to increased service utilisation, besides availability of services, is the proactive community case detection strategy that was part of the approach. Utilisation of the CIDT has been demonstrated to be a viable strategy to increase help-seeking for mental healthcare [34,35]. In future programmes the use of the CIDT should be combined with effective stigma-reduction interventions within the communities [36], in order to combine supply-side strategies with strategies that increase demand for mental healthcare.
+People with disorders being detected when attending facilities
+Once accessing health facilities with supervised mhGAP trained health workers, 3 out of 5 people with alcohol problems and 1 out of 4 with depression are detected when the knowledge and skills from training are still relatively fresh (6 months after training). Despite the detection rate remaining relatively stable for depression (1 out 5 patients at 2 years post-training), we see a big drop for AUD (1 out of 8 patients at 2 years), which is possibly due to high dropout rates over time, resulting in health workers losing faith in treating AUD. Although we see significant increase in detection of depression and AUD, many people with depression complaints still go undetected. This is not entirely unsurprising given the difficulties in diagnosing depression in primary care settings, also in high-income settings [37]. The global gaps in primary health workers’ detection of depression require further examination and potentially the development of new training and supervision strategies. One suggestion is to reframe the diagnosis of depression in primary care not as a binary approach, but as a staging approach to the identification and classification of mental disorder [38]. Finally, in our study the small change in depression is influenced by the high baseline detection rate for depression (nearly 9%), which is likely attributable to 1 health worker who had received mental health training in another location.
+With a cutoff score of 10 in primary care settings in Nepal, the PHQ-9 misclassifies approximately 6 participants as false positives for every 4 participants who are true positives [22]. This is comparable to false positive rates with the PHQ-9 in high-income settings. With this in mind, the identified detection rates at baseline, midline, and endline using the PHQ-9 likely underestimate the true detection rate given the high number of PHQ-9 false positives. Working with a PHQ-9 false detection rate of 60% and assuming that the primary health workers only identified true positives, then the upper limit for accurate detection of depression may have been 22% at baseline, 60% at midline, and 50% at endline. The actual detection rate likely falls somewhere below these rates and above the PHQ-9 results reported in our results section. Future studies should consider using structured diagnostic questions for confirmation of detection rates with primary health workers.
+People being diagnosed starting adequate treatment
+Nearly all (95%) people that health workers correctly detected with depression or AUD received minimally adequate treatment 6 months post-training. Twenty-four months after the training, we see a decline to 2/3 for depression and 3/4 for AUD. Importantly, these high rates support the feasibility of relatively short and focused training of health workers, which is at the heart of WHO’s mhGAP programme [39]. These results also re-emphasise the need for supervision to keep up good practice over time [40,41]. At the same time, it is worth noting that coding of treatment adequacy was based on ‘minimally adequate’ care. Unstable supply of psychotropic medicines during the early phase of the programme may explain some of the decline
+in adequate treatment. Further research is needed assessing the rate of optimal care (this study is currently ongoing, and will be published separately).
+People benefiting from treatment
+The combination of interventions provided through the MHCP—which includes psychotropic treatment, home-based care, and psychological treatments—has the expected beneficial effects for people with depression, AUD, and psychosis. At 12 months after treatment initiation, all cohorts see an 8%-16% reduction of functional impairment and 6%-10% symptom reduction (at midline, these values are 6%-15% and 9%-17%, respectively). For people with depression, 87% score below the cutoff of the validated symptoms checklist at 12 months post-treatment; for people with AUD this value is 32%. This study did not aim to evaluate the effectiveness of the provided treatments per se. Rather, it aimed to assess improved functioning and symptom reduction as indicators of feasibility of a community MHCP provided by non-specialists. Our findings support this feasibility, with the exception of the epilepsy cohort, which did not see significant improvement. That said, not having a control group remains an important limitation, especially given trends towards natural remission among people with depression and AUD [42]. The improvements among people with psychosis are similar to those in another recent study of mhGAP in a different rural region of Nepal [30]. Improvements among depressed patients are especially driven by the added value of psychological treatment by the community counsellors, whereas for patients with AUD, pharmacological treatment and psychoeducation by primary health workers appear to primarily explain the improvement [43]. The overall absence of treatment effects for epilepsy is surprising given established effectiveness of treatment as included in the mhGAP guidelines, as well as positive prior outcomes in Nepal [30]. There are a few possible explanations for this finding: (1) a relatively small sample size might have made for an underpowered study and (2) 40% of participants did not report any seizures in the month before baseline. The reasons for the lower change in number of seizures among the ‘lower’ castes and ethnic minority groups need to be studied further.
+The strength of this study is that it presents an evaluation of a real-world district-wide implementation of mental health services within community and primary healthcare platforms in a low-income country. The evaluation follows a theory of change that was developed at the outset of the programme [41], based on guidelines for the evaluation of complex interventions [44], and coordinated with studies in 4 other LMICs [18].
+There are several limitations to be noted. First, the use of screening tools (e.g., PHQ-9 and AUDIT) rather than structured diagnostic assessment risks misclassification of cases in the community and facility studies and an associated reduction of statistical power. Second, community- and facility-level impacts of PRIME for epilepsy and psychosis remain unknown, as these were not included in the study components. Third, while the assessment of adequacy of initiated treatment was done by a mhGAP trained psychiatrist using predefined criteria, we did not systematically assess the reliability of that assessment. Fourth, as noted above, the community surveys might have been underpowered. The sample size suggested by power analysis was reduced for budgetary reasons. We compensated for this by also evaluating changes in contact coverage using service utilisation data. Fifth, the study designs are observational and uncontrolled, which increases the risk for biases.
+This study has several implications for future implementation and research into scaling up mental healthcare in LMICs. First, for any population-level programme, it is essential to demonstrate changes in contact coverage. A programme like PRIME appears to improve contact coverage based on 12-month service utilisation data while failing to demonstrate such change using a representative community survey. Future studies evaluating contact coverage using
+representative samples might need to work with larger samples. Although over half of the people with psychosis appear to have been reached by the programme, there should be more focus on getting people into care, especially for depression, epilepsy, and AUD. Second, and related to the above, investments in making services available should be combined with efforts to increase demand for these services, for example by using proactive case-finding tools such as the CIDT [34]. A combined demand- and supply-side approach will optimise uptake and utilisation of care. Third, a brief mhGAP training to health workers appears adequate for drastically improving their capacity to detect cases of, and initiate minimally adequate treatment for, depression and AUD. At the same time, the attrition of detection rates for AUD over time calls for more focus on supervision and quality monitoring, and the depression detection rates still leave room for improvement, possibly by using different approaches to diagnosis. Fourth, while individuals with depression, AUD, and psychosis receiving mhGAP-based pharmacological and psychological treatment from non-specialist providers report clinical improvements, most of the changes have relatively small ESs, which calls for more focus on the quality of care in future implementation as a means of boosting clinical outcomes. Given the lack control groups, it is not possible to account for natural remission of symptoms as an explanation for these changes. Similarly, the lack of improvements within the overall epilepsy cohort requires further investigation. Taken together, these findings show encouraging improvements in effective coverage at the population level following the implementation of a local MHCP.
+Conclusion
+In efforts to respond to the enormous treatment gap for people with mental illness in LMICs, there is an urgent need for evidence regarding the feasibility of scaling up mental healthcare through community and primary healthcare platforms. PRIME is, to the best of our knowledge, the first programme to systematically evaluate the different assumptions about, and steps towards, making effective mental healthcare available at a population level. A primary indicator of success is effective coverage, defined as the proportion of people who need treatment who accessed services resulting in improvements in patient clinical and functional outcomes. Combining the results from the community, facility, and cohort studies, the programme appears to achieve effective coverage of 1 out of 34 participants with depression and 1 out of 23 participants with AUD—based on community and primary care services alone. Another important indicator is the extent of change that is the result of the implementation of the district MHCP. We demonstrated modest to large and targeted changes in contact coverage (ranging from 1 out of 13 participants with AUD to half of all patients with psychosis). Changes in health workers’ detection ranged from a small ES for change in health worker detection of depression at 24 months (d = 0.30) to a large ES for change in detection of AUD post-training (d = 1.6). We demonstrated that minimally adequate treatment was initiated at the lowest level for two-thirds of the cases with depression at endline, and up to 95% of the cases with AUD right after training. Finally, 3 months after patients initiated care, we observed small to moderate ESs for clinical outcomes (ranging from d = 0.25 for improved functioning among people with psychosis to d = 0.59 for reduction in symptoms for depression and AUD), changes that are maintained 12 months after starting treatment.
+These combined results make a strong case for the impact of a district MHCP in reducing the treatment gap and increasing effective coverage for priority mental disorders, while also pointing towards a set of strategies and new research questions that can contribute towards additional improvements for the future. Ultimately, populations in other low-income and fragile states with limited or non-existent mental health services desperately need models that build on the lessons learned in Nepal through PRIME’s public mental healthcare model.

Diabetes Obesity Metabolism - 2018 - Siskind - Glucagon‐like peptide‐1 receptor agonists for antipsychotic‐associated.txt

@@ -0,0 +1,83 @@
+1	| INTRODUCTION
+The life expectancy for patients with schizophrenia is more than 14-20 years shorter than for the general population,1,2 with 35% of excess deaths attributable to cardiovascular disease and diabetes.3 Patients with schizophrenia are at increased risk of developing cardio-metabolic disease, mediated by or coincident with obesity, for several reasons including a genetic predisposition for developing diabetes, reduced physical activity, poor diet and the use of antipsychotic medications.4,5
+Although the underlying mechanisms have not been fully elucidated, it is well-established that antipsychotic medications can lead to obesity, with clozapine and olanzapine having the greatest propensity for body weight gain.6 Among patients with schizophrenia, about half of those on clozapine and a third of those on olanzapine have metabolic syndrome.7
+Body weight gain is associated with poorer quality of life,8 reduced social engagement,9 and is the most distressing side effect reported to mental health helplines.10 Body weight gain also reinforces patients' negative views of themselves and may compromise adherence with treatment.10 Furthermore, being overweight or obese increases the risk of all-cause mortality with an association between body weight and higher mortality risk.11,12
+The current evidence for interventions addressing antipsychotic-associated obesity is limited. Physical activity interventions are compromised by low rates of uptake and acceptability,13 while many pharmacological treatments can result in unacceptable adverse events.14 For instance, sibutramine was withdrawn because of cardiovascular risks,15 while rimonabant was removed because of increased risk of depression, anxiety and suicide.16 Orlistat is associated with poor adherence because of steatorrhoea.17 Finally, there is only modest (and heterogeneous) body weight loss following the addition of met-formin18,19 or topiramate20 for obese and overweight patients on antipsychotics and/or those at risk for antipsychotic body weight gain.14
+As a result of these limitations, there has been increasing interest in glucagon-like peptide-1 receptor agonists (GLP-1RAs) to counteract the body weight gain associated with antipsychotic treatment in general,21 and clozapine and olanzapine treatment in particular.22,23 Glucagon-like peptide-1 (GLP-1) is an endogenous peptide, synthesized in the intestinal mucosa,24 which stimulates insulin secretion and decreases glucagon secretion in a glucose-dependent manner. It
+also delays gastric emptying and lowers food intake by promoting satiety.25
+GLP-1RAs have well-established glucose- and weight-lowering properties in patients with23 and without26 type 2 diabetes. GLP-1RA treatment is also associated with a lower risk of major adverse cardiovascular endpoints (composite endpoint including cardiovascular-related mortality, non-fatal myocardial infarction, and non-fatal stroke).27 In addition to daily injections, several GLP-1RAs are now available as weekly injections, which may improve adherence among patients with schizophrenia.
+To our knowledge, prior to conducting the comprehensive systematic review, at least three individual trials investigating the effect of GLP-1RAs (exenatide once-weekly or liraglutide once-daily) on antipsychotic-associated obesity28-30 had been published. A metaanalysis of participant-level data has the potential to identify whether the effects of GLP-1RAs vary for different antipsychotics and also to examine the influence of more clinically relevant participant-related factors than is possible in a meta-analysis of study-level data.
+In this study, we tested the hypotheses that
+•	GLP-1RAs would be superior to the control conditions for body weight loss, as well as all other anthropometric and cardio-metabolic outcomes;
+•	patients treated with clozapine or olanzapine would experience greater body weight loss with GLP-1RAs.
+2	| MATERIALS AND METHODS
+2.1	| Protocol and registration
+This study was registered with PROSPERO (CRD42017079791).31 We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations for the background, search strategy, methods, results, discussion and conclusions.32 Ethical approval was not required as all the included data had been previously published.
+2.2	| Search strategy
+The following databases were searched from inception to 24 October 2017: PubMed, PsycInfo, Embase, and the Cochrane Schizophrenia Group's Trials Register. Hand searches of references listed in included
+studies and other key publications were also conducted. Studies were limited to humans. Search terms included terms for antipsychotics and GLP-1RAs. There were no language limitations. PubMed search terms are provided in Table S1 (see the supporting information for this article).
+2.3	| Eligibility criteria and study selection
+We included all randomized controlled trials of patients on antipsychotic medications who were overweight or obese where a GLP-1RA was compared with placebo or usual care. All studies were independently screened at the title and abstract level by two authors (D. S. and M. H.). Studies that met the inclusion criteria on title and abstract review, or that could not be excluded on the basis of information provided in the abstract, were reviewed at full text level.
+2.4	| Data collection process
+Authors of the included studies provided access to de-identified individual participant data. Key authors of the included studies are also co-authors of this meta-analysis. These data were collated and analysed with validation by the corresponding authors of the included studies. Quality assessment was conducted by an author not involved in the included studies (M. H.).
+2.5	| Outcomes
+The primary outcome was the difference in endpoint body weight adjusted for baseline body weight and study between the GLP-1RA arm and control arm. We analysed the following secondary outcomes in the same way: metabolic syndrome components (waist circumference, blood pressure [BP], HDL, LDL, triglycerides [TGs] and fasting plasma glucose [FPG]), body mass index (BMI), HbA1c, homeostatic model assessment (HoMA), insulin, visceral fat and android-to-gynoid ratio (android adipose tissue surrounds the abdomen, chest, shoulder and nape of the neck, while gynoid adipose tissue surrounds the hips, breasts and thighs).
+Psychosis severity for individual patients was based on published inter-scale linkage thresholds for the PANSS, BPRS and GCI.33,34
+2.6	| Study quality
+We assessed study quality using criteria adapted from the Cochrane Collaboration guidelines32: (a) selection bias (random sequence generation and allocation concealment); (b) performance bias; (c) detection bias; (d) attrition bias; (e) reporting bias; and (f) other sources of potential bias including pharmaceutical company funding. Studies were deemed to be of low quality if they had three or more elements with a high risk of bias, while those of high quality had four or more elements with a low risk of bias.
+2.7	| Statistical analyses
+We conducted a one-step meta-analysis on individual participant data, where data from all included studies were modelled simultaneously, while adjusting for clustering of patients within included studies.35
+---------------------------------------WlLEY^95
+The primary and secondary outcomes were analysed as differences in endpoint values between intervention and control, and adjusted for baseline value and study as a random effect using ANCOVA with Bonferroni correction on SPSS version 24 for Mac OS. Where individual patient data were missing, we used the modified intention-to-treat model,29 where the last valid value after the baseline value was carried forward.
+We performed multiple linear regressions with endpoint variable as the dependent variable, including the baseline variable and, respectively, each of the following co-variables: demographics (age, sex), psychosis severity, metabolic variables (body weight, BMI, waist circumference, HbA1c, fasting blood glucose, HDL, LDL, TGs, systolic blood pressure [SBP], diastolic blood pressure [DSP], HoMA, insulin) and treatment variables (treatment arm, nausea, any adverse drug reaction and GLP-1RA agent). If any covariates were significant then they were all included in a multiple linear regression, using backward elimination. Adjusted R2 of the final model was calculated.
+We conducted sensitivity analyses to explore the impact of the specific antipsychotic used (patients on clozapine and/or olanzapine vs. patients on other antipsychotics) on treatment arm and endpoint metabolic variables, adjusted for the baseline variable, and conducted a meta-analysis for each metabolic variable using RevMan 5.3. We also carried out a sensitivity analysis by excluding patients with type 2 diabetes.
+Chi-square tests were conducted on the proportion of patients in the GLP-1RA and control groups who achieved >5% and >7% body weight loss. Number-needed-to-treat (NNT) was calculated for the proportion of patients with >5% or >7% body weight loss by dividing one by the risk difference.
+BMI was categorized as per WHO categories (overweight: 25-29.9, obese class I: 30-34.9, obese class II: 35-39.9, obese class III: 40 and above),36 and the proportion of GLP-1RA-treated patients and controls who shifted between categories from baseline to endpoint was analysed using a chi-square test.
+FPG was categorized as per ADA categories (normoglycaemic <5.6 mmoL/L, impaired FPG 5.6-6.9 mmoL/L, type 2 diabetes >6.9 mmoL/L). Chi-square tests between baseline and endpoint FPG categories were conducted for total participants, and those in the GLP-1RA and control arms.
+Adverse drug reactions (ADRs) were compared among treatment and control groups using chi-square tests with data available on nausea, diarrhoea, vomiting, other ADRs and any ADR, and a number-needed-to-harm was calculated for ADRs that were significantly different among GLP-1RAs and controls by dividing one by the risk difference. A regression analysis for body weight, adjusted for baseline body weight, study and nausea, was conducted to assess any potential impact of nausea as a mediating factor in body weight change.
+2.8	| Publication bias
+If the meta-analyses included 10 or more studies, we planned to test for publication bias using funnel plot asymmetry where low P values suggest publication bias.37
+296—*-Wl LEY-----------------------------------------------
+3	| RESULTS
+3.1	| Study selection
+Our search identified 56 unique articles. Of these, 43 were excluded at title and abstract level, leaving 13 articles for review at full text level. Three articles met the inclusion criteria28-30 with a total of 168 patients (GLP-1RA = 84, control = 84). Reasons for exclusion at full text level are provided in Supporting Information Figure S1 and Table S2 (see the supporting information for this article).
+3.2	| Study characteristics
+Studies were conducted in Denmark (n = 2) and in Australia (n = 1) (Table 1). Duration ranged from 12 to 24 weeks (mean 16.2 weeks, SD 4.0). Two studies used exenatide 2 mg subcutaneously (s.c.) once-weekly,28,30 and one study used liraglutide 1.8 mg s.c. once-daily,29 the standard maximum doses used for diabetes.38 All studies examined GLP-1RA for people on antipsychotic medications, with no notable changes in antipsychotic doses among participants. One study was restricted to participants receiving clozapine and olanzapine,29 and another to clozapine alone.30 The third study included a naturalistic patient sample treated with clozapine, olanzapine, aripiprazole, risperidone, paliperidone, quetiapine, ziprasidone, amisulpride and sertin-dole.28 After initial publication, an erratum on corrected metabolic blood markers was published for the third study, and these data were used in the current meta-analysis.39 Two studies were blinded and placebo-controlled,28,29 while the third was open label.30 All studies were of adults aged 18-65 years (mean 40.0 years, SD 11.1), 58.3% were male, and the mean BMI of participants was 35.4 kg/m2 (SD 5.7). All studies included patients with schizophrenia, while two also included schizoaffective disorder.28,30 One study also included patients with type 2 diabetes,30 while the other two specifically excluded type 2 diabetes.28,29 One study required patients to have impaired glucose tolerance.29 All studies provided data on body weight, BMI, FPG, HDL, TGs, SBP, DSP and HbA1c. Two studies provided data on insulin and HoMA,29,30 and two on android/gynoid ratio and visceral adiposity.28,29 Two studies (n = 97 and n = 28) showed significant effect on their primary outcome,29,30 while the third one (n = 43) was equivocal.28 Baseline characteristics of the combined dataset are provided in Table 2. All studies were rated to be of high quality (Supporting Information Table S3).
+3.3	| Primary outcome
+The mean adjusted difference in endpoint body weight among intervention and control groups was 3.71 kg lower for the intervention groups (2.44-4.99 kg, 95% CI) (Table 3). This was a statistically significant difference for treatment arm (p < 0.001), but not for study (p = 0.430).
+3.4	| Secondary outcomes
+Reductions in waist circumference, BMI, HbA1c, FPG, LDL and visceral fat were all significantly different between treatment and control
+(p values <0.001 to 0.03). Lower LDL and DSP were associated with study effect (Table 3).
+3.5	| Linear regression
+Treatment arm and the baseline variable were statistically significant in the multiple linear regressions of endpoint body weight, BMI and HbA1c (Supplementary Appendix A). Treatment arm, the baseline variable and the additional metabolic variable(s) provided in parentheses were statistically significant for endpoint waist circumference (baseline weight), endpoint FPG (baseline HbA1c), endpoint LDL (baseline TGs), endpoint TGs (baseline waist circumference) and endpoint visceral fat (baseline insulin).
+The variables (in parentheses) were significantly associated with the outcome; however, treatment type was not associated with changes in the following variables: endpoint HDL (baseline HDL), endpoint SBP (baseline SBP and DBP), DBP (baseline DBP and android/ gynoid ratio), HoMA (baseline insulin and visceral fat), insulin (baseline insulin and visceral fat) and android/gynoid ratio (baseline android/ gynoid ratio and TGs).
+Age, sex, psychosis severity, baseline SBP, nausea, any ADR and GLP-1RA agent were not significant in any of the linear regressions of endpoint variables. Adjusted R2 for the multiple linear regressions ranged from 0.284 for SBP to 0.958 for body weight (Supplementary Appendix A).
+3.6	| Sensitivity analyses
+For the sensitivity analysis by antipsychotic, patients on clozapine and/or olanzapine (n = 141) had a statistically significant reduction in body weight with GLP-1RAs (mean 4.70 kg, 3.13-6.27 kg, 95% CI), while those on other antipsychotics (n = 27) did not have statistically significant change in body weight (mean 1.5 kg, 1.47-4.47 kg, 95% CI). The difference between these two groups was statistically significant (p < 0.001). This pattern of statistically significant change in metabolic variables among patients on clozapine and/or olanzapine, but not other antipsychotics, was also seen for waist circumference, BMI, FPG and visceral fat. Both patients on clozapine and/or olanzapine, and those on other antipsychotics, had a statistically significant reduction in HbA1c (Figure 1, Supporting Information in Table S4). When patients only on clozapine were examined (n = 113), they had a 4.90 kg greater body weight loss (3.16-6.64 kg, 95% CI) with GLP-1RAs compared with controls. When patients only on olanzapine were examined (n = 25), they had a 4.70 kg greater body weight loss (1.15-8.25 kg, 95% CI) with GLP-1RAs compared with controls. The difference in comparative body weight loss between clozapine and olanzapine was not statistically significant (p = 0.845).
+When patients with type 2 diabetes were excluded, reduction in body weight with GLP-1RAs remained statistically significant (3.85 kg, 2.54-5.15 kg, 95% CI).
+3.7	| Percentage change in body weight
+A significantly greater proportion of patients on GLP-1RA treatment than controls had a body weight loss of >5% (36.9% vs. 10.7%,
+3.8	| Shift in BMI category
+Among patients on GLP-1RAs, 15 (17.9%) shifted down a BMI category, 69 (82.1%) remained in the same category, and none increased a category, while among controls seven (8.4%) shifted down a category, 72 (86.7%) remained in the same category, and four (2.4%) increased a category (x2 6.967, 2° of freedom, p = 0.031) (Supporting Information Table S5).
+3.9	| Shift in FPG category
+Among those with impaired FPG, 26 of 38 (68.4%) participants on GLP-1RAs had normal FPG at endpoint, while only 9 of 38 (23.7%) participants in the control arms had normal FPG at endpoint. Changes in FPG categories are provided in Supporting Information Table S6.
+3.10	| Adverse events
+Patients on GLP-1RAs reported significantly more nausea compared with controls (53.6% vs. 27.5%, p = 0.002), with a number needed to harm of 3.8 (2.4 to 9.7, 95% CI). Neither the presence of any ADR (76.8% vs. 62.9%, p = 0.073), diarrhoea, vomiting, nor other ADRs were significantly different between the two groups (Supporting Information Table S7). Nausea did not significantly impact the regression model for weight.
+3.11	| Publication bias
+We were unable to assess publication bias, as no analyses included 10 or more studies.
+4	| DISCUSSION
+This systematic review and patient-level meta-analysis suggests that GLP-1RAs can induce a clinically meaningful body weight loss in patients with schizophrenia on antipsychotic medications who are overweight or obese. Patients in the intervention arm lost 3.7 kg more body weight than controls. The NNT to achieve a body weight loss of at least 5% (considered clinically meaningful) was 3.8, while for loss of at least 7%, the NNT was 7.7. GLP-1RA treatment was also associated with greater reductions in BMI, FPG, HbA1c and visceral fat. Body weight loss was greatest for those on clozapine and/or olanzapine compared with other antipsychotics. Age, sex, psychosis severity, nausea, any ADR and GLP-1RA agent did not affect body weight or other metabolic variables.
+In terms of adverse events, nausea was more common in the GLP-1RA group, but was not associated with greater body weight loss, and thus unlikely to explain the findings. Antipsychotics, including clozapine and olanzapine, have antiemetic properties,40,41 which may have mitigated this adverse event. In addition, GLP-1RAs are not hepatically metabolized by cytochrome P450, and as such unlikely to interfere with elimination of antipsychotics.42
+Our findings are consistent with data in non-psychiatrically ill patients with23 and without type 2 diabetes.26 There is increasing acknowledgment of the role of GLP-1RAs as an efficacious pharmacological management strategy for the management of obesity, with suggestions that they are under-utilized in the general population.43
+Our finding of a reduction in visceral adiposity is also important, as this is an independent risk factor for cardiovascular disease,44 dia-betes45 and death.46 This is particularly relevant in schizophrenia, where antipsychotic use is both associated with increases in visceral fat and subsequent metabolic syndrome.47
+The improvements in FPG and HbA1c associated with GLP-1RA treatment are also of clinical importance, given the high rates of glucose intolerance (55%) and impaired FPG (21%) in patients on clozapine or other second-generation antipsychotics.48,49 Over one third of patients on clozapine develop type 2 diabetes.50 In turn, there is a higher mortality among those with serious mental illness and type 2 diabetes than those diagnosed with either type 2 diabetes alone, or serious mental illness alone.51,52 This finding highlights the advantages of body weight loss medications that are also approved antihypergly-caemic drugs with proven reductions in cardiovascular mortality in patients with high-risk type 2 diabetes.27
+Differences in endpoint body weight between GLP-1RA treatment and control interventions were greater for patients on clozapine and/or olanzapine, which is consistent with preclinical findings on olanzapine's and clozapine's disruption of the GLP-1RA pathway.22 This result is important, as clozapine remains the only antipsychotic indicated for patients with treatment-resistant schizophrenia, and has the best evidence for managing positive symptoms53 and reducing hospitalizations54 in this population. Body weight gain can be both a barrier to commencement of clozapine and a reason for its discontinuation. Our finding of a body weight loss of almost 5 kg more than in the control group among patients on clozapine was significantly greater than that reported for metformin in a recent meta-analysis of people on clozapine55 (-3.1, -4.9 to -1.4 kg, 95% CI) (p = 0.024). The potential superiority of GLP-1RAs over metformin is also supported by preclinical models. For instance, GLP-1RAs normalize
+glucose tolerance and decrease body weight in rats treated with clozapine, providing mechanistic justification of their therapeutic potential in this context.22 By contrast, metformin only partially attenuates glucose dysregulation in animal models of antipsychotic metabolic abnormalities.56 To date, there have been no head-to-head studies of the
+effects of GLP-1RAs versus metformin on body weight loss among overweight patients on antipsychotics.
+We were not able to include data on non-alcoholic fatty liver disease (NAFLD). NAFLD is a precursor for the development of incident type 2 diabetes and metabolic syndrome57 and has a mutual and bi-
+30^Wl LEY---------------------------------------------------
+directional relationship with these disease entities.58 There is recent evidence for the efficacy of liraglutide for non-alcoholic steatohepati-tis.59 Future studies of GLP-1RAs should include measures of NAFLD.
+A key strength of this study is the use of individual participant data. This approach allowed for consistent analytic techniques across studies, notably endpoint values adjusted for baseline values, as recommended by the European Medicines Agency.60 It also allowed sensitivity analyses by specific variables, notably antipsychotic, age, sex and study duration, and for correlations between change in BMI and baseline BMI.
+There are also limitations to this study. There were differences in study inclusion criteria. Although all studies used overweight or obesity as inclusion criteria, one study specifically recruited patients with prediabetes, while another also included patients with type 2 diabetes. Only one study included patients who were not on clozapine or olanzapine, and these patients may have differed in baseline characteristics. This limits both the power and the certainty of differences in the effect of GLP-1RAs on patients who were not on clozapine or olanzapine. One study was not blinded, increasing the risk of bias; however, sensitivity analysis by removal of this study did not markedly change the outcomes. It is unclear why DBP and HDL were significantly different because of study effect, but this result may have been related in part to the prediabetes entry criteria of the one study of liraglutide.29 None of the included studies could report whether the body weight gain was specifically attributable to antipsychotic use. The included studies did not use easily comparable psychotic symptomrating scales, making psychotic symptoms impractical to assess as an outcome. Study durations were too short to evaluate comparative risks of major adverse cardiovascular endpoints. We were only able to include 168 patients from three studies, which limits our ability to draw firm conclusions or infer clinical recommendations. We do not have data for older participants, and as such these results cannot be generalized to older adults on antipsychotic medications. Further studies are required in this population.
+In conclusion, our findings suggest a promising role for GLP-1RA treatment for body weight management in patients with schizophrenia treated with clozapine or olanzapine; however, there are insufficient data to comment on the role of GLP-1RAs for those on other antipsychotics. GLP-1RA agents are also well-tolerated, with nausea being the most common ADR. The availability of a once-weekly injectable formulation may also offer advantages when compared with traditional body weight loss or diabetic medications requiring daily administration. However, obviously the availability of an oral formulation would increase the ease of use. While several body weight loss agents have been withdrawn because of adverse cardiac effects, GLP-1RAs are associated with lowering of cardiovascular mortality.27 Our findings also suggest ancillary improvements in glucose homeostasis and visceral adiposity. While our data suggest that individuals taking clozapine or olanzapine may benefit most from GLP-1RAs with a less compelling argument for the use of GLP-1RAs for patients on other antipsychotics, this conclusion should be tempered by the fact that only one study included patients on antipsychotics other than clozapine and olanzapine. Further randomized clinical trials of GLP-1RAs in overweight or obese antipsychotic-treated patients with
+schizophrenia are required, particularly head-to-head trials comparing metformin and GLP-1RAs.

Donor-financing-of-global-mental-health-19952015-An-assessment-of-trends-channels-and-alignment-with-the-disease-burdenPLoS-ONE.txt

@@ -0,0 +1,36 @@
+Introduction
+Recently, the Institute for Health Metrics and Evaluation (IHME) at the University of Washington, Seattle (http://www.healthdata.org/) released the seventh edition of its Financing Global Health report[1]. A core objective of the report is to capture trends in development assistance for health (DAH) and government health expenditure with the aim to provide much-needed information to global health stakeholders about the levels and trends of global health financing.
+The Financing Global Health report splits funding across ten health focus areas, one of which is non-communicable diseases (NCD). Within the NCD health focus area, IHME further disaggregates donor funding into several more exact program areas, one of which is mental health. The Financing Global Health 2015 Report highlights that mental health receives little attention even though it is a major cause of disease burden-accounting for 6.5% of disability adjusted life years (DALYs) in low- and middle-income countries (LMICs) [2]. The lack of alignment between disease burden and funding has been discussed previously by this group [3].
+A valuable review of development assistance for mental health (DAMH) has been previously carried out by Gilbert and colleagues [4] who found that DAMH was less than 1% of total DAH. In addition to including a more extensive array of data sources, this paper will extend the work done by Gilbert and colleagues in two significant ways; first by contrasting DAMH against DAH for other disease categories (including HIV, TB, malaria and maternal and child health), and second by benchmarking allocated DAH against the core disease burden metric (disability-adjusted life year) as estimated by the Global Burden of Disease Studies (http://www.healthdata.org/gbd). This paper will explore DAH, and specifically DAMH, by health focus, geographical and income regions and over time. It will highlight the sources, channels and recipients of DAMH, and importantly, will report a standardised measure to clearly identify health financing gaps, development assistance (expressed in US dollars) per disability-adjusted life year-DAH per DALY.
+Methods
+DAH is the financial and in-kind contributions transferred from global health channels to low- and middle-income countries with the primary intent of maintaining or improving health. In order to track DAH, IHME collates information from audited financial records, project level data, and budget information from the primary global health channels. Tracked
+channels include bilateral aid agencies, such as the United States Agency for International Development and United Kingdom's Department for International Development; multilateral aid agencies, such as the World Bank and regional development banks; United Nations agencies, such as the World Health Organization and UNICEF; public-private partnerships, such as Gavi, the Vaccine Alliance and the Global Fund to Fight AIDS, Tuberculosis, and Malaria; and non-governmental organisations and private foundations. Resources disbursed through these organisations are tracked backward to assess the source of the funds and tracked forward to the recipient country. The diverse set of data are standardised and put into a single inflation adjusted currency (2015 US dollars), adjusted to reflect disbursements rather than simply commitments, and adjusted to remove double counting that occurs when organisations transfer resources between each other. Most important for this research, IHME also estimates the health focus areas and program areas targeted by each project [1, 5]. We extracted annual DAH estimates from 1990 through 2015.
+We tied these health financing estimates to health burden estimates for LMICs produced by the Global Burden of Disease 2015 Study (GBD 2015). GBD 2015 is a systematic and comprehensive framework that uses all available data to quantify mortality and morbidity in 188 countries from 1990 to the present. Mortality and morbidity are disaggregated into 301 medical conditions and causes of illness, including 19 mental and substance use disorders. Health loss due to mortality and morbidity are aggregated to quantify total health burden, measured using disability-adjust life years (DALYs). One disability-adjusted life year is one year of life lost due to premature mortality or several years of life lived with disability. DALY estimates, stratified by age and sex, are made for 1990 to 2015. Over 1,600 researchers from over 120 countries are involved in collecting data and analysing estimates for the GBD, which is coordinated by IHME [6±8].
+Results
+Health focus
+Total DAH in 2015 was estimated to be USD 36 billion. Of this, USD 110 million was estimated to be allocated to mental health. Fig 1 contrasts major health focus categories receiving development assistance over time. DAMH experienced a steady increase from USD 18 million in 1995 to USD 132 million in 2015, a 6-fold increase. Whilst this increase may appear substantial, it equates to only 0.4% of total DAH in 2015. NCDs, of which mental health is a subcategory, are allocated around 1% of total DAH. HIV receives the largest proportion of DAH and saw its allocation increase from USD 612 million to USD 11 billion over the 1995 to 2015 time period. HIV and maternal and child health each consume around 30% of the total DAH.
+Malaria experienced the greatest proportional gain in development assistance over this period increasing from USD 58 million to USD 2.3 billion.
+Sources and channels
+Over the 15 year period, 1990 to 2015, the United States government provided approximately USD 270 million of total DAMH; however, it was private philanthropy that was the most significant source (USD 435 million), accounting for one third of DAMH (Fig 2). NGOs and foundations channelled the overwhelming majority of DAMH (USD 780 million or approximately two thirds of total DAMH) over the 2000±2015 period. Most of the remaining DAMH was contributed by governments of high-income countries through bilateral aid agencies.
+When one examines the channels by which DAMH flows in detail, the World Health Organization distributed the second largest amount of DAMH (USD15 million) behind NGOs (USD54 million) in 2015 (see S2 Fig). Of the development banks, the African Development
+Africa Middle East (15%). East Asia and the Pacific received the smallest fraction of DAMH at 5%. In the same year, the distribution of DAMH across country income groupings was low (USD20 million), lower-middle (USD17 million) and upper-middle income (USD11 million) (S4 Fig). Proportionally the population distribution across these regions is low (42%), lower-middle (36%), and upper-middle (22%) (http://data.worldbank.org/news/new-country-classifications-2015).
+DAMH per DALY
+When DAMH is benchmarked against disease burden attributable to mental and substance use disorders from GBD 2013[2] by World Bank regions, a picture of inequitable distribution emerges (Fig 3). DAMH available per DALY of disease burden in 2013 ranged from USD 0.27 in East Asia and the Pacific to USD 1.18 in the Middle East and North Africa. Sub-Saharan Africa received USD 1.14 per DALY from mental and substance use disorder.
+Benchmarking development assistance against disease burden in LMICs allows for useful comparisons across disease categories. Fig 4 demonstrates there has been an increase in DAMH from USD 0.19 per DALY in 1995 to USD 0.85 per DALY in 2013, a 4-fold increase. Fig 5 demonstrates that HIV/AIDS has the largest ratio of funds to burden (USD 144 per DALY), around three times the amount of the second largest disease group recipient in 2013. Maternal and neonatal health, TB and malaria received between 32 and 48 USD of DAH per DALY in LMICs. Mental and substance use disorders and it
+Discussion
+Assessment of development assistance for health over a period of two decades reveals several observations. Most notably, it highlights significant increases in DAH across all major health groups. This has occurred over a period where health priorities have been tightly connected to
+Fig 4. DAMH per DALY across time, 1995-2013.
+doi:10.1371/journal.pone.0169384.g004
+the targets of the Millennium Development Goals. Consequently, the areas of HIV, TB and malaria in particular have seen substantial investment in terms of DAH. The closing of the MDG era has coincided with a revived investment and commitment to deriving global health estimates. The Global Burden of Disease Studies quantify health loss from hundreds of diseases, injuries, and risk factors, with the aim that information from these studies can be used to improve health systems and eliminate health disparities. It aims to align health systems with the needs of populations by assisting policymakers to identify the major health challenges facing their country. The joint use of estimates of disease burden and development assistance for health provides opportunity for a realignment of resource allocation.
+Apparent inequities extend well beyond the total proportion of development assistance allocated to mental health. Private philanthropy accounts for only a fraction of total DAH yet it is overwhelmingly the largest donor of DAMH—suggesting a lack of interest by governments to address mental health needs across the globe. In terms of recipients, there appears to be no clear association between need, in terms of absolute DALYs, and where DAMH is going.
+There are other important indictors which highlight the lack of resources in mental health. According to the World Health Organization Mental Health Atlas[9], there is only one psychiatrist per 200,000 people or more for about half of the world’s population. Around 80% of mental healthcare workers are based in inpatient and day care services. The capacity to build the workforce appears minimal with the same report estimating around 2% of physicians and nurses received at least 2 days of mental health training in the last 2 years. Furthermore, funding for research into mental illness is not on par with research funding allocated to physical conditions [10,11].
+South Asia received the largest portion of DAMH in 2013 in terms of absolute dollars. Of all World Bank regions, it has also experienced the largest percentage increase in DAMH since 1995. Whilst the direct drivers of these trends have not been documented it is interesting to note that South Asia, or more specifically India, has been the focus of a large and effective global mental health movement in recent years with significant progress being made in terms
+of research and policy. The allocation and distribution of developmental assistance is influenced by a complex interplay of geo-political factors, including political and strategic considerations [12]. The priorities of non-state development partners are regularly at odds with those of the national priorities [13]. Nonetheless, an understanding of the way both funding agencies and recipient governments or organisations view mental health will go partway in explaining the reasons behind the misalignment between disease burden and DAMH. Even in the face of economic arguments, as well as burden of disease evidence, governments in low and middle income countries have been slow to respond to the rising burden of mental and substance use disorders. The well-known study undertaken for the World Economic Forum estimated that the cumulative global impact of mental disorders in terms of lost economic output may amount to US$16 trillion over 20 years, equivalent to 25% of global GDP in 2010[14].
+However, the size of the burden for any group of disorders is insufficient, in its own right, to determine the magnitude of proportional investment within the health sector. Burden evidence needs to be combined with information on the cost-effectiveness of interventions to reduce the burden, especially in low and middle income countries. This information does exist for mental and substance use disorders. Work done for the Disease Control Priorities in Developing Countries third edition [15] and the WHO found that a scaled-up package of mental health interventions for key mental disorders in Sub-Saharan Africa and South Asia, would cost in the order of US$3±4 per head of population[15]. In addition to the availability of cost
+effective interventions, the return on investment in mental health is accumulating. The recent study by Chisholm and colleagues demonstrated that substantially scaling up effective treatment coverage for depression and anxiety disorders over the period 2016 to 2030 would conservatively lead to 43 million extra years of healthy life over the scale-up period. The economic value on these healthy life-years was estimated at USD 310 billion at net present value, with a benefit to cost ratio of 2-3±3-0 to 1 when economic benefits only were considered, and 3-3±5-7 to 1 when the value of health returns was also included[16].
+Even where the burden is high and cost-effective treatments exist, other factors influence governments and funders. It is beyond the scope of this paper to discuss these in detail but they include the importance of mental health as a public good, the societal impact of untreated mental illness (externalities), the need for regulation (including of service providers), protection from catastrophic costs and whether the private sector can provide mental health services. Using criteria such as these, an analysis for the World Bank found a strong case for government and public sector involvement in mental health treatment[17].
+Mental health has not, in most countries, become a priority commensurate with the extent of its burden and the potential to reduce the burden. Commenting specifically on the lack of action following the report for the World Economic Forum[14], Insel and colleagues argue that, mental illness is still perceived as an individual or family problem rather than a policy challenge with significant economic and political implications, and, in many low- and middleincome countries, treatment for mental illness is seen as an unaffordable luxury[18]. Tackling perceptions such as these will require a more sophisticated, multifaceted presentation of evidence to governments, funders and society than has been achieved to date. It is hoped that actions such as the inclusion of mental health in the Sustainable Development Goals[19] and commitments from major stakeholders in global health, such as those given at the April 2016 meeting, co-hosted by the World Bank and WHO, to make mental health a global health and development priority [20] will coalesce with mental health campaigns and movements within and across societies, to create the tipping point for mental health to at last become a global health priority.
+This paper demonstrates how it possible to track DAMH from global health channels to low- and middle-income countries. The primary limitation of this research relates to the underlying data used to generate estimates of DAMH. Budget, spending, and revenue data were collected for each major channel of development assistance. These data were disparate and vary greatly regarding the amount of project level data available and the information reported. In some cases, statistical models were used to estimate disbursement when only commitment data was available or to estimate disbursements for the most recent years, when reporting lags prevented project level reporting. In addition to this, and critical for this research, the disaggregation of development assistance for health across health focus areas, and identification of DAMH, is based primarily on keyword searches of project titles and project descriptions (S1 Table). These methods are not perfect as keywords searches relay on the comprehensiveness of the underlying project descriptions. While this means that these DAMH estimates should be considered approximations rather than precise estimates, these methods have been evaluated and vetted elsewhere [3,21,22], and the magnitudes and trends reported here conform to previous estimates. In addition to this, projects directed towards other sectors and health focus areas (e.g. poverty reduction, maternal and child health, and health system strengthening) which may indirectly finance the prevention or treatment of mental and substance use disorders are not included in this study, as there are not a comprehensive set of how much of government spending that is spent on mental health.
+Benchmarking development assistance for health per DALY provides only a single perspective on funding allocations. Allocating finite resources across sectors and health focus areas is
+complicated and requires a great deal of consideration beyond simply the underlying disease burden as discussed earlier. While decisions related to resource allocation should consider the cost-effectiveness of interventions, existing resources available, and a host of contextual and cultural issues [23], these factors do not preclude the DAH per DALY metric from being a valuable description of current resource allocations.
+The lack of alignment between disease burden and funding across disease categories raises the issue of whether the DAH, especially DAMH, is equitable and whether there is potential for large improvements in resource allocation. DAH, when assessed in a broader context, holds the potential to be a powerful indicator for progress in global health and global mental health.

Early nutrition influences developmental.txt

@@ -0,0 +1,67 @@
+Introduction
+Infancy and early childhood are sensitive and rapid periods of brain growth that coincide with the emergence of nearly all cognitive, behavioral, and social-emotional functions (Johnson, 2001). Throughout this period, the brain's eloquent networks are shaped and refined through processes that include myelination, dendritic arborisation and synaptogenesis, and synaptic pruning. These adaptive processes are modulated by neural activity and are responsive to environmental, genetic, hormonal, and other influences (Stiles and Jernigan, 2010). The development and pattern of myelination follows a well-described neuroanatomical arc (Brody et al., 1987), progressing in a posterior-to-anterior and centre-outwards spatiotemporal pattern that
+corresponding to maturing cognitive functions (McGee et al., 2005; Markham and Greenough, 2004). That is, there is a strong overlap in the emergence of a specific cognitive function and the myelination of brain regions and networks subserving that function (Fornari et al., 2007; Van der Knaap et al., 1991; Pujol et al., 2006). Beyond this temporal association, prior studies have further shown the importance of white matter and cortical myelination to cognitive development and brain plasticity (McGee et al., 2005; Pujol et al., 2004, 2006; Fields, 2008; Fornari et al., 2007), and altered myelination and white matter maturation in a variety of intellectual, behavioral, and psychiatric disorders (Bartzokis et al., 2003; Flynn et al., 2003; Davison and Dobbing, 1966; Wolff et al., 2012). We have further shown that early trajectories of myelination are associated with cognitive abilities and outcomes (O'Muircheartaigh et al.,
+2014; Deoni et al., 2014).
+The assembly and maintenance of the myelin sheath requires a carefully orchestrated delivery of nutrients, including lipids and fatty acids, proteins, minerals, and other micronutrients (Dobbing, 1964). Long-chain polyunsaturated fatty acids (LC-PUFAs), choline, iron, zinc, cholesterol, phospholipids, and sphingomyelin play essential roles in myelin elaboration, as key components of the myelin sheath and/or energy sources (Oshida et al., 2003; Saher et al., 2005; Hadley et al., 2016; Chang et al., 2009). Deficiencies in these nutrients throughout infancy can significantly alter myelin content, composition, and morphology, potentially disrupting normal brain function and impairing cognitive outcomes.
+Compositionally, human breastmilk provides many of the nutritional building blocks that support healthy physical growth, immune system development, and brain maturation (Kramer et al., 2008; Jacobi and Odle, 2012; Hoi and McKerracher, 2015; M'Rabet et al., 2008; Reynolds, 2001). This includes micro and macro-nutrients, short and long-chain PUFAs, phospholipids, neurotrophic factors, biofactors, and hormones that are important for myelination. While many of these nutrients are also provided by infant formula, their concentration often varies considerably from human milk, and does not mimic the changing nutritional composition of human milk across an individual feed (from foremilk to hindmilk), or from colostrum to mature milk (Ballard and Morrow, 2013). It is possible, therefore that given the importance of these nutritional components to brain development, nutritional differences between breast milk and infant formula, or between formula, may influence trajectories of brain myelination and, subsequently, affect cognitive development.
+With specific reference to brain myelination, breastmilk is an important source of long-chain PUFAs, including docosahexaenoic and arachidonic acid (DHA and ARA), the together comprise more than 20% of the brain's fatty acid content (Chang et al., 2009), and phospholipids such as phosphatidylcholine that make up 10% of the lipid weight of myelin. Approximately 40% of the lipid content of mature human milk is sphingomyelin (Blaas et al., 2011), a sphingolipid that plays a critical role in development of the myelin sheath (Oshida et al., 2003; Jana and Pahan, 2010). Breastmilk is also an important source of cholesterol, which is essential for myelin synthesis (Saher et al., 2005). Even in otherwise healthy children, prolonged deficits in these and other nutrients have been associated with developmental abnormalities and cognitive impairments. For example, prolonged essential fatty acid deficiency or low blood levels of ARA and DHA have been associated with learning disorders, ADHD, dyslexia, and autism spectrum disorder (Hadley et al., 2016).
+The nutritional composition differences between breast and infant formula milk may help to explain some of the observed difference in overall cognitive functioning and ability between breast and formula fed infants (Horwood and Fergusson, 1998). Even controlling for important confounds such as birth weight, pregnancy length, parent education level, and family socioeconomic status and demographics, the general consensus from prior studies is that children and adolescents breastfed as infants show improved performance on tests of cognitive functioning (Horwood and Fergusson, 1998; Anderson et al., 1999; Kramer et al., 2008; Mortensen et al., 2002; Huang et al., 2014). These results are also generally supported by brain imaging studies, which have shown increased white matter volume, total gray matter volume, and regional cortical thickness increases in association with breastfeeding duration and percentage of breastmilk in a infant's diet. These neuroimaging finds have further been associated with improved cognitive function as measured by IQ (Ou et al., 2015; Isaacs et al., 2010; Kafouri et al., 2013; Luby et al., 2016). Although these studies have been performed predominately in older children and adolescents, our group's prior work (Deoni et al., 2013a,b) extended these findings to infants, showing cross-sectional differences in early brain myelination between exclusively breastfed, exclusively formula-fed, and mixed-fed infants and toddlers. These differences were found to present prior to one year of age
+and extend throughout childhood, and were associated with duration of breastfeeding.
+An important limitation of past neuroimaging (MRI) studies, however, has been their cross-sectional nature with children pooled across large age-ranges, making it difficult to draw causative conclusions. In addition, formula-fed children are often treated as a single group without consideration of the potential differences in formula composition. These limitations generally stem from the retrospective nature of most studies, with nutritional composition information often not remembered or readily available. Though infant formula is tightly regulated (e.g., http:// www.fda.gov/ForConsumers/ConsumerUpdates/ucm048694.htm), there exists measurable differences in micronutrient, PUFA, and phospholipid content across different infant formulas.
+To investigate nutritional influences on longitudinal infant and child brain development in a naturalistic setting, we longitudinally characterised myelination in a large group (n = 150, 57 females) of healthy and neurotypically developing children from 3 months to 9 years of age. A total of 452 total MRI and neurocognitive datasets were acquired on these children. These children were drawn from a larger study of normal brain development and were selected since knowledge of infant feeding habits, duration of exclusive breastfeeding, and main infant formula composition was known. Brain myelination was quantified using a multicomponent relaxometry (MCR) technique termed mcDESPOT (Deoni et al., 2008), which decomposes the measured MRI signal into contributions from distinct sub-voxel anatomical water pools (MacKay et al., 1994). Through the acquisition of multiple T1 weighted and T1/T2 weighted spoiled and fully-balanced steady-state images with different flip angles, mcDESPOT applies a 3-pool tissue model (Deoni et al., 2013a, b) to quantify the T1 and T2 relaxation and volume fraction properties for water pools associated with intra- and extra-cellular water, water trapped within the lipid bilayers of the myelin sheath, and a non-exchange free water pool (i.e., cerebral spinal fluid). The volume fraction of the myelin-associated water, termed the myelin water fraction (MWF) is used as a surrogate measure of myelin volume, and has been verified via comparisons with histology (Wood et al., 2016), and used previously to investigate trajectories of early brain maturation (Deoni et al., 2012), myelin-function relationships throughout childhood (O'Muircheartaigh etal., 2013; O'Muircheartaigh et al., 2014), and myelin loss in adults with multiple sclerosis and other demyelinating disorders (Kolind et al., 2013, 2012, 2015). For children up to 5 years and 8 months of age, cognitive function and development was measured using the Mullen Scales of Early Learning (MSEL) (Mullen, 1995), a population-normed tool that provides standardised measures of fine and gross motor control, expressive and receptive language, and visual processing. In addition to domain specific scores, computed early learning composite (ELC) and verbal and non-verbal development quotients (VDQ and NVDQ) composite values reflect overall cognitive, verbal, and non-verbal functioning. Each of these age normalized composite values has a mean of 100 and standard deviation of 15.
+In addition to comparisons of brain and cognitive development trajectories associated with exclusive breast and formula-fed children, we further stratified the formula-fed children based on the main formula composition they received over the first 3 months of life and examined developmental differences between them. This analysis allowed us to more specifically investigate the role of nutritional composition on early brain growth. Finally, we extended this analysis to investigate the influence of individual nutrients on developmental myelin trajectories by examining the associations between specific formula nutrient levels and growth curve parameters.
+Overall, we find that compared to exclusive breastfeeding for 3 months, children who exclusively received formula milk have lower overall neurodevelopment, including both neuroimaging measures of myelination and measures of cognitive performance that persist into later childhood, even with groups matched for important socioeconomic and demographic factors. In addition, significant deviations in development are evident across children who received different formula compositions.
+Further, individual nutrient analysis suggests an important role for DHA, ARA, folic acid, sphingomyelin, iron, and phosphatidylcholine in brain development. These results further stress the importance of proper early nutrition for optimal brain development and, by consequence, cognitive outcomes in healthy children.
+Materials & methods
+Infant participants
+Infants included in this study were drawn from a large and ongoing longitudinal study of normal brain and behavioral development: the Brown university Assessment of Myelination and Behavior Across Maturation (BAMBAM) study (Deoni et al., 2012). BAMBAM currently includes more than 500 children recruited between birth and 5 years of age, and combines neuroimaging (MRI) measures, comprehensive observational and parent report measures of cognitive and behavioral development, on-going medical history information, biospecimen collection, and anthropometry. To obtain longitudinal measures of development, children are scanned and cognitive assessed at 6-month increments from time of recruitment until 2 years of age, and yearly thereafter. As a general study of neurotypical development, infants and young children with major risk factors for developmental, behavioral, or developmental disorders are excluded during enrolment. These risk factors included in utero exposure to alcohol, tobacco, or illicit recreational drugs; premature (<37 weeks gestation) or multiple birth; abnormalities on fetal ultrasound; complicated pregnancy (e.g., preeclampsia, gestational diabetes); 5 min APGAR scores < 8; NICU admission; history of neurological disorder or trauma (e.g., head injury, epilepsy); psychiatric or developmental disorders in first-degree relatives (including maternal depression requiring medication). Ongoing screenings, including the modified checklist for autism (MCAT) and child behavior checklist (CBCL) (Bilenberg, 1999; Chlebowski et al., 2013), and updated medical history information have been used to remove enrolled children with clinically concerning behaviors, diagnosed medical conditions, or head trauma following initial enrolment.
+A combination of retrospective and prospective infant nutrition data was acquired from parents using detailed medical histories and parent interview. This included type of infant formula used; percentage of breastfeeding; and length of exclusive breastfeeding. This information was updated at each study visit, which occurred approximately every 6 months for children under 2 years of age, and yearly for older children. Using this information, children were categorized as either exclusively infant formula-fed or exclusively (at least 90 days) breastfed. Children who were fed a combination of breastmilk and formula were excluded from this analysis. Infants within the exclusively formula-fed group were further sub-divided based on parental reports of the main formula composition they received in at least 80% of feedings throughout the child's first 3 months. All infant formulas consumed by children in this
+study were commercially available in the US.
+Using these criteria, 88 (34 female) exclusively formula-fed infants and young children were selected into group #1. This number included 21 (9 female) children who received formula #1; 28 (10 female) who received formula #2; and 39 (15 female) who received formula #3. A sample of 62 (23 female) exclusively breast-fed infants were also selected and matched to the overall formula-fed children with regards to mean age at scans (p = .24), pregnancy length (p = .39), birth weight (p = .52) and length (p = .09), male:female ratio (p = .85), parent marital status (p = .66), maternal and paternal education levels (p = .9 and p = .9, respectively), family size (p = 1), and the mean inter-scan interval (time between each set of repeat scans, p = .29). Group demographics are provided in Table 1. There were no significant differences in these demographic characteristics between the individual formula groups, mal-e:female ratio (p = .26), gestation (p = .17), birth weight (p = .08), birth length (p = .5), maternal or paternal education (p = .64), family size (p = .85), or marital status (p = .98). Two-tailed student t-tests were used to compare group mean age, pregnancy length/gestation duration, birth weight, birth length, and inter-scan interval. Chi-squared tests were used to compare group parental education level, marital status, and family size.
+A total of 231 scans were obtained on the breastfed children, and 221 on the formula-fed children (n = 42 for formula #1; n = 81 for formula #2; and n = 98 for formula #3). A pictorial display of the longitudinal imaging points and ages for each child is provided in Fig. 1. All child ages were corrected to a 40-week gestational age by subtracting the difference between 40 weeks and the child's actual gestation duration from the child's age.
+Imaging methods and analysis
+A multimodal imaging protocol was performed to assess brain morphology and myelination. mcDESPOT (multicomponent Driven Equilibrium Single Pulse Observation of T1 and T2) (Deoni et al., 2008) was used to quantify the myelin water fraction (MWF), a surrogate marker of myelin content or volume, throughout the brain. All infants were scanned during natural and non-sedated sleep using acoustically-reduced mcDESPOT imaging protocols described previously (Deoni et al., 2012) that comprise 8 T1-weighted spoiled gradient recalled echo (SPGR) images; 2 inversion (IR-) prepared SPGR images; and 16 T1/T2 weighted steady-state free precession (SSFP) images. Total imaging times ranged from 16 min for the youngest infants to 24 min for the older 4-year old and older children.
+All data were acquired on a Siemens 3T Tim Trio scanner equipped with a 12-channel head RF array. To minimize intra-scan motion, children were swaddled with a pediatric MedVac vacuum immobilization bag (CFI Medical Solutions, USA) and foam cushions. Scanner noise was reduced by lessening the peak gradient amplitudes and slew-rates, and using a noise-insulating scanner bore insert (Quiet Barrier HD Composite,
+UltraBarrier, USA). MiniMuff pediatric ear covers and electrodynamic headphones (MR Confon, Germany) were also used (Dean et al., 2014). Children were continuously monitored with a pediatric pulse-oximetry system and infrared camera. Data used for this analysis had no visible motion-artefacts present in their acquired data, however, 12 datasets (5 from the breastfed group and 7 across the formula-fed groups) were rejected for either incomplete data (2) or visible ghosting and ringing artefacts (10).
+Following data acquisition and inspection for image artefacts, conventional mcDESPOT preprocessing was performed consisting of image alignment (Jenkinson et al., 2002), non-brain signal removal (Smith, 2002), and correction for main and transmit magnetic field (B0 and B1) inhomogeneities (Deoni, 2011). A three-pool tissue signal model (the myelin-associated water; intra-extra axonal water; and a non-exchanging free-water pool) was then fit to the mcDESPOT data to derive voxel-wise MWF maps (Deoni et al., 2013a,b) using a stochastic region contraction approach (Deoni and Kolind, 2015).
+Each child's map was then non-linearly aligned to an existing study specific template using the Advanced Normalization Tools software package (Avants et al., 2011) using a previously described procedure (Deoni et al., 2012). Briefly, the high flip angle T1 weighted SPGR image from each child was non-linearly aligned to one of 14 age-specific templates (constructed at 3, 6, 9, 12, 15, 18, 21, 24, 30, 36, 42, 48, 64 and greater than 60 months), which have similar image size and tissue contrast. This transformation was then applied to the child's quantitative MWF image. An overall study template in approximate MNI space was also previously constructed from these age templates, with pre-computed transformations between it and each age template and this transformation was then applied to the MWF image.
+White matter masks, corresponding to 5 bilateral regions (frontal, temporal, occipital, parietal, and cerebellar WM) as well as the body, genu, and splenium of the corpus callosum were generated from the JHU white matter atlas (Oishi et al., 2011), registered to the study template, and superimposed onto each child's MWF map. Mean values for each region were calculated for each child and used for subsequent developmental analysis and trajectory modeling.
+Analysis of myelination trajectories
+To examine group-wise developmental differences between the breast and formula-fed infants, a non-linear mixed effects modeling approach was used to fit a modified Gompertz growth model (example shown in Fig. 2) (Dean et al., 2015) to the regional MWF data, with the form:
+Fig. 2. The modified Gompertz growth model used for all brain growth analysis labelled with relevant model parameters. Here, beta defines the onset of myelination; gamma is the initial rate of myelination; alpha is the MWF value at the shoulder point, or transition from rapid to slower myelination; and delta is the secondary slower rate of myelination.
+MWF(age) = aexp( — /} x ageexp — (/ x age + 3 x age})
+As shown previously, the modified Gompertz model provides the most robust and reliable fit to developmental MWF data compared to other models (Dean et al., 2015). Each of the 4 Gompertz curve parameters were compared between the breast and formula-fed groups using an unpaired t-test with significance defined as p < .001 (p < .05 corrected for the 32 regional and parameter comparisons).
+Examining this data further, we fit Gompertz growth models independently to children exclusively fed each of the three formula compositions (details provided below). Each model parameter was compared using an analysis of variance followed by a post-hoc Tukey test to determine which of the infant formula groups differed. Significance for these analysis was defined as p < .00052 (p < .05 corrected for the 96 comparisons performed).
+Cognitive assessments and analysis
+Alongside MR imaging, general cognitive ability and skills were evaluated in each child under 5 years and 8 months of age within 7 days of scanning using the Mullen Scales of Early Learning (Mullen, 1995). For older children, the Wechsler Intelligence Scale for Children, 5th Edition (WISC-V) was used. Due to the difference in cognitive assessment tool, we restricted our analysis here to only the MSEL data. The MSEL is a population-normed tool that provides domain-level assessment of fine and gross motor control, receptive and expressive language, and visual
+reception. In addition to age-normalized T-scores for each domain, the early learning composite (ELC) and verbal and non-verbal development quotients (VDQ and NVDQ, respectively) composite scores may be calculated that reflect overall cognitive ability, and verbal and non-verbal functioning. Longitudinal group differences (breast vs. all infant formula-fed, and between each formula brand) in ELC were examined using mixed effects modeling assuming a linear trend with age.
+Formula nutrient analysis and analysis
+To examine the potential relationship between specific nutrients to aspects of development, the nutritional composition of each infant formula composition was determined. Alpha-lactalbumin, Beta-lactoglob-ulin, ARA, DHA, Calcium, Phosphorus, Sodium, Potassium, Copper, Magnesium, Vitamin B12, and Folic Acid were measured in the analytical laboratories of Asure Quality, Auckland, New Zealand. The phospholipid profile (Phosphatidylcholine, Phophatidylinositol, Phosphatidylserine, Phosphatidylethanolamine, Sphingomyelin) of each product was determined at the analytical laboratories of Neotron, Italy using the method of Giuffrida et al. (2013). This method was validated for the quantification of Phosphatidylcholine, Phosphatidylethanolamine and Sphingomyelin. The method for sphyngolmyelin had a quantification limit of <200 mg/ kg.
+Nutrients that differed substantively (with a difference greater than 25% in concentration between the minimum and maximum value) between the 3 infant formulas were identified as: ARA, DHA, folic acid, phosphatidylcholine, and sphingomyelin (Table 2). Associations between these nutrient values and aspects of development were investigated by constructing a series of 4 general linear models (GLMs) that modeled each Gompertz model parameter as an outcome variable, and each nutrient value as a predictor variable. For the Gompertz parameters, all children were included in the same mixed-effects model.
+Results
+Fig. 3 contains the group-mean longitudinal MWF trajectories for the exclusively breast and all formula fed infants. In all investigated brain regions, we find differential patterns of development, with breastfed children qualitatively exhibiting a prolonged period of rapid development between 500 and 750 days of age, with an overall increase in myelin content by 2 years of age that persists throughout childhood. While the formula-fed group appears to show increased MWF before 1 year of age, they suffer a slower initial rate of MWF development between 1 and 2 years of age, and fail to reach the overall MWF magnitude of the breastfed group. Exploring these trajectory differences quantitatively (Table 3), in each brain region, there are statistically significant differences between all Gompertz growth model parameters in the frontal, temporal, and occipital white matter, and the body and genu of the corpus callosum. In the remaining regions (parietal and cerebellar white matter, splenium of the corpus callosum) there were significant differences between the a (asymptotic MWF value) and p (initial MWF onset) Gompertz terms; S (secondary rate of MWF development) was found to be significantly different in parietal white matter and splenium. In each of these regions, y (initial rate of MWF development) was not found to significantly differ between the groups.
+Examining differences between children who received different formula compositions, we find significant qualitative (Fig. 4) and quantitative (Table 4) differences in developmental patterns throughout the brain. In general, results from the ANOVA analyses revealed significant differences across the majority of brain regions examined and in almost all Gompertz model parameters. In particular, we note that children who received formula compositions with higher levels of DHA, ARA, choline, and sphingolipids (formulas #2 and #3) showed increased levels of myelin development. Of note, Formula #1, which showed to slowest myelin development, has the lowest concentration of DHA, ARA, and sphingomyelin, but has the highest concentration of iron and vitamin B12. Iron deficiency has previously been associated with cognitive impairments in older children (Lozoff and Georgieff, 2006).
+To determine if differences in cognitive maturation were also present in our sample of children, a linear mixed effects model was fit to the repeated Mullen composite scores (ELC, VDQ, and VNDQ). Results (Fig. 5 and Table 5) support our own prior results (Deoni et al., 2013a,b) as well as those of numerous cognitive studies comparing breast and formula-fed children (Horwood and Fergusson, 1998; Anderson et al., 1999; Kramer et al., 2008; Mortensen et al., 2002; Huang et al., 2014). Specifically, we find that while the mean trend for both groups fell within the normative range (85-115), there were statistically significant differences in the rate of cognitive change with age (slope) and a general increase in the mean (intercept) between the breast and formula-fed groups. Breastfed children exhibited an overall increase in ELC, VDQ, and NVDQ scores, and increased rates of development in ELC and VDQ, and an attenuated decrease in NVDQ with age. Thus, early differences in cognitive function were found to persist, and in the case of VDQ and NVDQ increase, into childhood.
+Repeating this same analysis for each infant formula, we find (Fig. 6 and Table 6) an overall correspondence between brain development profiles and trajectories of cognitive maturation. Specifically, children who received Formula #1, which shows the slowest myelination profile across the majority of brain regions, also have the most pronounced decline in cognitive function across early childhood. Formula #2, which had the closest myelination trend to breastfeeding also exhibit cognitive trends that are most consistent with breastfeeding. These results suggest not only the importance of early nutrition to brain and cognitive development, but also suggest a strong link between brain structure and cognitive performance.
+Investigating the influence of specific nutrient concentrations on the myelination model trajectory parameters (Table 7), we find that whilst each of ARA, DHA, folic acid, iron, choline, sphingomyelin, B12, and phosphatidylcholine contribute to myelination, sphingomyelin and phosphatidylcholine appear to have the most diffuse influence throughout the brain with the remaining nutrients more associated with development in focal brain areas.
+Discussion
+The impact of nutrition on human infant brain myelination has traditionally been indirectly investigated via studies of cognitive performance or using evoked potentials (Pivik et al., 2007), with few neuroimaging studies performed throughout infancy and early childhood (Deoni et al., 2013a,b; Luby et al., 2016). This study, therefore, adds to
+the existing literature examining the role of early life nutrition and feeding choice in infancy, presenting the first longitudinal neuroimaging results demonstrating differences in profiles of myelination and cognitive development between children exclusively breast or formula fed, and between children who received different infant formulas. These data suggested an important role for DHA, ARA, sphingomyelin and choline in early brain development, which was subsequently confirmed by examining the associations between these nutrients and brain growth model parameters.
+On a general level, our results indicate that exclusive breastfeeding for at least 3 months is associated with improved myelination diffusely throughout the brain by 2 years of age, including early and late maturing brain regions and networks associated with a broad array of cognitive and behavioral skills. Supporting this structure-function link, we also show improved overall cognitive ability and rates of cognitive development, including verbal and non-verbal functioning, in breast-fed children versus those who received only infant formula. Examining the longitudinal trends within our data we find that these structural and cognitive differences become evident by approximately 18 months of age (depending on brain region), and persist at least into early childhood (at least 5.5 years of age). Observed differences in myelination, an essential
+element of the brain's white matter structure (Fields, 2010; O'Brien and Sampson, 1965), may be predictive of previously observed white matter volume and integrity changes in older children and adolescents who were breastfed as infants (Isaacs et al., 2010). The importance of myelination to brain connectivity may also link our findings to prior reports of altered functional activation and connectivity in breastfed infants (Ou et al., 2015).
+In order to more specifically link observed breast vs. formula-fed differences to nutrition, as opposed to other potential socioeconomic or demographic aspects, we also contrasted the brain and cognitive developmental profiles in children who received different formula compositions (Figs. 4 and 6). Here we found significant and consistent differences in the profiles of myelination and cognitive maturation, with children who had the lowest myelin development overall having the worst cognitive scores and vice-versa. Of note, the formula compositions associated with the highest myelin levels and cognitive scores also had the highest concentration of long-chain PUFAs (DHA and ARA), choline, folic acid, sphingolipids (sphingomyelin) and phosphatides (phosphatidylcholine). This finding is in strong agreement with prior nutrition literature. Long-chain poly-unsaturated fatty acids (LC-PUFAs), in particular, ARA and DHA, help promote neuronal growth and white
+matter development (Innis, 2007). Preclinical studies of animals withheld AA and DHA via early weaning have reduced myelin basic protein expression, consistent with reduced myelin content (Kodama et al., 2008; Bruno and Tassinari, 2011). While folic acid deficiency is more often associated with neural tube defects (Bower, 1995), preclinical studies have also shown that postnatal deficiencies can negatively affect the fatty acid composition of myelin (Chida et al., 1972). Choline, a precursor to phosphatidylcholine as well as sphingomyelin, and choline-containing phosphoglycerides, comprise more than 10% of the lipid weight of myelin (Norton and Cammer, 1984). In in vitro studies of choline deficiency, significant reductions in phosphatidylcholine (—49%) and sphingomyelin (—34%) concentrations were found compared to cells grown in a choline-rich medium (Yen et al., 1999). These results are mirrored in in vivo human studies, demonstrating a 30% decrease in circulating phosphatidylcholine levels following a 3-week choline-deficient diet. Finally, both phosphatidylcholine and sphingomyelin are critical components of myelin, with dietary supplementation of sphingomyelin previously shown to improve myelination in a pre-clinical model (Oshida et al., 2003).
+In contrast, formula compositions high in iron, but lower in LC-PUFAs and sphingolipids, appear to be associated with slower and reduced overall myelination. Although the role of iron in myelin synthesis is not yet fully understood, both animal and human infant studies have revealed associations between iron deficiency and hypomyelination, reduced oligodendrocyte functioning, and decreased myelin basic protein concentrations. Iron may also play a specific role in myelin synthesis through oligodendrocyte energy metabolism, and fatty acid synthesis. Children with prolonged iron deficiency also suffer a variety of behavioral and cognitive impairments (Saloojee and Pettifor, 2001; Lozoff and Georgieff, 2006; Congdon et al., 2012). However, while iron deficiency has been well studied, little is known regarding potential outcomes associated with iron over supplementation. Breastmilk contains little iron
+content, and iron deficiency may arise after 4 months in exclusively breastfed infants (Kramer and Kakuma, 2012; Ballard and Morrow, 2013). However, healthy non-anaemic infants supplemented with iron exhibit reduced growth (Dewey et al., 2002) and increased fever and illness (Pasricha et al., 2013).
+There are important caveats to our examination of different infant formulas, including: 1. The retrospective nature of our investigation; and 2. The high variability of these nutrients. While formula and feeding information for this study was acquired prospectively, nutritional composition analysis was performed retrospectively using a single time assay. Thus, the specific nutritional formulations may have changed in the 6 years since the earliest imaging data was acquired. It is also important to note that it is not possible, using these observational data alone, to infer which particular nutrient (or combination) is most associated with preferential myelination trajectories. Such information is likely only be provided by pre-clinical models in which individual nutrients may be carefully varied and the effects followed.
+Given the temporal delay between infant feeding and the first appearance of differences in both myelination and cognition between the breast and formula-fed children, it is likely that prolonged breastfeeding durations, follow-on complementary feeding and other environmental influences (e.g., parental interaction) are important but unexamined contributors to our results. In past cross-sectional analysis examining associations between breastfeeding duration (and including complementary feeding past 6-months) and MWF, we showed prolonged breastfeeding was associated with increased myelination in the cerebellum, internal capsule, and parietal and temporal white matter (Deoni et al., 2013a,b). In addition to prolonged breast-feeding, numerous other environmental conditions have previously been associated with differences in cognitive development and outcomes in children, such as family socioeconomic status (SES) (Noble et al., 2015), parental education (Cromwell and Panksepp, 2011), parent-child interaction (Swain et al.,
+2007), physical activity (Best, 2010), and sleep quality (Peirano and Algarín, 2007). Although we aimed to mitigate these SES-related factors by matching children on the basis of maternal and paternal education levels, family size, and parental marital status, it is difficult to discount their contribution without accurate and detailed assessments of each.
+A variety of environmental and economic conditions have previously been associated with differences in cognitive development and outcomes in children, such as family socioeconomic status (SES) (Noble et al., 2015), parental education (Cromwell and Panksepp, 2011), parent-child interaction (Swain et al., 2007), physical activity (Best, 2010), and sleep quality (Peirano and Algarín, 2007). Although we aimed to mitigate these SES-related factors by matching children on the basis of maternal and paternal education levels, family size, and parental marital status, it is difficult to discount their contribution without accurate and detailed assessments of each. However, while the effect of these socioeconomic differences to cognitive outcomes in breastmilk and formula fed children is increasingly documented (Walfisch et al., 2013), the expected differences between formula-fed children may be less.
+An additional, and as yet unresolved, question related to neuro-cognitive outcomes associated with breastfeeding is whether they arise due to the specific nutritional, hormonal, and other constituents of breastmilk per se; if they are driven by maternal-child interaction and
+other environmental differences (Walfisch et al., 2013); or are a result of a combination of the two (Reynolds, 2001). Our study does not attempt to resolve this quandary, but does provide important support for the role of early nutrition, with specific emphasis on nutrients either involved in myelin synthesis or compositional components of myelin. This is particularly evidenced by the differences in development observed amongst the exclusively formula-fed children.
+In this work, we have focused on de novo myelination given its fundamental role in learning and cognition (Nagy et al., 2004; Fields, 2008) and its previously demonstrated sensitivity to nutrition (Lozoff and Georgieff, 2006; Innis, 2007; Bruno and Tassinari, 2011; Kodama et al., 2008). However, other developmental processes, including synapto-genesis and synaptic pruning are also important contributors to brain connectivity, brain function, and cognition throughout this age period. Like myelination, these processes may also be differentially influenced by nutrition (Kafouri et al., 2013; Luby et al., 2016). An MR imaging measure related to synapse density is cortical thickness, commonly quantified using conventional anatomical imaging and brain segmentation (Fischl, 2012). Analysis and examination of cortical thickness differences was not performed here owing to the challenges of accurate gray-white matter segmentation in young children, particularly those under 1 year of age. A further possibility is that the increase in myelin content measured here
+does not reflect more myelin per axon, but more myelinated axons overall. One approach to investigating this further is the use of neurite orientation dispersion and density imaging (NODDI) (Zhang et al., 2012) and myelin g-ratio imaging (Dean et al., 2016), which would inform on both axonal density and mean axon myelin thickness. High resolution structural and NODDI imaging data were, unfortunately, not collected across all children included in this analysis and, thus, investigations of these additional metrics remains a topic for future investigations.
+Conclusions
+While the exact mechanisms that underlie the previously demonstrated brain myelination and cognitive advantages differences in children, adolescents, and adults who were breastfed as infants remain unclear, our results presented here add to the growing evidence and consensus that early and exclusive breastfeeding is associated with improved neurodevelopment, including de novo myelination, and cognitive outcomes. Our longitudinal findings further suggest that early developmental differences persist into childhood and may predict changes previously identified in adolescents and adults. Furthermore, different compositions of infant nutrition appear to result in different
+658
+patterns of myelin development, with some being closer to the myelin trajectory associated with breast-fed infants than others. With respect to potential nutritional contributors, our analysis highlights the importance of known neuro-associated nutrients, including long-chain polyunsaturated fatty acids as well as the important myelin components phosphatidylcholine and sphingomyelin to early neurodevelopment.

Editorial.txt

@@ -0,0 +1,21 @@
+The debate about the impact of the economy on suicide risk has progressed from untested theories to more complex epidemiological studies and much idiographic data in between. The subject illustrates the interaction of societal effects with the individual’s personal risk profile and vulnerability. Early theorists proposed that economic recession could increase suicide rates because of the stress and hardship that poverty creates (Brenner, 1979; Stack, 1981) as well as the potential loss of social status and connectedness (Durkheim, 1897). Some attribute suicide to the interplay of economic and social factors (Lester, 2001), while others focus solely on the economic contribution such as higher income decreasing the opportunity cost of suicide (Hamermesh & Soss, 1974). Conversely, it has also been proposed that the suicide rate would decline during times of economic hardship, because individuals could blame the macro-economy for feeling miserable instead of themselves (Henry & Short, 1954). Much research has examined the relationship between personal income level, the macro-economy, and suicide rates, and a few studies have also examined the interaction of regional or national economic state with psychiatric illness.
+How does an economic crisis affect suicide rates and which economic variables play a role? Does an economic downturn impact all social groups, regions within a nation, and different countries similarly? How does psychiatric illness and its treatment interact with economic variables and potentially affect suicide rates? An examination of the available literature answers some of these questions and illuminates the consideration of the best approaches for addressing the increased suicide rates seen during recent times of economic hardship as in the 2008 recession.
+The initial literature involves idiographic studies that consist of case reports of suicide apparently related to eco
+© 2017 Hogrefe Publishing
+nomic factors. In the 1997 financial crisis in South Korea (Watts, 1998) a woman indicated that she was going to kill herself because the devaluation of the Korean currency meant she could no longer afford her son’s college tuition. This case and many other reports of so-called economic suicides lack a systematic review of the person’s psychiatric condition and other factors that contribute to risk, meaning we do not know why this specific life stress was so deadly for this person.
+Reports on groups at particularly high risk of suicide are sometimes more informative since more general factors can be identified. Suicide rates in Canadian Inuit people, which are three-to-four times higher than Canada’s average suicide rates, have also been considered in the context of multiple socioeconomic factors (Leenaars, Anawak, & Taparti, 1998). The impact of the fur trade collapse and high unemployment rates are mentioned as potential contributing factors but are not studied systematically in this idiographic report. Increases in Australian farmer suicide rates - which are higher than those of the average rural population and male national rates - correspond with declines in trade over the same time period (Page & Fragar, 2002). An increase in male unemployment - among increases in violence, substance misuse, and alcohol consumption - was postulated as a potential psychosocial stressor to explain the increase in male suicide rates in England and Wales during 1975-1990 (McClure, 2000). Unemployment and worsening poverty associated with neoliberal structural adjustment following the 1998 economic crisis in South Korea were posited as contributing factors of the increased suicide rates observed (Khang, Lynch, & Kaplan, 2005). Similarly, benefits of economic growth - such as decreased unemployment - were proposed as explanatory factors of the observed decrease in suicide rates in China during 2009-2011 (Wang, Chan, & Yip, 2014). Unemployment, change in income, and household debt were posited as fac-
+Crisis (2017), 38(3), 141-146
+DOI: 10.1027/0227-5910/a000487
+tors contributing to the immediate rise in suicide following the 2008 European financial crisis (Karanikolos et al., 2013). Studies of this sort can suggest links between factors and suicide, but cannot estimate the attributable risk of the suggested economic factors because other factors may be contributing to the higher suicide rates.
+Analysis of suicide rates and economic variables at an epidemiological level may allow consideration of more variables like psychiatric disorder rates, treatment, regional income levels, unemployment rates, and demographics, but this analysis carries the risk of the ecological fallacy and does not permit conclusions to be drawn about causality. An examination of whether differences in the structure of agricultural production explained inter-state variation in suicide rates in India found positive relationships between the percentage of marginal farmers, cash crop production, and indebted farmers and suicide rates (Kennedy & King, 2014). Increased suicides in rural Japan led to the consideration of the effects of industrialization and urbanization as contributing factors; higher suicide rates were found in areas with a sparse population and a non-prosperous economy (Kurosu, 1991). Higher suicide rates and economic variables - unemployment and urbanization - have been studied in the context of other epidemiological studies, and relationships have been observed between negative economic impact and increasing suicide rates (Álvaro-Meca, Kneib, Gil-Prieto, & Gil de Miguel, 2013; Otsu, Araki, Sakai, Yokoyama, & Voorhees, 2004; Preti & Miotto, 1999; Thomas & Gunnell, 2010; Yip, Law, & Law, 2003). By contrast, an epidemiological study of Ontario farm suicides failed to find any associations between economic indicators - including number of farm bankruptcies, net farm income, and loan and unemployment rates - and farm suicide rates (Pickett, Davidson, & Brison, 1993). However, such studies cannot answer the question of whether there is a causal relationship nor do they provide any useful estimate of an economic factor’s potential importance in terms of attributable risk for suicide as an outcome.
+More helpful are time series studies and correlative studies of regional differences in suicide rates, per capita income, unemployment rates, and economic state. Time series studies have the advantage of examining the subsequent impact of changes in potentially relevant variables to seek a possible causal relationship. A causal change must precede an outcome attributable to that cause. The difficulty in concluding that such a relationship exists is due to the need to have measured all the relevant variables and interactions.
+A decline in the economy at a national or regional level is generally associated with higher suicide rates in time series studies and cross-sectional regional correlative studies. Associations between suicide rates and economic
+variables have been studied in many countries. National suicide rates in the United States during business cycles, 1928-2007, rose during economic recessions and fell during economic expansions (Luo, Florence, Quispe-Ag-noli, Ouyang, & Crosby, 2011). Greater increases in suicide rates were observed in the countries most affected by the Asian economic crisis of 1997-1998, with higher unemployment being most strongly associated with suicide rate increase (Chang, Gunnell, Sterne, Lu, & Cheng, 2009). Similar findings have been reported in other time series studies (Andrés, 2005; Brenner, 1979; Ceccher-ini-Nelli & Priebe, 2011; Corcoran & Arensman, 2010; Corimer & Klerman, 1985; Kwon, Chun, & Cho, 2009; McKeown, Cuffe, & Schulz, 2006; Motohashi, 1991; Park, Lee, & Kim, 2003; Pompii et al., 2014; Ruhm, 2000; Stuckler, Basu, Suhrcke, Coutts, & McKee, 2009; Tapia Granados, 2005; Vigderhous & Fishman, 1978; Zhang et al., 2010). A minority of studies found the opposite relationship (Neumayer, 2004) or no association between economic fluctuations and suicide rates (Hintikka, Saarinen, & Viinamaki, 1999; Rancans, Salander Renberg, & Jacobsson, 2001).
+Other perspectives have emerged from comparison of the effect of economic variables on suicide rates across different countries. The association between unemployment and suicide rates has been shown to be stronger in the United States compared with other countries, in which the effect of unemployment is weak or nonexistent (Yang & Lester, 1995). Economic factors were found to play a role in influencing US suicide rates but not Taiwan suicide rates in the period 1952-1984 (Yang, Lester, & Yang, 1992). The authors attribute this difference to the fact that being poor in Taiwan is not shameful, whereas in the United States being poor or becoming poor involves more of a loss of place in society and has a greater effect in a consumer-oriented society. Similarly, a lack of common socioeconomic predictors - unemployment rates, annual percentage change in gross national product, female labor force participation, and divorce rates - of suicide rates has been observed in the United States and Japan (Lester, Motohashi, & Yang, 1992). While a negative impact of unemployment on overall suicide rates was found in both Japan and the United States, the relationships between suicide and the other socioeconomic variables differed in the two countries. The correlation of GNP with suicide rates is negative in Japan and positive in United States. Female labor force participation correlation with suicide rates is negative in the United States and positive in Japan. The impact of employment conditions on suicide differed between Hong Kong and Taiwan: Suicide rates fell in Hong Kong but increased in Taiwan as employment conditions improved (Chen, Yip, Lee, Fan, & Fu, 2010). Such national differences have been attributed to national income
+level: unemployment having a positive relationship with suicide rates in high-income countries, but a negative association in low-income countries (Noh, 2009). Annual growth rates for industry and health-care expenditures are additional economic factors that distinguished European countries with higher suicide rates (Ferretti & Coluc-cia, 2009). Countries belonging to the high suicide rate group had lower health-care expenditures, a lower at-risk-of-poverty rate, higher percent total unemployment, and higher annual growth rates compared with the countries belonging to the low suicide rate group.
+How income level plays a role in affecting suicide rates during economic hardship is unclear. Some studies find that higher income is associated with higher suicide rates (Hamermesh, 1974; Jungeilges & Kirchgãssner, 2002) while others find that higher income is associated with lower suicide rates (Brainerd, 2001; Chuang & Huang, 1997; Hamermesh & Soss, 1974; Neumayer, 2003). Others report suicide rates are insensitive to income levels (Andrés, 2005). Findings for unemployment as an economic predictor are similarly mixed. Some studies find that higher unemployment rates are associated with higher suicide mortality (Blakely, Collings, & Atkinson, 2003; Brainerd, 2001; Chuang & Huang, 1997; Hamermesh & Soss, 1974; Neumayer, 2003), while others find no impact of unemployment rates on suicide rates (Andrés, 2005; Hamermesh, 1974; Kunce & Anderson, 2002).
+Noneconomic factors such as rates of psychiatric illness and their treatment levels should also be considered since untreated psychiatric illness is known to be present in most suicide decedents, and psychiatric illness can affect employment status and income. US county-level suicide rates are inversely related to median income with wealthier counties having lower suicide rates (Gibbons, Hur, Bhaumik, & Mann, 2005). Importantly, authors noted that higher suicide rates in rural areas were associated with fewer antidepressant prescriptions, lower income, and relatively more prescriptions for older antidepressants, tricyclic antidepressants, a possible index of how up to date doctors were in terms of medical education on new medications. The findings suggest an impact of access to affordable, adequate medical care of major depression on suicide rates. Lower per capita income reduces health-care resources available to people. Lower per capita income may also mean fewer and poorer health-care resources in a community. Higher-income areas may also have better emergency medical care, increasing the chance of survival after a suicide attempt (Neumayer, 2003). Higher suicide rates correlate with higher levels of rurality (Singh & Siahpush, 2002). Rural areas may be vulnerable owing to a smaller tax base as a result of both fewer people and lower per capita income, in addition to being less attractive to doctors as a place of work and living.
+How demographic subgroups are differentially affected can reveal other factors that affect suicide rates such as social cohesion, religion, and the stigma of psychiatric illness. Until recently the main breadwinner in a household has been a male aged 25-65 years. Because the responsibility for supporting the family falls most heavily on this demographic subgroup, it is the group that may feel the most stress when an economic decline adversely affects their capacity to earn the same level of income as before the recession. Suicidal behavior of older people has been shown to be more sensitive to fluctuations in unemployment compared with the suicidal behavior of younger people (Hamermesh & Soss, 1974), perhaps because the chances of re-employment are lower for older people, and the loss of income and social status is greater. More broadly, an economic decline may force a change in the social group to which the family belongs and impact the family’s housing, schooling for their children, vacations, automobile ownership, clothing, and many other social-defining characteristics. It is therefore of note that many studies identify spikes in suicide rates to be more pronounced in males (Aihara & Iki, 2002; Berk, Dodd, & Henry, 2006; Brainerd, 2001; Corcoran & Arensman, 2010; Huang, 1996; Inoue et al., 2007; Pompii et al., 2014; Preti & Miotto, 1999; Rancans et al., 2001; Schapiro & Ahlburg, 1982; Yang, 1992), and recently particularly in middle-aged males (Andrés, 2005; Corcoran & Arensman, 2010; Jungeilges & Kirchgãssner, 2002; Khang et al., 2005; Luo et al., 2011; Pompii et al., 2014; Schapiro & Ahlburg, 1982).
+Life stressors can trigger a major depressive episode or other psychiatric disorders in vulnerable individuals (Van Heeringen, 2012). An economic downturn may be such a stressor (Dooley, Catalano, & Wilson, 1994). Significant increases in the prevalence of major depression corresponding with economic hardship have been observed in cross-sectional studies (Economou, Madianos, Peppou, Patelakis, & Stefanis, 2013; Lee et al., 2010). Measures of low-economic status, such as low income and unemployment, have been found to be associated with a higher incidence of suicidal thoughts (Gunnell, Harbord, Singleton, Jenkins, & Lewis, 2004), increased risk of suicide (Gerdtham & Johannesson, 2003), and higher attempted suicide rates (Economou, Madianos, Peppou, Theleritis, et al., 2013; Ostamo, Lahelma, & Lonnqvist, 2001). Along with unemployment, fear of losing one’s job has adverse effects on psychological health (Reichert & Tauchmann, 2011) and exacerbates depression and suicidal thinking (Gunnell, Platt, & Hawton, 2009). An economic recession can also trigger a review of workforce needs by companies, and individuals with impairment due to a psychiatric illness may be more likely to be laid off at such times. This process would be detected as a disease by economic
+decline interaction. Some studies have examined such an effect. For example, about half of the association between unemployment and increased suicide risk was attributable to a confounding mental illness in a New Zealand sample (Blakely et al., 2003).
+Evidence exists to support a complex relationship between economic conditions and suicide. Outcome depends on both (a) adverse economic factors that can reduce per capita income and (b) the reduced tax base that can degrade quality and quantity of health care that a community can offer its citizens. From the other perspective, individuals with psychiatric illness can find it more difficult to find and hold better-paid jobs or any job. Finally, an economic downturn can have a disproportionately adverse economic effect on certain demographics like males 25-65 years of age, who are the household’s main income source, and on those with psychiatric illness, or older individuals, whose capacity to compete in the job market is not as good and who are also more vulnerable in terms of stress-triggered psychiatric or other medical disorders.

Effect-of-the-Brazilian-cash-transfer-programme-on-suicide-rates-a-longitudinal-analysis-of-the-Brazilian-municipalitiesSocial-Psychiatry-and-Psychiatric-Epidemiology.txt

@@ -0,0 +1,46 @@
+Introduction
+Worldwide suicide is a public health problem causing almost one million deaths every year [1]. Among countries, Brazil ranks eighth in the incidence of suicide, with an age-standardized rate of 5.8 per 100,000 inhabitants, in 2012 [1].
+Inside Brazil, variations may be observed between regions, with higher rates in the Southern region (9.8/100,000 in 2012), followed by the Centre-West (7.6/100,000 in 2012) [2]. Incidence is approximately three times higher amongst men than women, and there is a higher rate amongst old people (8.0/100,000 in 2012) [2].
+The association between suicide and socio-economic factors has been well-documented in the literature [3, 4]. A recent review assessed the effects of poverty on the entire spectrum of suicidal ideation and behaviour in low and middle-income countries and found that 62% of studies reported an association of suicide with worst economic conditions, income shocks and unemployment [5]. At the aggregate level, studies have found an inverse association between public spending on social policies and suicide mortality in Europe [6] and the United States [7]. In Brazil, a positive association has been seen between the suicide rate and the income inequality, where every 10-point decrease in the Gini Index results in a 5.5% reduction in the suicide rate. An inverse association of municipal suicide rates and average
+per capita income was also reported [2]. It is known that suicide is associated with mental health, in particular, mood disorders, such as depressive conditions [8]. As poverty is associated with mental health [9], this could be a possible mechanism to explain the link between suicide and poverty.
+It seems that poverty and income inequality may have some important impact on suicide rates [2, 10, 11] and therefore, actions focused on decrease poverty such as cash transfer programmes can impact on suicide [12]. In Brazil, the conditional cash transfer programme (CCT), branded as Programa Bolsa Familia (PBF), established in 2004, is an important socio-economic intervention that aims to attenuate the effects of absolute poverty through a minimum cash transfer for beneficiary families, and to break the intergen-erational cycle of poverty through investment in education and health conditionalities [13].
+The effects of the PBF on population health in Brazilian municipalities have been investigated, and the PBF has been associated with the improvement of nutritional status [14], and health outcomes such a as infant mortality [15] and leprosy [16]. However, no studies have yet been conducted to evaluate its effects on suicide. It is important to investigate whether a conditional cash transfer programme, such as the PBF, helps to attenuate the incidence of suicide to support the adoption of poverty reduction strategies, enabling potential changes and improvements to the programme, such as encouraging the inclusion of mental health
+care in its conditionalities. This study, therefore, aims to assess the effect of PBF coverage on suicide rates in Brazil. The mechanisms linking BFP to suicide are conceptualized in Fig. 1.
+Method
+We conducted an ecological longitudinal study, which combines an analysis of multiple observation units with a temporal trend design. We used panel data for the 5507 Brazilian municipalities which existed at the time of analysis. All the municipalities were examined through repeated observations over the 9 years from 2004 to 2012.
+All data came from the Health Informatics Department of the Brazilian Ministry of Health [17], including the mortality data for each municipality that was collected from the Mortality Information System (Sistema de Informagao sobre Mortalidade: SIM). Socio-economic and demographic variables were obtained from the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatistica: IBGE) [18] and the Primary Care Information System (Sistema de Informagao da Atengao Basica: SIAB) [17]. PBF coverage was obtained from the Ministry of Social Development’s database [19].
+Suicide, the outcome variable, was defined as the cause of death recorded as “intentional self-harm” according to the
+10th Edition of the International Classification of Diseases (ICD-10), codes X60-X84. The suicide rate was calculated at the municipal level and standardized for age in 5-year intervals using the direct method and taking the World Health Organization (WHO) population as a reference. For each municipality and year of analysis, we calculated a general and gender-stratified suicide rates. Suicide rates were calculated for individuals aged 10 years or above since this event is infrequent before this age.
+The Programa Bolsa Familia (PBF)—the main exposure—is a conditional cash transfer programme, established in 2004. In 2012, cash transfers were made through a basic benefit of 70 Brazilian Reals (R$), aimed at extremely poor families, that is families with a monthly family income of up to R$ 70.00. Furthermore, there was a variable benefit of 32 Brazilian Reals granted to families with children, adolescents, pregnant and breastfeeding women, who had a per capita household income of up to R$ 140.00 [13]. Health conditionalities include the monitoring of vaccinations and nutritional surveillance of children under 7 years old, as well as pre-natal care for pregnant and postpartum women. In education, these include 85% school attendance for children and adolescents aged 6-15 years old and 75% attendance for young people aged 16 and 17 years [13]. In our study, we used the PBF coverage of the target population, calculated by dividing the number of families who receive PBF by the total number of eligible families, transformed into percentages and multiplied by 100 [18]. The coverage was then categorized according to three levels (< 30%; > 30% and < 70%; and > 70%). Then we analysed whether the persistence in the time of high coverage would have an effect on the rates of suicide and hospitalization for suicide attempts. Thus, the duration of high PBF coverage of 70% or more was categorized into 3 time periods (coverage < 70% for all years; > 70% for 1 year; > 70% for 2 years; > 70% for 3 or more years).
+We included socio-demographic, economic and social welfare variables as controls: percentage of employed population—percentage of the employed population aged 16 years or above; rate of urbanization—percentage of people living in cities; percentage of people with low education levels—percentage of people with incomplete primary education (up to 8 years of schooling); percentage of people on a low income—percentage of population resident with a monthly per capita household income of up to R$ 140.00; percentage of separated people—percentage of people who declared themselves divorced; percentage of households with one resident—percentage of private households occupied by only one resident; percentage of individuals who declared themselves to be Pentecostal. We also included a health care variable as control: Family Health Programme (Programa Saude da Familia: PSF) coverage—the ratio between the total number of people registered on the
+programme divided by the municipal population. The selection of all these variables was based on evidence of their association with suicide, as seen in the literature [4, 8, 9].
+Statistical analysis
+We used models of negative binomial regression to assess the association between PBF coverage and suicide rates. The time variable was introduced into the models to control for the effect of time on political changes and secular trends which might affect all the municipalities [20]. The estimated regression was (Yit) = ai + p1BFit + pXnit + yt + uit, where, Yit refers to the number of suicides divided by the population residing in municipality i in year t; ai was the fixed effect for municipality i which captures all the unobserved time-invariant factors; BFit was Bolsa Familia Programme coverage for municipality i in year t; Xnit the value for each co-variable in the model, including all the socio-economic, demographic and social welfare determinants, in municipality i in year t; yt was the specific effect of time; and uit the error. To assess whether municipality size might influence the results, we conducted stratified analysis by municipalities with a population lower than or equal to 10,000 inhabitants, between 10,000 and 50,000, and greater than or equal to 50,000 inhabitants.
+The Hausman specification test was conducted to analyse the robustness of results in relation to the regression model selected for the panel data. We used the Akaike (AIC) and Bayesian Information Criteria (BIC) to establish the model that best fitted the data was produced by Poisson or negative binomial regressions [20]. All the statistical analyses were conducted using Stata (version 12).
+This study used exclusively secondary data from the public domain, therefore; approval from a Research Ethics Committee was not a requirement.
+Sensitivity analysis
+We checked the robustness of our results performing several sensitivity analyses. First, we tested using several alternative model specifications with random and fixed effects and Poisson models (S1 Appendix, Table A). Second, models specifications with sequential addition of covariates were tested (S2 Appendix, Table B). Third, we repeated the best-fit model including only municipalities with accurate vital information [21] (S3 Appendix, Table C). Fourth, we tested different classifications of PBF coverage (S4 Appendix, Table D). Fifth, because of the potential for misclassification of external causes and suicide, we tested the effect of BF Programme over suicide rates adjusted to ill-defined causes of death (S5 Appendix, Table E). Sixth, we conducted analyses on hospitalizations for attempted suicide to verify common data trends between mortality and suicide attempted,
+given that the literature suggests that suicide and suicide attempts have similar phenomenological characteristics [8] and may have common contextual determinants (S6 Appendix, Table F). Seventh, we have also performed sensitive analyses including “hospitalizations rates due to psychiatry problems” as a controlling variable (S7 Appendix, Table G).
+Results
+From 2004 to 2012, there was a 4% increase in suicide rates. The PBF coverage increased by 46.64%, accompanied by improvements in socio-economic conditions, with the percentage of people on low incomes falling by 26.76%, the employment rate rising by 7.26% and those with low education levels falling by 21.83%. About the socio-demographic data, the average urbanization grew over this period (7.02%), as did the proportion of divorced people (59.75%). About healthcare, the average Family Health Programme coverage in the municipalities rose from 60% in 2004 to 77.23% in 2012 (Table 1).
+Table 2 shows the crude and adjusted associations between levels of municipal PBF coverage and suicide rates, presenting a statistically significant dose-response relationship, even after the adjustment for the socio-economic, demographic and social welfare co-variables. Suicide rates were significantly lower in municipalities with coverage between 30 and 70% (RRcrude 0.966; CI 95% 0.960-0.972) and over 70% (RRcrude 0.942; CI 95% 0.936-0.947) compared with low coverage municipalities (less than 30%).
+An increase in the percentage of the population with low incomes, with low educational levels, separated, in families with one resident was associated with an increase in suicide rates. The percentage of people employed, declared to
+be Pentecostal and urbanization rate by municipality were negatively associated with suicide rates. Similar results were obtained when the analysis was repeated stratifying by the size of municipal population (less than 10,000; between 10,000 and 50,000; and over 50,000), maintaining the magnitude and direction of association found between suicide rates and PBF coverage. There was no statistically significant association between suicide rates and Family Health Programme coverage (Table 2).
+When we assessed the association between the duration of high PBF coverage of 70% and suicide rates, the results demonstrated that an increase in the duration of high PBF coverage of 70% or more was associated with a fall in suicide rates. The effect increased as the number of years with the persistent high coverage increased (Table 3).
+In relation to gender, an increase in PBF coverage duration of 70% or more was associated with a fall in suicide rates amongst women (Table 3).
+Sensitivity analyses
+Sensitivity analyses demonstrate that our findings are robust as none of these sensitivity analyses changed our main findings. Alternative model specifications (S1 Appendix, Table A) demonstrate the stability of the results. We found that controlling for different factors, such as fixed or random effects and different covariates, did not change our results (S1 and S2 Appendix, Table A and B). Different classifications of PBF were tested, and our results were the same for all classifications (S4 Appendix, Table D). Evaluating proportion ill-defined cause adjusted in the model to control by mistakes in the classification of suicide, demonstrated that an increase in PBF coverage is associated with a reduction in suicide rates in Brazilian municipalities, even following
+adjustment for this variable (S5 Appendix, Table E). Repeating the analyses only including municipalities considered to have accurate vital information the results remained similar to the analyses including all the Brazilian municipalities (S3 Appendix, Table C). Evaluating the effect of PBF coverage on hospitalizations for attempted suicide rates in Brazil, demonstrated that an increase in PBF coverage is also associated with a reduction in hospitalizations for attempted suicide rates in Brazilian municipalities (S6 Appendix, Table F). We have added in the model the hospitalizations due to psychiatry problems rates, and our results remained similar (S7 Appendix, Table G).
+Discussion
+The results of our study demonstrate that an increase in PBF coverage was associated with a reduction in suicide rates in Brazilian municipalities, even following adjustment for socio-economic, demographic and social welfare factors. The effect of the PBF increased when, alongside high coverage (equal to or greater than 70%), this level of coverage was maintained for several years. We also conducted robustness sensitivity testing, and the main results were maintained in both suicide and hospitalizations for attempted suicide rates. Furthermore, this effect was also maintained following stratification for municipal population size. In relation to gender, PBF had an effect on reducing suicide rates in women, but not in men. These findings support the hypothesis that a well established and with high coverage CCT
+programme have an effect on reducing deaths from intentional self-harm. Our study main finding, that high PBF coverage is associated with lower suicide rates, is in line with a previous finding, suggesting that a cash transfer programme in Indonesia reported a reduction of approximately 10% in suicide incidence in the sub-districts that implemented the programme [12].
+Cash transfer may attenuate the effects of poverty, by improving better mental health and consequently reducing suicide [22]. Several studies have suggested that poverty may influence the incidence of suicide [2, 10], with one study undertaken in Brazil reporting that both absolute poverty, measured by average per capita household income, and income inequality, measured by the Gini Index, are related to an increased suicide rates in Brazilian municipalities [2]. Greater investment in public welfare is associated with reduced suicide rates [6, 7], and the population’s confidence in social welfare policies exercise a protective effect on suicidal behaviour [6]. Cash transfer programmes not only reduce poverty, but has also been associated with other factors that influence suicide, such as a reduction in depression symptoms [22] and common mental health disorders [23], and in perceived hope and optimism [22] among beneficiaries.
+In Brazil, the PBF may influence suicide rate by fulfilling its main objectives of immediate poverty alleviation through the transfer of benefits to poor and extremely poor families and via investment in human capital through education and health conditionalities. Transferring money may provide greater financial stability, which helps to reduce the stress
+related to economics [24] and increased well-being [23, 24]. Regular income transfers may also support a reduction in those factors that may precipitate the occurrence of suicide, such as alcohol consumption and diagnosis of other mental health disorders [23-25]. Education-related conditionalities may act in a prospective manner, supporting increased schooling, increasing social empowerment and inclusion in the job market. There is evidence that low educational levels and unemployment are associated with increased risk of suicide [4]. In this way, programmes aimed at mitigating these factors, such as the PBF, may also contribute to reducing suicide rates. Furthermore, health conditionalities lead to increased access to health services (Fig. 1).
+Another important finding was that the observed reduction of suicide rates increased with the duration of high PBF coverage at the municipality level, indicating the importance of the continuity of social interventions to strengthen its effect [11]. Regarding gender, duration of PBF coverage at municipality level only reduced suicide rates among women, but not among men. Women are the principal beneficiaries of many social protection programmes, underpinned by the notion of the greater vulnerability of women in economic crises, and the fact that studies demonstrate that cash received by women tends to be invested in resources that promote family well-being [26]. In Brazil, PBF is preferentially awarded to women [13]. Benefits received by women may influence the family dynamic, strengthening female self-esteem and decision-making power [25]. In Brazil, qualitative evidence has shown that PBF may contribute to female empowerment, promoting greater autonomy and visibility for women in the community [27, 28] as legitimated representatives and family spokespeople [27]. An association between cash transfer programmes and reduced intimate partner violence suffered by women have also been reported [25, 29]. Cash transfer programmes have an impact on increased overall well-being and in the self-reported happiness of women [29].
+In our study, socio-economic factors—percentage of people with low levels of education and percentage on a low income—were associated with higher suicide rates in the Brazilian municipalities, supporting the findings that suicide rates may be reduced by anti-poverty polices, such as PBF, given that this aims to increase family income and to break the poverty cycle. Furthermore, higher employment rates had a protective relationship with suicide rates.
+As we used municipal level data, we evaluated the effect of social policy at an aggregate level and our inferences cannot be extrapolated to the individuals, under the risk of committing an ecological fallacy. However, we are aware of the possible spillover effects of BFP. Besides the effect on the target population receiving BF, the economic improvement in the municipalities given a high percentage of inhabitants receiving the benefit, possibly affect the overall
+economic situation of these municipalities [30]. The quality of SIM data is one possible limitation, due to potential under-recording of suicide data. However, a previous study showed that SIM data is of good quality for about 80% of Brazilian municipalities [31]. We also tested the robustness of the PBF effect on suicide rate restricting our analysis for municipalities with accurate vital information, and we found similar results.
+One strong feature of the study was its use of longitudinal, panel data analysis, than traditional cross-sectional data analysis, which enabled us to explore the influence of the PBF and the social and economic contextual characteristics in the same municipalities over time, strengthening the evidence for the relationship we found [20]. We also realized sensitivity analysis done gave much the same effect estimates, suggesting that our findings are robust. Suicide contributes to increasing the burden of mortality from causes that have been gaining prominence in Brazil’s mortality hierarchy and our findings also have the potential to produce robust evidence on the impact of poverty and social interventions on suicide rates, not limited to health care. Our results suggest that economic circumstances may be associated with suicide incidence in Brazil and that the implementation of conditional cash transfer programmes has the potential to decrease self-inflicted deaths, particularly in programmes which use conditionalities to break the poverty cycle. However, as our results cannot be extrapolated for the individual level, we suggest that new researches at the individual level test the association between cash transfer and suicide rates, to strengthen the evidence of the potential role of PBF in reducing suicide.
+Compliance with ethical standards
+Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.

Effectiveness of a Dutch community-based alcohol intervention Changes in alcohol use of adolescents after 1 and 5 years.txt

@@ -0,0 +1,53 @@
+1.	Introduction
+Underage drinking is a major public health problem in Western society. In the Netherlands, adolescent alcohol use ranks among the highest in Europe. At the age of 14, 39% of Dutch adolescents are recent drinkers, i.e., they had at least one drink in the month prior to investigation (Van Dorsselaer et al., 2010). A young age of onset is associated with a greater risk of alcohol abuse 10 years later (Behrendt et al., 2009). Moreover, there are several risks involved in drinking alcohol at an early age, such as unprotected sex, accidents and brain damage (Bonomo et al., 2001; Hingson et al., 2003a,b; Tapert et al., 2002). Therefore, from a public health viewpoint, prevention of alcohol use in young adolescents is crucial.
+Especially in the Dutch Achterhoek region, a rural area in the eastern part of the Netherlands, the prevalence of alcohol use among adolescents was high. Health monitors performed by the Community Health Service in 1997 and 2002 showed a negative trend in the Achterhoek: the age of onset became lower, adolescents drank more often and they drank more alcohol consumptions per occasion (De Rover et al., 1998, 2002, 2003). Drinking alcohol was part of the culture at that time, and drinking alcohol by adolescents was considered normal by the community. Therefore, in 2005, the local authorities and several local organisations decided to develop the community intervention “Alcohol moderation among adolescents in the Achterhoek”. The aim was to promote alcohol moderation among adolescents aged 10-19 years, in order to reduce the harmful effects. This has been the start of one of the first community-based interventions for alcohol reduction among adolescents in the Netherlands.
+The effect of the intervention on knowledge, attitude and social norm of parents has already been demonstrated (De Vlaming
+et al., 2008). Moreover, the intervention has been acknowledged by the Dutch Centre ofHealthy Living as theoretically well-founded (Database Healthy Living, 2015). However, until now, the effect of the community intervention “Alcohol moderation among adolescents in the Achterhoek” on the drinking behaviour of adolescents has not been examined.
+Worldwide, the scientific literature on community-based interventions for prevention and reduction of alcohol use among adolescents is relatively scarce and shows mixed results (Anderson et al., 2009; Bagnardi et al., 2010; Foxcroft et al., 2003; Giesbrecht, 2003; Hallgren and Andreasson, 2013). Evaluation studies of community-based interventions do face difficulties regarding the time frame and scientific standards. For example, community interventions are often initiated by local organisations instead of researchers, which reduces the influence of the researcher in creating a ‘controlled’ setting, and increases the risk ofbias. In addition, measuring long-term effects (i.e., 4 years or longer) is important since it takes a long time before community interventions are developed and implemented, and it takes even longer before changes in behaviour or health status can be demonstrated. To this end, it has been argued that more methodically sound research is required, measuring both short- and long-term effects.
+Therefore, the aim of this study was to evaluate the effectiveness ofthe communityintervention“Alcoholmoderation amongadoles-cents in the Dutch Achterhoek region” on alcohol use by adolescents in the second grade and fourth grade of Dutch high school. It was hypothesised that the community intervention would be superior to the reference condition in reducing the prevalence of recent drinking and binge drinking on the short and long term (1 and 5 years, respectively). Superiority was expected, in particular, in adolescents in the second grade compared to adolescents in the fourth grade of high school, as Dutch adolescents in the second grade are all underage, whereas adolescents in the fourth grade are a mixture of underage adolescents and adolescents who already reached the legal drinking age (16 years at that time). In addition, we performed stratified analyses for age, gender, educational level and ethnicity to gain more insight into possible sources of heterogeneity.
+2.	Methods
+2.1.	Design and data collection
+In order to evaluate the effectiveness of the community intervention “Alcohol moderation among adolescents in the Achterhoek”, a quasi-experimental (non-randomised) pretest posttest design was used, based on three independent cross-sectional surveys in the intervention and reference region. The change in adolescent alcohol use in the Achterhoek region (intervention region) was compared to the change in the Noord-Veluwe region (reference region) in the same period. The repeated cross-sectional surveys were part of the regular electronic health monitor system (E-MOVO), performed in October/November, 2003, 2007 and 2011 by the Dutch Community Health Service as described elsewhere (Croezen et al., 2009). Data were collected in the second and fourth grade of Dutch high schools using a detailed Internet questionnaire, under supervision of instructed teachers following a standardised protocol.
+The questionnaire contained approximately 100 standardised questions concerning social-demographic factors, school, health-status and lifestyle, including alcohol use (Dutch National Health Monitor, 2015). Ethnicity was measured by asking where the parents were born, in accordance with the definition of Statistics Netherlands (2015a). Educational level was measured as type of education that adolescents were following at the time of the survey and classified as low (VMBO) or high (HAVO/VWO). Recent alcohol use was measured by asking how many times adolescents had consumed an alcoholic beverage in the past four weeks, with 13 predefined response categories ranging from 0 times to 20 times or more. Recent binge drinking was measured by asking how many times adolescents had consumed 5 or more alcoholic beverages at one occasion in the past four weeks, with 7 predefined response categories ranging from never to 9 times or more. Self-report measures of adolescents on alcohol use are reliable and valid methods to measure alcohol use (Del Boca and Darkes, 2003), although they might underestimate heavy alcohol consumption (Northcote and Livingston, 2011). We had no data available on the onset of alcohol use.
+2.2.	Intervention “Alcohol moderation among adolescents in the Achterhoek”
+The Dutch community intervention “Alcohol moderation among adolescents in the Achterhoek” was one of the first large-scale, intensive and long-lasting interventions inthe Netherlands which aimed to stop the trend of increasing alcohol use in adolescents. This intervention has been described in detail elsewhere (Izeboud et al., 2008). In short, the community intervention was comprised of a range of activities in orderto promote alcohol moderation among adolescents aged 10-19 years, targeting their environment and adolescents themselves. Health education, regulation, and enforcement were integrated and implemented in multiple settings, i.e., homes, schools, sport clubs, youth work, bars and dance clubs. The intervention was developed and carried out by the eight municipalities in the Achterhoek region, the regional Addiction Service, the Police and the Public Prosecution Service, underthe guidance of the Community Health Service. The Community Health Service and the regional Addiction Service selected evidence-based programmes (such as “Alcohol: another story”) or developed intervention activities based on scientific knowledge in close collaboration with the National Institute of Mental Health and Addiction and local communities. Some examples of intervention activities are mass media campaign (radio broadcast, posters, TV commercials etcetera), parent-child evenings at school, regulations at schools and at sport clubs, instruction of barkeepers of community centres, sport clubs, bars and dance clubs, health education by the school nurse, cartoon battle at high schools and the “fine or chance card” for adolescents who were fined for an alcohol-related crime. Substantial attention was paid to preventing the onset of alcohol use under the age of 16, the legal drinking age at that time. Several prevention strategies were focused on raising awareness among parents on the relation between braindevelopment and alcohol use of adolescents, as well as parenting skills, e.g., rule setting. The implementation of intervention activities started in 2006 and, after two prolongations, ended in December, 2012. The aim of this study was to assess the overall impact of the combined interventions and not the effects of individual strategies. The primary target population consisted of approximately 37,000 adolescents aged 10-19 years living in the eight municipalities of the Achterhoek region in January, 2006 (Statistics Netherlands, 2015b).
+2.3.	Reference region
+The reference region was a rural area west of the intervention region, with enough distance to avoid contamination from the intervention region to the reference region (Fig. 1). In the reference region, which consisted of six municipalities, “regular policy” was continued throughout the study period. This also included the regular national Dutch alcohol legislation and policy of that time (2003-2011), including the development of local initiatives for alcohol prevention. We do not consider this as a threat tothe results of our study, as most alcohol initiatives inthe Netherlands hadasmallerscale, a lowerintensity andashortertime framethan our intervention “Alcohol moderation among adolescents in the Achterhoek”.
+2.4.	Analyses
+Our hypothesis was that the change in alcohol use of adolescents would be significantly larger in the intervention region compared to the reference region. In addition, we expected that the effect would be more prominent in the second grade than in the fourth grade. Therefore, all analyses were stratified by grade. Data were analysed using SPSS, version 21. Overall, the response to the repeated cross-sectional surveys was high. As shown by a response study performed in 2007, 82% of schools participated inthe surveys andwithin participating schools, 95% ofthe adolescents participated (Croezen et al., 2009). This resulted in an analytical sample of 5881, 5502 and 5920 adolescents in the intervention region and 3122, 3053 and 3211 adolescents in the reference region in 2003, 2007 and 2011 respectively. Missing data varied from 0 to 606 missings (1.5%) per variable and consequently subjects with missing data were not included in the analyses. Descriptive analyses per region were conducted to identify possible differences in gender, educational level and ethnicity. For the main analyses, ‘recent alcohol use’ was defined as at least one drinking occasion in the past four weeks and ‘recent binge drinking’ was defined as at least one drinking occasion with 5 or more alcoholic beverages in the past four weeks, in accordance with national standards (Dutch National Health Monitor, 2015). To this end,the scales were recoded intodichotomous variables 0 = ‘no recent alcohol use’ versus 1 = ‘recent alcohol use’ and 0 = ‘no recent binge drinking’ versus 1 = ‘recent binge drinking’.
+For the main analyses, we compared the change in alcohol use in the period 2003-2007 and 2003-2011 in the intervention region with the reference region. Linear regression was used to obtain (adjusted) percentages as the outcome. Although logistic regression is the common method for binary outcomes, we primarily applied linear regression to obtain (adjusted) effect estimates; this enhances straightforward interpretation and it has been argued that this is statistically appropriate for the limited range of percentages and effect estimates in our data (Hellevik, 2009). The model used to obtain (adjusted) effect estimates contained an indicator variable for intervention (I =1 for intervention region, I =0 forcontrol region) and time period (T with subscript for the year 2007 and 2011; 2003 served as reference). The covariates gender, educational level and ethnicity were added as potential confounders as indicated below. In this model, the intervention effect is estimated by the coefficient ^12 and $13 of the product terms region*year, for the short and long term effects,
+respectively:
+3. Results
+Y = $0 + $1 * I + $2 * T2007 + $3 * T2011 + $12 * I * T2007 + $13 * I * T2011
++ $4 * gender + $5 * educational level + $6 * ethnicity.
+Crude estimates were obtained using the model without covariates, and stratified analyses were done to obtain adjusted percentages forthe second and fourth grade separately. Adjusted estimates were obtained from the predicted values from the model, using the mean values of the confounders as predictors. Logistic analyses weredone insimilarways,usingthe product term region x yearforthe intervention effect and adjusting for the same variables. Additionally, to gain insight into possible sources of heterogeneity of effect estimates, the analyses were repeated using strata for age, gender, ethnicity and educational level, adjusting for confounders where appropriate.
+3.1.	Characteristics ofthe study population
+The sociodemographic characteristics of the study population are presented in Table 1. The mean age was slightly over 14 years (SD 1.2), with more than half of the sample in the second grade. In the fourth grade, 41% of adolescents was 16 years of age or older, the legal drinking age of that time. For all three survey years, age, gender, educational level and ethnicity were similar in the intervention and control group; although some of the differences were statistically significant, adjustments for these covariates did not substantially alter the effect estimates and these small
+differences do not raise serious concerns on confounding by these factors.
+3.2.	Effects on alcohol use
+Table 2 and Fig. 2 present the prevalence of recent alcohol use and binge drinking in the second grade in 2003, 2007 and 2011. Generally, a strong decline in alcohol use could be seen. Over the whole period of 2003-2011, the prevalence of recent alcohol use in the intervention region declined from more than 50% to less than 20% (crude percentages). After one year of intervention, the change in the adjusted prevalence of recent alcohol use was significantly stronger in the intervention region (-26%), compared to the reference region (-15%). On the long term, these results remained, but were not strengthened: after five years of intervention, the change in prevalence of recent alcohol use in the intervention region was -39%, which was significantly stronger, compared to the reference region (-30%). The same pattern was seen for recent binge drinking. After one year of intervention, the change in the adjusted prevalence of recent binge drinking was significantly stronger in the intervention region (-14%), compared to the reference region (-8%), and this effect remained until 2011 (albeit non-significant in the logistic analysis). In fact, the high prevalences of alcohol use and binge drinking, which were observed before the start of the community intervention “Alcohol moderation among adolescents in the Achterhoek” in 2003, were ‘normalised’ to the same level as the reference region by the year 2007 and further declined similarly to the reference region until 2011. However when looking at the fourth grade (Table 3 and Fig. 2), the change in the intervention region was not significantly different from the change in the reference region - both regions showed a substantial, but similar decline in recent alcohol use and binge drinking in the period 2003-2011.
+3.3.	Stratified analyses
+Figs. 3 and 4 show the effect estimates of the intervention on recent alcohol use and binge drinking stratified by several variables which are possible sources for heterogeneity. It is clear that the
+effect of the community intervention “Alcohol moderation among adolescents in the Achterhoek” is concentrated in the second grade; as mentioned above, in the fourth grade no significant effects can be observed. The picture for age is in line with this. A significant effect can be seen for 13- and 14-year-olds, which is consistent with the effect in the second grade. The picture for 15- and 16-year-olds is, on average, also consistent with the fourth grade: generally no effect can be observed, although the effect estimate for alcohol use of 15-year-olds on the short term seems somewhat increased. For ethnicity, gender and educational level the effect estimates in the strata are similar to the overall effect estimates.
+4. Discussion
+This quasi-experimental evaluation study provides evidence that the community intervention “Alcohol moderation among adolescents in the Achterhoek” was effective in reducing the alcohol use of adolescents in the second grade of Dutch high school. After one year of intervention, the decline in the prevalence of recent alcohol drinking and binge drinking was 11% (P<0.01) and 6% (P<0.01) stronger in the intervention region as compared to the reference region. This effect was restricted to the second grade and remained, but was not strengthened, after five years of intervention. No clear subgroup effects or confounding were observed for ethnicity, gender or educational level.
+During the study period, there was an overall decline in alcohol drinking. This decline in alcohol use over the past years is a well-known phenomenon in the Netherlands. National data of 12- to 18-year-olds, gathered in similar ways as our data and including similar outcome variables, also show a decline in recent alcohol use: from 58% in 2003 to 51% in 2007 to 43% in 2011 (Van Laar et al., 2011; Verdurmen et al., 2012). This is comparable to the decline in our reference region: from 60% in 2003 to 48% in 2007 to 37% in 2011 (crude percentages of the second and fourth grade combined). The decline of alcohol use in the Netherlands in the past decade can be attributed to the Dutch policy towards adolescent drinking. In general, the policy has been lenient and it is only since 2006 has adolescent drinking become an issue on the public health agenda
+in the Netherlands, and has national policy become more stringent. Our community intervention was initiated as one of the first large initiatives and with our quasi-experimental evaluation study we were able to demonstrate an effect on top of the national trend.
+Most evaluation studies on alcohol prevention and reduction describe family- and school-based interventions rather than community-based interventions. A meta-analysis of family interventions on alcohol initiation and frequency of alcohol use (Smit et al., 2008) suggests family interventions to be effective in reducing adolescent alcohol consumption. However, just three of the studies in the meta-analysis reported the long-term effect of the intervention and all studies were conducted in the US. More recently, Foxcroft and Tsertsvadze (2012) reviewed multicomponent alcohol interventions, defined as interventions conducted in multiple settings, for example in both school and family settings. Out of 20 multicomponent alcohol interventions, 12 were effective in preventing alcohol abuse in young people, up to three years of follow-up. Also the majority of these studies (17 out of 20) were conducted in the US. In the Netherlands, the Preventing heavy Alcohol use in Adolescents (PAS) study showed that the combined school- and family-based intervention reduced the likelihood of onset of weekly drinking and the frequency of drinking after 10 and 22 months (Koning et al., 2009). Generally, community-based alcohol interventions tend to be scarce and less well described. Two examples of relatively well described community interventions
+are Project Northland, one of the first community interventions which was effective in reducing alcohol use of adolescents (Perry et al., 1996, 2002) and the Italian ‘Alcohol, less is better’ community project (Bagnardi et al., 2010), which found a reduction of 1-2 drinks per week in the intervention communities compared to the control communities after 2.5 years of intervention activities. Our study is a promising supplement to the current, still modest, evidence on community-based alcohol interventions.
+The majority of studies, including the Dutch PAS, found that alcohol interventions are primarily effective in underage adolescents. This is partly in line with our results. As hypothesised, our stratified analysis showed that the effect of the intervention was clearly concentrated in the second grade (13 and 14 year olds) and there was no effect in the fourth grade (15 and 16 year olds). Possibly, underage adolescents in the fourth grade were subject to peer pressure of classmates who had already reached the legal drinking age of that time (16 years). Peer pressure is a well-known factor influencing alcohol use of adolescents (Bot et al., 2005).
+In our study, the 1-year effect of the community intervention remained, but was not strengthened, after 5 years. Although in theory a stronger effect could be expected after a longer follow-up time (see also Section 1), the literature shows mixed results. In the above mentioned meta-analysis of Smit et al. (2008), a stronger effect was found after a longer follow-up time. However the opposite has also been reported, e.g., the Dutch PAS study found that the effect on
+one of their outcome measures (heavy drinking) disappeared on the longer term (Koning et al., 2009). More research is needed to understand the mechanisms that lead to stability, strengthening or reduction of the effects of alcohol interventions on the longer term.
+There are some limitations of our study which should be mentioned. Firstly, the use of existing data of the Dutch regular electronic health monitor system caused a time gap between the baseline measurements in 2003 and the start of the intervention in 2006. Because of this, other factors may have influenced the measurements in 2007, however this is unlikely to be different in the intervention and reference region. Another disadvantageous implication of using existing monitor data is that we did not have all relevant variables available such as alcohol related harms (Hallgren et al., 2012) and the frequency and quantity of specific beverages. The latter is a more specific measure to assess changes in the estimated volume of alcohol consumed and might have been more adequate for demonstrating important changes in the distribution of drinking.
+Secondly, our quasi-experimental study design lacks randomisation at the individual level and the community intervention could not be cluster-randomised over more regions. Therefore we are not completely sure that the effects found in this study are due to the intervention, and not to differences (inequivalence) between the intervention and reference region. Various region-level factors of environmental or cultural nature may influence trends in alcohol consumption. However to date, it is well known that randomisation is often unfeasible for community interventions and therefore the use of quasi-experimental designs has been advocated (DesJarlais et al., 2004; Victora et al., 2004). Moreover in our study, the baseline characteristics of the intervention group and the reference group were very similar and appeared not to be strong predictors or modifiers, therefore we do not expect that bias due to inequivalence has occurred.
+Thirdly, a multi-level design was not applied since adolescents in this region are, besides school, part of many other settings where alcohol is consumed i.e. homes, sports, night life and youth work. However by treating each pupil as an independent observation unconnected to the class and school environment, we may have underestimated the standard errors to the estimates.
+Fourthly, it is a limitation that our study design did not allow to disentangle the effect of the individual components of the community intervention. This makes it difficult to clarify why the intervention worked and which mixture of intervention activities was most effective.
+Finally, there are some potential limitations to the generalizability of our results. The intervention effects as observed in this study may depend on region- or country-specific characteristics such as a highly tolerant drinking culture or a lenient policy towards adolescent alcohol consumption. Therefore, it is unclear to which extent the results would hold for other regions or countries.
+There are also some important strengths to our study. Firstly, the use of a reference region made it possible to isolate the effects of the intervention from other influences in the time-period. The selected Noord-Veluwe region is a reliable reference, since ‘regular policy’ was provided throughout the study period, and the geographical distance to the intervention region was large enough to prevent contamination.
+Secondly, the main programme strategies were theoretically well founded; for example, the integration of health education, regulation and enforcement (Alcohol and Public Health Policy Group, 2010), the implementation in multiple settings (Foxcroft and Tsertsvadze, 2012) and the focus on adolescents as well as their environment. Especially strategies for parents were considered important and included knowledge transfer, raising awareness and increasing parenting skills (Van der Vorst et al., 2006; Smit et al., 2008). Although such evidence was scant at the start of
+our community intervention (2005), several publications during the past decade have demonstrated the effectiveness of these strategies.
+Thirdly, the high response of adolescents in the repeated cross-sectional surveys is a strength, since it yielded a high and representative number of cases for our study.
+Fourthly, we measured effects on the short term as well as the long term, after five years of intervention. This is much longer than the time frame of most studies.
+Finally, our study was part of a comprehensive evaluation of the community intervention “Alcohol moderation among adolescents in the Achterhoek”, which also included an extensive process evaluation (Database Healthy Living, 2015) and an effect evaluation among parents of adolescents (De Vlaming et al., 2008). Our results fit well into the greater picture of these evaluations.
+Some processes might have facilitated the favourable outcomes. One of these is the joint decision making between health promoters and local communities. Substantial effort was put in building relationships, lobbying and explaining the new scientific knowledge to the local communities. Although the health promoters initiated most plans, local communities were involved at an early stage, and they implemented and financed a large part of the community intervention.
+Our community-based intervention contributes to the growing body of evidence on community efforts aiming at reducing alcohol use. The evidence-base of community approaches for alcohol reduction has been debated in the past (Anderson et al., 2009). However, our study provided evidence that the prevalence of alcohol use and binge drinking can be substantially lowered by such efforts, even in communities where drinking alcohol at a young age is part of the culture and is considered normal. This broadens the perspective to community approaches, i.e., organised bottom-up by the initiative of local authorities or other local organisations, and combining the strategies of health education, regulation and enforcement. As suggested elsewhere, alcohol policy seems to be most effective on behavioural change when the three approaches are mixed and combined integrally (Alcohol and Public Health Policy Group, 2010). Especially in environments where drinking alcohol is the norm, a broad and integrated approach is important in order to be able to turn the tide. Therefore, we think that our results are of great importance for policymakers and local organisations who want to reduce alcohol use of adolescents in an effective and efficient manner.

Effects of Housing First approaches on health and well-being of adults who.txt

@@ -0,0 +1,87 @@
+BACKGROUND
+Access to housing is an important determinant of health, with homeless people having substantially increased morbidity and mortality compared with the housed population.1 2 For instance, a recent systematic review found that all-cause mortality in homeless populations in high-income countries is between 3 and 11 times higher than their housed counterparts.2 This excess mortality appears to persist even after accounting for socioeconomic
+deprivation and comorbidity.3 Homelessness may have a direct impact on health, through the physical and psychosocial hazards associated with rough sleeping or temporary accommodation (such as excessive cold, heat or damp; physical and sexual violence and other forms of crime); lack of basic amenities and social goods (such as washing facilities); stigma and social isolation and difficulties in accessing healthcare services.1 5 It is also strongly associated with other experiences deleterious to health, such as poverty (especially child poverty), adverse childhood experiences and substance misuse.6 7 The association between homelessness and health is also bidirectional, since poor physical or mental health can increase the risk of unemployment, relationship breakdown and housing loss.8
+Homelessness is increasing across Europe.10 Recent increases in homelessness may be linked to economic trends, cuts to public services and welfare benefits and changes in the availability and affordability of housing.11 Rehousing homeless (roofless or houseless) persons, or persons at risk of homelessness (insecure housing),12 may therefore be an important health intervention.13-15 One approach to increasing housing stability is Housing First (HF).
+HF is defined in contrast to the traditional ‘Treatment First’ model, which provides temporary accommodation alongside services to address health needs, particularly substance use. The client then progresses to transitional housing before achieving permanent housing; this is conditional on adherence to treatment for mental health and problematic substance use.16 17 The ‘HF’ approach aims to assist clients to access permanent housing as an initial step in addressing homelessness. Housing provision is not contingent on compliance with health treatment or substance abstinence. Additionally, HF includes ongoing support, through Intensive Case Management or Assertive Community Treatment approaches.17 18 There are a number of established HF projects in North America and Scandinavia, and governments in other countries, including France and the UK, have shown interest in rolling out the model.19-24
+HF may improve health, via the mediating factors of increased housing stability and access to support services (figure 1). However, critics have suggested that HF may adversely affect health, since engagement with health services is not compulsory
+and, it is argued, there is therefore a lack of incentive to adhere to treatment or abstain from problematic substance use.
+Although prior literature reviews of the impacts of HF have been conducted,26-29 these reviews did not meet the reproducibility standards of a systematic review and did not undertake meta-analysis. Moreover, new data on the health impacts of HF are now available. This paper reports the findings of a systematic review of the health effects of the housing provision aspect of HF. The review addresses a current gap in the literature by using a clear definition of the intervention, including recent studies, and conducting the first meta-analyses of health outcome data.
+METHODS
+We constructed an initial logic model linking HF to health from relevant literature sources (figure 1).17 25 We then systematically reviewed evidence of the health effects of HF to test the hypothesis that rapid provision of permanent, non-abstinence-contingent housing to homeless people, leads to health improvement in this vulnerable population compared with housing provision without these features.
+The scope, inclusion criteria and methods of the review are outlined below and in box 1. The review protocol was registered on the PROSPERO database.30 The intervention was defined in this review as ‘rapid provision of permanent, non-abstinence-contingent housing’. The inclusion of additional supports (such as Intensive Case Management or Assertive Community Treatment) was not used to define the intervention here, as our aim focused on housing. We had intended to compare interventions adhering with the wider principles of HF with interventions providing only housing; however, all studies found included some form of additional support, so this subgroup analysis was not possible. Given all interventions included both rapid provision of permanent, non-contingent housing and additional support, they are therefore labelled ‘HF’, whether or not they were identified as such in the literature.
+We restricted study types to randomised controlled trials (RCTs), to minimise risk of bias and allow synthesis of data from directly comparable studies. Given a number of RCTs were known to have been conducted, we focused on these as the best available evidence. Primary outcomes were quantitative measures of health, well-being and quality of life; a secondary outcome was housing stability.
+Search strategy
+The search strategy was developed in collaboration with a University of Glasgow librarian. The following databases were searched: EMBASE, MEDLINE, PubMed, PsycINFO, Cochrane Central Register of Controlled Trials (CENTRAL), Social Sciences Citation Index and Biosis. Databases were searched using Homeless Persons, Housing and Public Housing as MeSH terms, alongside keywords homeless*, housing and ‘housing first’. Filters were used to select RCTs.31 32 The full search strategy for each database is found in online supplementary file 1.
+Searches were restricted to studies published from 1992 (when Pathways to Housing was founded and the intervention first initiated) up to the date of the search (15 May 2017) in peer-reviewed journals. Reference lists of previous reviews were checked for additional studies.
+Screening and selection of studies
+Only studies published in English in peer-reviewed journals were included. Only studies which reported a primary health outcome (box 1) were included. Search results were screened by title by one reviewer (AJB) to remove obviously irrelevant citations. Abstracts and full texts were screened independently by two reviewers (AJB and ET). Any discrepancies were resolved by consensus.
+Data extraction and risk of bias assessment synthesis
+Data on key study characteristics, intervention details and reported outcome data were extracted by one reviewer (AJB) and checked by a second (ET). Outcome measures from studies were grouped by domain: mental health; quality of life; substance use; non-routine use of healthcare services; housing stability and other health-related outcomes.
+To avoid double counting of data, where sampling overlap was stated or suspected for any single outcome or where findings were reported in multiple papers, data were selected to prioritise larger combined samples or allow calculation of standardised effect estimates for comparison with other papers.33
+The Cochrane Risk of Bias Tool V2.0,34 was used by one researcher (AJB) to assess potential bias for each of the outcomes, and checked by a second (ET). If high risk of bias was reported in at least one domain of bias for an outcome, the outcome was given an overall ‘high’ rating.
+Population: adults (16 years and older) who meet at least one of the European Typology for Homelessness and Housing Exclusion (ETHOS) criteria: roofless, houseless, living in insecure housing, living in inadequate housing.
+Intervention: providing the homeless person with access to housing through:
+►	Assistance in locating and entering housing.
+►	Subsistence of rental costs to maintain permanent tenancy.
+The housing provided was defined as:
+►	Intended to be permanent—no intention by providers to end or transfer tenancy, counting sustained tenancy as the intended outcome.
+►	Not contingent on adherence to treatment or substance abstinence.
+►	Rapid, with the process of securing and entering housing initiated at first contact with the homeless person and with the aim of beginning tenancy promptly.
+Comparators: treatment as usual groups; although we note that this includes many diverse alternative homeless services and interventions.
+Outcomes: the primary outcomes, chosen to reflect the aim and research questions, were quantitative measures of health and well-being. These were grouped into five domains:
+►	Mental health—including self-reported mental health and clinical assessment of mental ill health.
+►	Self-reported health and quality of life—questionnaires and interviews recording perspectives.
+►	Substance use—including self-reported occasions of substance use and self-reported problematic substance use.
+►	Non-routine use of healthcare services—including episodes of hospitalisation and use of emergency services.
+►	Other, unanticipated measures of health and aspects of well-being associated with health and mental health.
+Secondary outcome: housing stability. This included any measure of housing which reflected the stated goals of the intervention of ending homelessness. The use of this domain in the review was based both on the hypothesised causative mechanism leading to changes in health and also its expected availability in almost all studies.
+Study design: randomised controlled trials
+Data synthesis
+Calculations of standardised effect sizes were conducted manually in Microsoft Excel.35 Standardised mean differences were calculated to compare continuous variables. These were interpreted as ‘small’ to ‘large’ effect size using Cohen’s classification.36. Incidence rate ratios were calculated for counts of use of health services, a risk ratio for attaining stable housing and ratios of rate ratios for the substance use subgroup outcome. Where effect sizes were reported only by subgroups and not the whole trial population, data were pooled where possible, otherwise subgroups were presented separately in forest plots.
+Forest plots were used to present standardised effect estimates for each outcome domain using Review Manager V5.3.37 A random effects model was used to calculate pooled effect size estimates, 95% CIs and heterogeneity, as we assumed that effect sizes and variation would differ across studies. Where meta-anal-ysis was not possible these were reported narratively in the relevant domain.
+Findings were summarised using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidance to assess certainty of results for each meta-analysed outcome.38-42
+RESULTS
+Searching returned 494 records after removal of duplicates (figure 2). Following full-text screening, 25 eligible papers were identified for inclusion; these papers report results from four studies, all based in Canada and the USA (see online supplementary file 2 for included papers and online supplementary file 3 for exclusions).
+The four studies included in this review are outlined in table 1. The context and ‘treatment as usual’ provision varied across the cities and nations represented in these studies but were not always clearly and fully reported. All participants were homeless or insecurely housed; inadequate housing was not included in the studies retrieved. Beyond the inclusion criteria, there was
+some variation in the implementation of the HF model. All studies reported a measure of housing stability alongside one or more primary outcome measures. All results are summarised in table 2.
+Risk of bias
+The overall risk of bias was assessed as high for each outcome reported across all four studies (see online supplementary file 4 for all domains and table 2 for overall rating). Bias due to missing outcome data was rated as high if there were no data to
+evaluate how effectively the effect of loss to follow-up had been addressed.
+Primary outcomes
+Mental health
+All four studies reported mental health outcomes; these were categorised as ‘self-rated mental health’ (n=3: At Home, Chicago Housing for Health Partnership (CHHP) and Housing Opportunities for Persons with AIDS (HOPWA)) and ‘severity of mental health symptoms’ (n=3: At Home, Pathways Housing First (PHF) and HOPWA). Two studies provided data eligible for meta-anal-ysis of self-rated mental health (At Home and CHHP)43 44; a very small improvement was seen in intervention groups compared with treatment as usual (SMD = 0.07; 95% CI -0.19 to 0.33; p = 0.60, I2=82%; figure 3A). Additionally, HOPWA reported
+no statistically significant difference between groups.45 Both groups saw improvements in all studies.43-45 A small improvement in mental health symptom severity at 24 months in the At Home study was reported (SMD=-0.05; 95% CI -0.31 to 0.22; p = 0.73; I2 = 82%).46 47 Pathways HF participants saw no significant differences between groups in symptoms over 24 months (F=0.348; p=0.85; no effect direction reported).48 Improvements were seen in both intervention and TAU groups of the HOPWA study in depression and perceived stress, with no statistically significant differences between the two conditions.45
+Self-reported health and quality of life
+Several measures were reported in the domain of self-reported health and quality of life. Self-rated physical health was reported in three studies (At Home, CHHP and HOPWA).43-45
+Meta-analysis of two studies showed no detectable difference (SMD = 0.00; 95% CI -0.09 to 0.09; p = 0.94; I2 = 0%; figure 3B). Participants in both intervention and TAU groups of the HOPWA study reported improvements in self-rated physical health, with no statistically significant difference between groups.45 Two measures of quality of life were found in the At Home/Chez Soi study, but not repeated elsewhere. Pooling the two age group subgroups showed a small difference in mean change of generic quality of life between treatment and control groups from baseline, favouring TAU (SMD=-0.03; 95% CI -0.13 to 0.06; p = 0.50; I2 = 0%) and a small difference in condition-specific quality of life, favouring intervention (SMD = 0.18; 95% CI -0.09 to 0.46; p=0.19; I2 = 83%; not shown).43
+Substance use
+Two studies reported substance use outcomes (At Home and PHF).43 46-49 Data from PHF were reported as showing no significant differences in either alcohol or drug use at 24 months, but no direction of effect was indicated and so these could not be used in meta-analysis.48 Across 48 months, a greater reduction of heavy alcohol use (defined as using alcohol on >28 days in 6 months) in intervention groups compared with control is reported in the study by Padgett et al,49 with no clear difference in drug use. Pooling the two age group subgroups of the At Home/Chez Soi study showed a very small overall difference in self-reported problematic substance use, favouring HF (ratio of rate ratios = 0.96; 95% CI 0.72 to 1.28; p = 0.77; I2=61%; not shown)43; both groups saw decreases in reported problems.46 47
+Health service use
+All studies reported a measure of health service use. In meta-anal-ysis (n=2: CHHP and HOPWA), intervention participants experienced fewer hospitalisations (incidence rate ratio (IRR) = 0.76; 95% CI 0.70 to 0.83; p<0.00001; I2 = 0%; figure 3C).44 45 A small difference was seen in time spent hospitalised, also favouring intervention (n=3: At Home, CHHP and PHF; SMD = -0.14; 95% CI -0.41 to 0.14; p = 0.32; I2 = 83%; figure 3D).8 44 47
+A greater reduction was seen in intervention groups over control groups in number of emergency department visits (n=2: At Home and CHHP; IRR=0.63; 95% CI 0.48 to 0.82; p = 0.0006; I2 = 95%; figure 3E).44 46 HOPWA participants saw no significant difference between intervention and control groups in likelihood of one or more emergency department visit in each of three 6-month time periods (F = 0.63; p = 0.5977),45 and the CHHP intervention group saw a small reduction in likelihood of one or more emergency department visit in the 18-month period over control (risk ratio (RR) = 0.92; 95% CI 0.81 to 1.04).44
+Housing stability
+All four studies reported measures of housing stability, either recording a proportion of total days reported as ‘stably housed’ or a proportion of the population in stable housing at the end of the trial period. In all four studies, the intervention group was found to have large increases in housing stability over TAU.43-48 The combined effects estimate indicated that participants receiving HF are two and a half times more likely to be stably housed after 18-24 months (n=3: At Home, CHHP and HOPWA; RR=2.46; 95% CI 1.58 to 3.84; p<0.00001; I2 = 94%). A large standardised mean difference for time spent housed during trial was also seen, favouring intervention (n=2: At Home and PHF; SMD = 1.24; 95% CI 0.86 to 1.62; p<0.0001; I2=90%; see online supplementary file 4).
+Subgroups reported
+Subgroup comparisons were only conducted in the At Home/ Chez Soi study (see online supplementary file 4). In age comparisons, the older group (aged >50 years) had better outcomes than the younger group (18-49 years old) in a number of areas, such as self-rated mental health, mental health symptom severity, substance use and quality of life.43 50 Participants with less severe mental health and problematic substance use experienced slightly better outcomes.46 47 Participants housed together in dedicated accommodation blocks (referred to as the ‘congregate model’) experienced greater improvements than those in ‘scattered site’ housing, in mental health, quality of life and problematic substance use, among other outcomes.51 Across all subgroups reported, intervention participants saw large increases in housing stability.
+Other outcomes
+Several further outcomes that were related to health were recorded. These are listed in online supplementary file 5. Several small, uncertain effect sizes were observed, favouring HF in most cases, with two of the At Home subgroups experiencing small, uncertain effects favouring treatment as usual.43 51 Two studies reported HIV survival and viral load but the findings were conflicting.45 52
+DISCUSSION
+Summary of findings
+Our systematic review found that HF resulted in large improvements in housing stability; with unclear short-term impact on health and well-being outcomes. For mental health, quality of life and substance use, no clear differences were seen when compared with TAU. HF participants showed a clear reduction in non-routine use of healthcare services, over TAU. This may be an indicator of improvements in health.
+Comparison with existing literature
+The combination of a strong, positive impact on housing with little additional impact on mental health and substance use, compared with TAU, is consistent with the findings of other reviews.26-29 Our meta-analyses provide a clear picture of improvements in hospitalisation and emergency department visits, which has not yet been reported in other reviews. Inclusion of only RCTs gives greater confidence that these results are less susceptible to bias. Previous reviews have questioned whether abstinence-contingent housing may lead to greater reductions in problem substance use than HF, although at the cost of housing stability.28 However, our results found reductions in problem substance use for both HF and TAU, with no clear difference between them. This is consistent with non-randomised observational evidence suggesting greater effectiveness of HF than TAU in this respect.16 53
+Prior research on HF suggests that the consumer choice framework allows homeless clients greater perceived control, security and mastery of circumstances, leading to greater improvements in mental health and quality of life.54 55 A lack of clear difference seen across the RCTs analysed here may be due to several factors, including the heterogeneity of sample participants, differences in provision of attached services, differences in application of consumer choice and the relatively short-term observation period.
+Strengths and limitations of this review
+Our systematic review has several strengths. We conducted a comprehensive search across several databases, which aimed to include all of
+Review
+the relevant studies. The strict use of a clearly predefined protocol, with explicit inclusion and exclusion criteria, has allowed us to bring together all relevant evidence in a transparent manner. This includes drawing on theoretical understanding to define a clearly identifiable and replicable intervention. The use of the logic model allowed testing of the theoretical impact of HF on health through housing stability as a mediator.
+This systematic review had some limitations. The scope of this review was primarily limited by the focus on quantitative data from RCTs, and the largest study, a trial of At Home/Chez Soi, carried substantial weight and was the main determinant of effect estimates in rate of emergency department visits, and time spent stably housed. Although trials are underway elsewhere (eg, the Un Chez-Soi d’abord study in France56 57), the data included in this review were exclusively from North America and the participants were all selected on the basis of complex health needs (such as mental illness, substance abuse or chronic physical illness) as per the principles of HF.16 17 This may limit the generalisability of our findings internationally, as well as to homeless people without complex health needs. Other published data from non-randomised studies are available and may provide further insights into health outcomes, but these studies are at a higher risk of bias. Future qualitative enquiry to identify mechanisms associated with changes in health outcomes could help optimise the benefits of HF.
+Across all studies there were high ratings of risk of bias in several areas. Available data were limited to a 24-month follow-up period, providing observations of only short-term outcomes (figure 1). The uncertainty of effect size and direction of the primary health outcomes prevents accurate testing of the hypothesised intermediate and long-term effects of housing stability.
+A further systematic review, comparing HF with other interventions, for example, abstinence-contingent housing, housing vouchers, residential treatment and case management (without housing), was published after the completion of this review. This did not consider health outcomes but reported similar results for housing stability.58
+Implications for research and implementation
+Further questions are prompted by this review which could be addressed by ongoing evaluation of the HF model. Clear reporting of the intervention characteristics (for primary research) and inclusion criteria (for systematic reviews) should be a starting point in future research to ensure testing of an identifiable and replicable model.59 Further observations of longer follow-up periods would give greater confidence of impacts on long-term health.
+The subgroup analyses of the At Home/Chez Soi study showed several differences in effects for different age groups and health needs. It is unclear if these findings reflect genuine differences60; further research would be required to determine if there is greater effectiveness of the intervention for particular groups of homeless persons.
+To address some of these concerns, a further systematic review could synthesise the wider evidence base and allow generation of hypotheses about explanations for heterogeneity in reported effects. These data could then be used to refine aspects of HF with the aim of optimising potential beneficial impacts of HF investment. Evaluation of the relative contribution of key principles of HF to its effectiveness would be an important next step. In addition, a clearer differentiation and comparison of the treatments broadly grouped under treatment as usual in this review could show whether better interventions exist for certain groups.
+This review adds strength to the calls to adopt HF as an ‘evidence-based’ housing model, having shown consistent improvements in the housing stability of vulnerable homeless persons. Concerns that HF could result in higher rates of problematic substance use than treatment as usual are contradicted by these data. Alongside this, HF could reduce use of non-rou-tine health services, with potential cost savings. Subgroup analysis, although only reported in one study, suggests that housing stability is improved regardless of the age or health needs of the clients, while improvements in health might be differentially seen across groups. According to the logic model in figure 1, the improvements in housing stability associated with HF might be expected to result in intermediate and long-term positive impacts on these and other health outcomes, beyond the timescales considered in this review.
+CONCLUSION
+HF approaches appear to be highly effective in reducing homelessness among vulnerable participants. However, in several direct measurements of short-term health outcomes, the impact of HF is not clear. HF can be seen to reduce non-routine use of healthcare services, which may be an indicator of better health outcomes. Further evidence could be valuable in assessing the long-term effects of improved housing stability on health. HF could be implemented with strong confidence in its success as a housing intervention, alongside some confidence in a lack of immediate adverse effects on health, but with caution in relying on this model for certainty in improved health outcomes.

Effects of media stories of hope and recovery on suicidal.txt

@@ -0,0 +1,141 @@
+Introduction
+The way that media reports about suicide has received considerable attention in suicide prevention research over the past 5 decades. Most of the research has focused on harmful media impacts—the Werther or imitation effect.1,2 The Werther effect refers to the relationship between sensationalist or repetitive reporting of suicide and subsequent increases in suicide in the population. There is now strong empirical support for the Werther effect. A 2020 systematic review and meta-analysis found
+that media reports about celebrity suicides were associated with an 8-18% increase in suicides within 2 months of the reporting.3 This highlights that media reporting of suicide is a powerful environmental exposure that can have an impact on suicides. But importantly, poor reporting of suicide is potentially amenable to intervention through the implementation of media guidelines. For this reason, recommendations about news reporting on suicide are now a standard component of national suicide prevention strategies.4
+Research in context
+Evidence before this study
+Three systematic reviews and meta-analyses on harmful media reporting about suicides and subsequent suicides in the general population have been published (the Werther effect). One was published in 2005, another in 2012, and the third one in 2020. The most recent meta-analysis found that media reporting of celebrity suicides was associated with an 8-18% increase in suicides in the population in the following 1-2 months.
+The increase was even stronger for the same suicide method as reported in the media (18-44%). No systematic reviews or meta-analyses have assessed the effects of media messages designed to be protective against suicidal behaviour (the Papageno effect). We searched PubMed, Embase, PsycInfo, Scopus, Web of Science from inception up to Sept 6, 2021, using the search terms ((suicid* OR self-harm OR help-seeking) AND (Werther* OR Papageno* OR ripple* OR copycat OR imitat* OR contagio* OR suggesti* OR lived-experience) AND (media OR newspaper* OR print OR press OR radio* OR televis* OR film* OR movie OR book* OR documentar* OR internet OR cyber* OR web* OR music* OR drama* OR message* OR news* OR announcement* OR video OR broadcast* OR song* OR play* OR theat* OR story* OR narrative)). No language restrictions were applied. The first study investigating protective media information—an observational study—was published in 2010. This study showed a small negative association between protective media reports and suicide. Eight subsequent studies tested different types of protective interventions using a randomised controlled trial design. The interventions that were tested included media stories featuring individuals mastering their suicidal crises, and stories featuring prevention experts and peers of suicidal individuals speaking about suicide prevention and help seeking. All the eight studies included some baseline indicator of vulnerability of participants, and four of them
+explicitly examined protective effects among people with some degree of vulnerability to suicide. The individual studies showed differing results—four showed no statistically significant changes in suicidal ideation after the exposure to a narrative of hope and recovery whereas another four showed some beneficial effect. Among six studies assessing help-seeking attitudes, one reported a statistically significant beneficial effect while five reported no changes. There are no existing evidence syntheses on this topic.
+Added value of this study
+To our knowledge, this is the first systematic review and metaanalysis to pool together studies examining the association between protective media and suicidal behaviour. Our pooled analysis showed evidence that exposure to narratives of hope and recovery resulted in a small, but statistically significant reduction in mean suicidal ideation when compared with active controls. There was insufficient evidence to conclude that narratives of hope and recovery were associated with differences in help-seeking attitudes and intentions between groups.
+Implications of all the available evidence
+To our knowledge, this is the first study providing combined evidence across published trials that media narratives of hope and recovery from a suicidal crisis have a small protective effect on suicidal ideation on the short term (up to 4 weeks after exposure) among individuals with some degree of vulnerability to suicidal ideation or behaviour. These narratives pose little risk in populations with some degree of vulnerability to suicide. Large-scale trials to test the efficacy of these types of interventions on suicidal and help-seeking behaviours are needed.
+Although the harmful effects of media are documented, less is known about protective effects. In the past 10 years, researchers have begun to investigate how media exposure might be used to reduce suicide risk.5-10 If some specific media portrayals serve as a basis for subsequent suicidal behaviours, as indicated by the Werther effect, other narratives, particularly those featuring individuals who tell stories of overcoming suicidal crises without engaging in suicidal behaviour might reduce suicidal behaviours by decreasing the risk of acting out suicidal thoughts. Findings from the first study6 in the topic area suggested that news items featuring individuals who survived suicidal crises were associated with a small decrease in suicides. This suicide-protective media effect has been termed the Papageno effect.6,7
+Nearly all the studies of the Papageno effect have been randomised controlled trials (RCTs), which differ from ecological research designs used to study the Werther effect and provide a stronger basis for making causal inferences. These studies have frequently used ideation
+as the main endpoint of interest, which is a pragmatic endpoint for trials especially for assessing the safety of suicide-related messaging.8,10-14 A further important endpoint in these studies has been help-seeking attitudes and intentions.5 Some of these studies suggest that stories of hope and recovery have the strongest effect on individuals with some degree of vulnerability.10,12 This might be due to stronger identification with the stories, or higher perceived usefulness among individuals with personal experience of suicidality or suicide attempts.15 Effects on individuals with some degree of vulnerability are of particular importance to suicide prevention. Stories of suicide have been shown to be particularly harmful among those with some vulnerability,16 which raises the question of whether narratives of hope have similar harmful effects. Considering the small samples analysed in previous studies, meta-analytic approaches, particularly analyses using individual participant data (IPD), can answer the question about harmfulness in vulnerable audiences.
+In this systematic review and IPD meta-analysis,17 we combined participant-level data from RCTs on the Papageno effect to quantify the effect of personal stories of mastery of suicidal crises on individuals with some degree of vulnerability. We focused on two endpoints which have received the most attention: first, suicidal ideation (the primary outcome); and second, helpseeking attitudes and intentions (the secondary outcome). In two sensitivity analyses, we assessed whether findings for vulnerable individuals were generalisable to the entire intervention groups including people with low vulnerability, and whether findings were generalisable to narratives featuring peers and professionals rather than individuals speaking about their own recovery process.
+Methods
+Search strategy and selection criteria
+In this analysis, we searched PubMed (including MEDLINE), Scopus, Embase, PsycInfo, and Web of Science using the following search terms ((suicid* OR self-harm OR help-seeking) AND (Werther* OR Papageno* OR ripple* OR copycat OR imitat* OR contagio* OR suggesti* OR lived-experience) AND (media OR newspaper* OR print OR press OR radio* OR televis* OR film* OR movie OR book* OR documentar* OR internet OR cyber* OR web* OR music* OR drama* OR message* OR news* OR announcement* OR video OR broadcast* OR song* OR play* OR theat* OR story* OR narrative)). No language restrictions were applied. We searched titles, abstracts, and keywords from inception until Sept 6, 2021, for each database. The search was intentionally broad to capture all related studies. Google Scholar was used to identify grey literature, using the search terms “suicide and media”. Furthermore, we checked the Canadian Agency’s for Drugs and Technologies in Health (CADTH) Grey Matters guidance for a comprehensive list of clinical trials registries to identify any unpublished or further trials of interest.18 We also screened the reference lists of identified reviews and recent editorials or comments for further references. Finally, we searched the reference lists and did a cited-reference search for all included studies using Google Scholar. For a detailed overview of the search strategy see the appendix (p 3).
+Titles and abstracts were screened independently by two authors (TN and SK) using Mendeley. Eligible studies were then selected by full-text articles review by the same authors. At both stages of screening, disagreements were resolved by consensus.
+Our eligibility criteria were framed using the Population, Intervention, Comparison, Outcomes, and Study (PICOS) design tool. Studies were included that used data from the general population (P), analysed a media intervention illustrating hope and recovery from a suicidal crisis (I), used active controls not exposed to suicidal media content (C), measured suicidal ideation or help-seeking attitudes or intentions as an outcome (O), and used an RCT design (S).
+The media interventions needed to satisfy all the following criteria: first, have a focus on suicidal ideation in the absence of near-fatal or fatal suicidal behaviours; second, feature a personal narrative of hope and recovery; and third, involve media exposure only and not include any other components (eg, skills training). We were primarily interested in stories featuring hope and recovery from the perspective of an individual experiencing a suicidal crisis or ideation, but we also included stories from other perspectives (eg, stories emphasising recovery but featuring peers or professionals). Regarding the content of the control group, restrictions were applied in that it had to be a non-suicide related intervention and comparable to the intervention (eg, similar length, style, and format).
+Studies were excluded if they did not feature a clearly positive story of hope and recovery, had no control group, or had a control group exposed to suicide-related stimulus material. Furthermore, we excluded studies if they did not measure suicidal ideation or help-seeking attitudes or intentions. No restrictions were placed on the instruments used to assess outcomes, the publication dates, or follow-up periods.
+Data collection and analysis
+We contacted the lead or senior authors of all original studies to obtain participant-level data for our metaanalysis. All authors provided their data and the data documentation. These authors were included as coauthors in this study and approved the estimates from their original data.
+From each original dataset, we extracted participantlevel data on age, gender, trial group allocation, baseline scores for suicidal ideation, help-seeking attitudes or intentions, suicide vulnerability, and follow-up scores for suicidal ideation and help-seeking attitudes or intentions. Datasets were checked for consistency and completeness, and where appropriate, data recoding was done to ensure consistency (such that the vulnerability and outcome variables were scored in the same direction). We also extracted study-level data on the number of trial groups, the content of the intervention and control narratives, their length (in time), the scales used to measure ideation and help-seeking attitudes or intentions, how baseline suicide vulnerability was defined, where the trial was done, and follow-up times. Extractions were done independently by SK, TN, and MJS. Discrepancies were discussed and resolved.
+Risk of bias assessment
+Risk of bias within studies was based on the Cochrane risk-of-bias tool for randomised trials.19 This tool assesses bias in RCTs. It includes five domains: first, bias arising from the randomisation process; second, bias due to deviations from the intended interventions; third, bias due to missing outcome data; fourth, bias in outcome measurement; and fifth, bias in selection of reported
+results. Each domain is coded into low, some, or high risk ofbias. Studies were coded as being overall at low risk of bias if all five domains were coded as low risk. Studies were coded as being at some risk of bias if at least one domain was coded to be at some risk, but no domains were coded as being at high risk. Studies were coded as high risk if any domain was at high risk. Domain three largely determined the overall assessment. The coding of bias was done by two independent researchers (unconnected with this study or any of the original studies) for both outcomes. Any discrepancies were discussed and resolved by consensus.
+Risk of bias across studies, primarily due to missing results in the synthesis (publication bias), was assessed visually using contour-enhanced funnel plots and statistically with Egger’s regression test for funnel plot asymmetry of the study-specific estimates.
+Synthesis methods
+For studies with multiple intervention groups, we recoded the data to make a two-group comparison (intervention vs control). We combined intervention groups because we were primarily interested in a direct comparison between Papageno interventions and active controls (rather than examining whether some types of Papageno interventions have greater efficacy than others). We did a primary analysis and two sensitivity analyses using studies rated at low or some risk of bias. The primary analysis was done on studies that tested personal narratives of how to cope with a suicidal crisis and was done using participants experiencing vulnerability. Our first sensitivity analysis used the same set of studies as the primary analysis but included all individuals (ie, those not experiencing vulnerability in addition to those with vulnerability). Our second sensitivity analysis took a broader view of the narratives— that is, including studies testing professional as well as peer narratives of hope and recovery. In this analysis, we included only individuals experiencing vulnerability.
+For all analyses, we estimated the pooled standardised mean difference (SMD). We did this using the two-stage IPD meta-analysis approach. In the first stage, we extracted from the intervention and control groups of each study the mean, its standard deviation, and sample size of each outcome using complete-case methods. We then calculated study-specific SMDs and standard errors (the difference in means between groups divided by the pooled standard deviation). In the second stage, we used these data to estimate the pooled SMD using randomeffects restricted maximum likelihood estimation. Heterogeneity for all models was assessed using the 12 statistic. Values around 25%, 50%, and 75% were interpreted as low, moderate, and high heterogeneity, respectively.20 All analyses were done using Stata (version 16.1) and R (version 3.6.1).
+This study is registered with PROSPERO,21 number CRD42020221341. We made one amendment to the design
+of the study after registration. Instead of stratifying the analyses by vulnerability of study participants, we decided to focus solely on individuals with some degree of vulnerability in the primary analysis. This decision was due to the high relevance of findings in this subsample. Findings for all participants are reported as sensitivity analysis. A further protocol change was that we removed laboratory experiments from the eligible designs to focus only on RCTs. No changes in terms of studies included resulted from this change. We report our study using the Preferred Reporting Items for Systematic Review and Meta-Analyses of Individual Participant Data (PRISM A-IPD) guidelines.22 Ethical approval was obtained from the Ethics Review Board of the Medical University of Vienna (review number 1481/2020).
+Role of the funding source
+There was no funding source for this study.
+Results
+Our search yielded 7347 records, and after removing the duplicates, 3920 records remained for screening (figure 1). After the removal of ineligible records, we retained 25 records that we assessed for eligibility by reading the full text; 17 of these were excluded for specific reasons (figure 1). The appendix (pp 4-6) contains a complete description of the excluded studies. No additional ongoing or completed trials of possible relevance were identified in the search of clinical trials registries as listed in CADTH’s Grey Matters. This left eight unique studies8,10-14’23’24 for our qualitative synthesis and meta-analysis (table and appendix p 7). IPD data were sought and obtained from all eight studies. No issues of data integrity were identified.
+Two studies23,24 were done in Australia and six studies in Austria.8,10-14 The studies had a total of2350 participants randomly assigned to either the intervention or control groups. At baseline, participants had a mean age of 32 years (SD 14, range 18-97) and 60% were female. Because five studies had multiple intervention groups,8’11’12’14’23 there was an imbalance in the allocation to intervention and control groups. In total, 1518 participants (65%) were allocated to the intervention groups and 832 (35%) to the control groups. Detailed information on each study (ie, number of participants, demographic profile, and unavailability of outcomes) is available in the appendix (p 8).
+Vulnerability to suicide before the intervention was recorded in a variety of different ways (table 1). In four studies,8,10’23’24 vulnerability was measured using a suicidal ideation questionnaire and recorded as a binary variable (low vs high) that was split into two groups at the median. One study12 measured vulnerability using a question about suicide attempts in the past year. One study13 measured vulnerability using the Patient Health Questionnaire-9, with scores above 14 indicating vulnerability. One study14 assessed vulnerability with a question about current suicidal thoughts. A final study11
+had no baseline assessment of vulnerability but measured identification with the suicidal person featured in the media story, which has been shown to constitute a relevant factor in vulnerability and media effects.9,15
+Suicidal ideation was measured using the Adult Suicidal Ideation Questionnaire (two studies23,24), the Reasons for Living Inventory (one study10), its subscale, the Survival and Coping Beliefs subscale (four studies8,12-14), or the Implicit Association Test (one study11). Helpseeking intentions were measured using the General Help-Seeking Questionnaire (four studies13,14,23,24) and help-seeking attitudes with the Short Attitudes Towards Seeking Professional Help Scale (two studies10,12). Details on the interventions are in table 1. Follow-up outcome data were collected immediately after the media exposure for four studies,8,11,13,14 1 week after for two studies,10,12 and 4 weeks after for two studies.23,24 Four studies8,10,12,24 were judged to be at low risk of bias and four11,13,14,23 at some risk. No studies were at high risk of bias (appendix p 9), and therefore all studies were included in our analyses.
+For suicide ideation, scales in the original studies were scored so that lower scores were associated with lower levels of suicidal ideation. For the primary analysis, six studies met the inclusion criteria and follow-up data were available for 569 (90%) of 633 participants with baseline ideation scores above the median (345 [55%] participants were allocated to the intervention group and 288 [45%] to the control group). The pooled SMD for this group indicated a small reduction in mean suicidal ideation in the intervention group of -0-22 (95% CI -0-39 to -0-04, p=0-017, six studies; figure 2A). Low levels of heterogeneity were observed in this analysis (12=5%). For the first sensitivity analysis that used all participants regardless of the baseline vulnerability, data were available from the same six studies for 1138 (86%) of 1317 participants (717 [54%] allocated to the intervention group and 600 [46%] to the control group). The pooled SMD for this group was -0-06 (95% CI -0-24 to 0-11, p=0-49, six studies; figure 2B). Moderate heterogeneity was observed (12=49%). For the second sensitivity analysis, which broadened the types of narratives under investigation, all eight studies were included, and baseline data were available for 876 (87%) of 1009 participants who had baseline ideation scores above the median (643 [64%] allocated to the intervention group and 366 [36%] to the control group). For this group, the pooled SMD was -0-13 (95% CI -0-28 to 0-01, p=0-064, eight studies; figure 2C) and the heterogeneity was low (12=0%).
+For help-seeking attitudes and intentions, scales were scored so that higher scores indicated stronger help-seeking attitudes or intentions. For the primary analysis, four studies met the inclusion criteria and follow-up data were available for 362 (86%) of 420 participants who had baseline ideation scores that were above the median (247 [59%] allocated to the intervention group and 173 [41%] to the control group). The
+Figure 1: Study profile
+pooled SMD showed no significant difference between the groups (SMD=0-14, 95% CI -0-15 to 0-43, p=0-35, four studies; figure 3A). Moderate heterogeneity was observed for this set of studies (12=36%). For the first sensitivity analysis that included all participants, follow-up
+data were available for 739 (79%) of 939 participants (556 [59%] allocated to the intervention group and 383 [41%] to the control group). The pooled SMD for this analysis was 0-12 (95% CI -0-10 to 0-35, p=0-28, four studies; figure 3B) and the heterogeneity was moderate (12=50%). For the second sensitivity analysis, follow-up data were available for 583 (82%) of 710 participants who had baseline ideation scores above the median (459 [65%] participants were allocated to the intervention group and 251 [35%] to the control arm). The
+pooled SMD was 0-04 (95% CI -0-16 to 0-25, p=0-69, six studies; figure 3C). The heterogeneity in this analysis was low (12=24%).
+We found no evidence of publication bias in any of the analyses. For the outcome suicidal ideation in the primary analysis, scores were to the left of the null line, but none were outside the contours representing the 1% significance level (figure 4A). All other estimates were distributed symmetrically around the null line with none falling outside the 1% significance level.
+Egger’s test for asymmetry was non-significant for all analyses.
+Discussion
+To the best of our knowledge, this is the first systematic review and meta-analysis about media portrayals of stories of hope and recovery from suicidal crises on suicidal ideation and help-seeking attitudes and intentions. The evidence is that, among individuals with some degree of vulnerability, personal stories of hope have a small protective effect on suicidal ideation up to 4 weeks after exposure.
+The protective effect highlights that these narratives are unlikely to be harmful for individuals with some degree of vulnerability to suicide in the general population. This is
+important because this group is the key target for many media-based suicide prevention efforts. There are several examples of well intentioned media narratives that sought to educate the public and reduce suicide, but which have tragically resulted in an increase in suicides.25,26 These narratives have not typically focused on hope and recovery, but featured a specific suicide or suicide method. This highlights the importance of identifying narratives that do not put individuals at risk of harm.
+Our IPD meta-analysis did not find evidence of an association between personal stories on help-seeking attitudes and intentions. The wide confidence interval of the combined estimate pointed in the expected direction but was based on a small number of studies. Thus, we lack sufficient evidence to make a reliable determination
+about Papageno messages on this outcome. The only original study that found a statistically significant improvement in help-seeking intentions used a narrative that was specifically set up to tackle gender stereotypes to enhance help seeking in men,24 who have higher suicide rates but show less help seeking than women.27 To influence help-seeking attitudes and intentions, a focus on barriers to help seeking in the context of surviving a crisis might be necessary.
+Although the efficacy of messages of hope and recovery from a suicidal crisis was small, there are some reasons to be optimistic about their real-world impacts when implemented at scale. First, universal interventions with low efficacy can still be important if they can be delivered to a large proportion of the population. Media interventions are a good example of this because they can have substantial reach into the population. Narratives of hope and recovery are already readily available, and they are widely acceptable to stakeholders including individuals bereaved by suicide or with personal experience of suicidal ideation and attempts.28 Development and implementation of this type of intervention requires fewer resources than other interventions.29
+Another reason for optimism is that effects were observed in randomised trials where the outcomes were
+measured within a short period of time. Demonstrating efficacy at this stage should be a necessary precondition to further develop public health messages. It is unlikely that media messages will be effective in the real world if they have no effect in the short term. But in contrast to the included RCTs measuring single exposures, in the real world, successful public health advertising works through repetition and over prolonged periods.30 Finally, it is increasingly common to target advertising to a specific population using social media. Many media consumers are self-selecting themselves into media streams that reflect their interests, meaning it is possible to deliver relevant media messages to many people in the target audience at the same time.
+Another key implication of our study is the issue of harm. There are several examples of well intentioned media messages that have caused more harm than good.25,26,31-34 In contrast to this, stories of hope and recovery do not appear to show harmful effects among those who are vulnerable from the general population and could have some benefits. The Papageno effect cannot replace media narratives about fatal suicides if the topic meets the media criterion of newsworthiness, but it provides a potentially novel and safe way forward to educate the public about suicide prevention that can help
+to shift the focus from narratives of despair to a more focused portrayal of how to cope with adversity. Many media guidelines are now rightfully cautious about reporting on suicide.4 Our findings suggest that these guidelines could safely be revised to support the reporting of stories of mastery of a suicidal crisis. Further studies specifically analysing the effects of media stories of mastery on a wide range of audiences, using a variety of narratives that are tailored to the specific target groups and to the primary impact domain under investigation (eg, suicidal ideation or help-seeking intentions), are needed, as are large-scale trials that test any effect of such stories on suicidal and help-seeking behaviour.
+Strengths of this meta-analysis were the specific focus on stories of hope and recovery from a suicidal crisis and the inclusion of a broad set of interventions consisting of stories of different lengths and media types. The inclusion criteria were intentionally conservative in that they focused on evaluating studies with a message of hope and recovery rather than including studies with broad narratives aiming to promote suicide prevention (specifically those studies featuring suicides or details about suicide methods). Most importantly, the narratives all avoided presenting information on specific near-fatal or fatal suicidal behaviours, something which is known to cause harm. Another strength was that we selected only those studies that used active controls. This is a conservative approach because it isolates the effect of the message from other factors such as the way the message was delivered. None of the included studies were at high risk of bias. Finally, cross-study heterogeneity effects were low and there was no evidence of publication bias as evidenced by contour enhanced funnel plots and Egger’s tests.
+There are several limitations of our study. First, this meta-analysis relied on suicidal ideation and helpseeking attitudes and intentions as outcomes, rather than suicidal and help-seeking behaviours. It is not possible to generalise findings to suicidal behaviours due to the low specificity of suicidal ideation. It appears likely, however, that an intervention that reduces suicidal ideation in a substantial proportion of a population would reduce some suicides.35 To the best of our knowledge, there are currently only four published studies that have assessed associations between portrayals of suicidal ideation (typically covering stories of recovery) with suicides.6,32,36’37 Three ofthese studies6,36,38 suggested fewer suicides following stories about suicidal ideation, whereas one study32 did not identify any association. Second, the study could only examine those interventions that have been tested in RCTs, most of which were short, one-off interventions. Third, all source studies included in the meta-analyses were done by members of the study team. We attempted to address the limitation of evaluating our work by using established guidelines for the conduct and reporting and by being transparent about this. Furthermore, the quality
+assessment of included studies was done by independent researchers. Fourth, we assessed effects in a group of individuals with some degree of vulnerability to suicide, but it remains unclear how vulnerable these individuals were. Fifth, although we found no evidence of publication bias, our analyses included a maximum of eight studies. Egger’s test might have had insufficient power to detect publication bias if it were present. Nonetheless, the contour enhanced funnel plots showed good symmetry, which strengthens the argument that publication bias is not a factor in the retrievable literature. Sixth, all the studies we identified were done in either Australia or Austria and, therefore, the results might not generalise beyond these settings. Seventh, the outcome related to help seeking included attitudes and intentions, which are not the same constructs. For future studies, we recommend investigating help-seeking intentions (rather than attitudes) with broadly applicable questionnaires (eg, the General Help-Seeking Questionnaire). Eighth, data on race and ethnicity were not available. Finally, media portrayals can affect various other domains beyond help-seeking attitudes and intentions and suicidal ideation (eg, stigmatisation or prevention-related knowledge), which we did not study.
+To our knowledge, this is the first meta-analysis on the effect of portrayals of personal mastery of suicidal ideation on suicidal ideation and help-seeking attitudes and intentions. Our results suggest that these narratives reduce suicidal ideation in audiences with some vulnerability to suicide in the general population, providing a strong case for their use for suicide prevention.
+Articles
+3
+4
+5
+6
+7
+8
+9
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+Niederkrotenthaler T, Braun M, Pirkis J, et al. Association between 21 suicide reporting in the media and suicide: systematic review and meta-analysis. BMJ 2020; 368: m575.
+WHO. Preventing suicide: a resource for media professionals.
+Geneva: World Health Organization, 2017.	22
+Niederkrotenthaler T, Reidenberg DJ, Till B, Gould MS. Increasing help-seeking and referrals for individuals at risk for suicide by decreasing stigma: the role of mass media. Am J Prev Med 2014;	23
+47 (suppl 2): S235-43.
+Niederkrotenthaler T, Voracek M, Herberth A, et al. Role of media reports in completed and prevented suicide: Werther v. Papageno 24 effects. Br J Psychiatry 2010; 197: 234 43.
+Niederkrotenthaler T, Voracek M, Herberth A, et al.
+Papageno v Werther effect. BMJ 2010; 341: c5841.
+Till B, Arendt F, Scherr S, Niederkrotenthaler T. Effect of educative	25
+suicide prevention news articles featuring experts with vs without personal experience of suicidal ideation: a randomized controlled trial of the Papageno effect. J Clin Psychiatry 2019; 80: 17m11975.
+Till B, Strauss M, Sonneck G, Niederkrotenthaler T. Determining 26 the effects of films with suicidal content: a laboratory experiment.
+Br J Psychiatry 2015; 207: 72-78.	27
+Till B, Tran U, Voracek M, Niederkrotenthaler T. Papageno vs.
+Werther effect online: randomized controlled trial of beneficial and	28
+harmful impacts of educative suicide prevention websites.
+Br J Psychiatry 2017; 211: 109-15.
+Arendt F, Till B, Niederkrotenthaler T. Effects of suicide awareness material on implicit suicide cognition: a laboratory experiment.	29
+Health Commun 2016; 31: 718-26.
+Niederkrotenthaler T, Till B. Effects of suicide awareness materials on individuals with recent suicidal ideation or attempt: online	30
+randomised controlled trial. Br J Psychiatry 2020; 217: 693-700.
+Niederkrotenthaler T, Till B. Effects of awareness material featuring individuals with experience of depression and suicidal thoughts on 31 an audience with depressive symptoms: randomized controlled trial. J Behav Ther Exp Psychiatry 2020; 66: 101515.
+Till B, Tran US, Niederkrotenthaler T. The impact of educative news articles about suicide prevention: a randomized controlled trial.	32
+Health Commun 2020; 36: 2022-29.
+Niederkrotenthaler T, Arendt F, Till B. Predicting intentions to read suicide awareness stories: the role of depression and characteristics 33 of the suicidal role model. Crisis 2015; 36: 399-406.
+Niederkrotenthaler T, Stack S. Media and suicide: international perspectives on research, theory, and policy. NY: Routledge, 2017.	34
+Burke DL, Ensor J, Riley RD. Meta-analysis using individual participant data: one-stage and two-stage approaches, and why they may differ. Stat Med 2017; 36: 855-75.
+Niederkrotenthaler T, Spittal MJ, Till B, et al. Effects of stories of 35 hope and recovery on suicidal ideation and help-seeking intentions: systematic review and meta-analysis. 2020. https://www.crd.york. ac.uk/prospero/display_record.php?ID=CRD42020221341 (accessed Sept 17, 2021).	36
+Sterne JAC, Savovic J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019; 366: l4898.
+Deeks JJ, Higgins JP, Altman DG, Cochrane Statistical Methods 37 Group. Analysing data and undertaking meta-analyses. In: Higgins JPT, Thomas J, Chandler J, et al (eds). Cochrane handbook for systematic reviews of interventions. 2nd edn. Chichester, UK:	38
+John Wiley & Sons, 2019.
+Canadian Agency for Drug Technologies in Health. Grey matters: a practical tool for searching health-related grey literature. 2021. https://www.cadth.ca/grey-matters-practical-tool-searching-health-related-grey-literature-0 (accessed Nov 10, 2021).
+Stewart LA, Clarke M, Rovers M, et al. Preferred reporting items for a systematic review and meta-analysis of individual participant data: the PRISMA-IPD statement. JAMA 2015; 313: 1657-65.
+Ftanou M, Ross A, Machlin A, et al. Public service announcements to change attitudes about youth suicide: a randomized controlled trial. Arch Suicide Res 2021; 25: 829 44.
+King KE, Schlichthorst M, Spittal MJ, Phelps A, Pirkis J. Can a documentary increase help-seeking intentions in men?
+A randomised controlled trial. J Epidemiol Community Health 2018; 72: 92-98.
+Niederkrotenthaler T, Kirchner S, Till B, et al. Systematic review and meta-analyses of suicidal outcomes following fictional portrayals of suicide and suicide attempt in entertainment media. EClinicalMedicine 2021; 36: 100922.
+Schmidtke A, Hafner H. The Werther effect after television films: new evidence for an old hypothesis. Psychol Med 1988; 18: 665-76. Canetto SS, Sakinofsky I. The gender paradox in suicide.
+Suicide Life Threat Behav 1998; 28: 1-23.
+Suicide Attempt Survivors Task Force. The way forward: pathways to hope, recovery, and wellness with insights from lived experience. Washington, DC: National Action Alliance for Suicide Prevention, 2014.
+Austin LS, Husted K. Cost-effectiveness of television, radio, and print media programs for public mental health education. Psychiatr Serv 1998; 49: 808-11.
+Wakefield MA, Durkin S, Spittal MJ, et al. Impact of tobacco control policies and mass media campaigns on monthly adult smoking prevalence. Am J Public Health 2008; 98: 1443-50.
+Bridge JA, Greenhouse JB, Ruch D, et al. Association between the release of Netflix’s 13 Reasons Why and suicide rates in the United States: an interrupted time series analysis.
+J Am Acad Child Adolesc Psychiatry 2020; 59: 236 43.
+Niederkrotenthaler T, Stack S, Till B, et al. Association of increased youth suicides in the United States with the release of 13 Reasons Why. JAMA Psychiatry 2019; 76: 933 40.
+Sinyor M, Williams M, Tran US, et al. Suicides in young people in Ontario following the release of “13 Reasons Why”. Can J Psychiatry 2019; 64: 798-804.
+Cooper MT Jr, Bard D, Wallace R, Gillaspy S, Deleon S. Suicide attempt admissions from a single children’s hospital before and after the introduction of Netflix series 13 Reasons Why.
+J Adolesc Health 2018; 63: 688-93.
+McHugh CM, Corderoy A, Ryan CJ, Hickie IB, Large MM.
+Association between suicidal ideation and suicide: meta-analyses of odds ratios, sensitivity, specificity and positive predictive value.
+BJPsych Open 2019; 5: e18.
+Pirkis JE, Burgess PM, Francis C, Blood RW, Jolley DJ.
+The relationship between media reporting of suicide and actual suicide in Australia. Soc Sci Med 2006; 62: 2874-86.
+Sinyor M, Schaffer A, Nishikawa Y, et al. The association between suicide deaths and putatively harmful and protective factors in media reports. CMAJ 2018; 190: e900-07.
+Sinyor M, Williams M, Zaheer R, et al. The association between Twitter content and suicide. Aust N Z J Psychiatry 2021; 55: 268-76.
+www.thelancet.com/public-health Vol 7 February 2022
+e168

Evaluation-of-a-collaborative-care-model-for-integrated-primary-care-of-common-mental-disorders-comorbid-with-chronic-conditions-in-South-AfricaBMC-Psychiatry.txt

@@ -0,0 +1,64 @@
+Background
+Mental disorders are on the rise globally, and are often co-morbid with other chronic conditions, being two to five times more prevalent in people with chronic physical health conditions than the rest of the population [1-4]. Mental-physical comorbidities are associated with greater decrements in health outcomes [1] and increased health care utilization costs [5]. The need for chronic disease management to include treatment for co-existing common mental disorders (CMDs) is thus increasingly viewed as a priority in the global challenge of care for multi-morbidity [6].
+South Africa is one of a growing number of low- and middle-income countries (LMICs) experiencing a rising burden of multi-morbid chronic conditions [7]. This is a consequence of the transition of HIV to a chronic condition with the scale-up of antiretroviral treatment; as well as the intensifying non-communicable disease (NCD) burden. In response the South African Department of Health has pioneered an Integrated Clinical Services Management (ICSM) approach that strives to service the majority of patients in primary health care (PHC) at a single delivery point using integrated clinical chronic care guidelines [8]. Integration of mental health care is part of ICSM in South Africa, but has been shown to be inadequate [9]; with a treatment gap of 75% for CMDs in South Africa [10]. While there is evidence of the effectiveness of collaborative care models for the treatment of common mental disorders (CMDs) comor-bid with chronic physical conditions from high-income countries [11], there is little evidence of the effectiveness of task-shared collaborative care models for physical and mental multi-morbidity from LMICs.
+The aim of this study was to evaluate an integrated collaborative care package of care for chronic patients with co-existing depressive and alcohol use disorder (AUD) symptoms that strengthened identification and management of these CMDs under real world conditions through strengthened referral pathways in one case study district in South Africa [12]. The study forms part of the PRogramme for Improving Mental Health CarE (PRIME) research consortium concerned with the development, implementation and evaluation of integrated packages of care for priority mental disorders at PHC level in five LMIC countries [13]. The specific objectives of this study were to assess primary outcomes of whether the collaborative care package: i) improved provider identification of depressive and AUD symptoms in chronic care patients and ii) reduced depressive symptoms and improved functioning in screen positive chronic patients identified and referred for care within the task shared stepped up collaborative care model. Secondary outcomes assessed effects in relation to health equity criteria.
+Methods
+Setting
+The study site was in the Matlosana municipality in Dr. Kenneth Kaunda District (DKK) in the North West Province. The district population was estimated to be approximately 796,823 at the time of the study, while the catchment areas where this study was located comprised over 90,000 people serviced by four primary health care facilities varying in size, serving between 2353 and 6058 chronic care patients per month. Further details of the DKK district can be found elsewhere [12].
+Description of the collaborative mental health care package
+Details of the collaborative care package that was developed through the formative phase and evaluated by this study are described in greater detail in Petersen et al. [12]. In brief, it comprised the following five components: i) PHC nurses functioned as case managers and were oriented to the ICSM, trained in clinical communication skills to facilitate person-centered care, and provided with supplementary mental health training in basic adult care guidelines (known as Adult Primary Care in South Africa) [14]; ii) Doctors were oriented to the importance of mental health and upskilled to prescribe antidepressant medications; iii) Referral pathways for psychosocial counselling for patients with mild to moderate depressive symptoms were strengthened with the introduction of clinic-based lay counsellors trained and supervised to deliver individual and group-based counselling drawing on cognitive behavioural therapy techniques which have international evidence of effectiveness [15]; and v) A referral form to monitor nurse referrals to the counsellor was introduced.
+Research design
+The research design was pragmatic with the intervention delivered independently of the evaluation. Given the complex nature of the collaborative care package, the evaluation comprised two main components: i) a Facility Survey to assess effects of the intervention on provider detection of depression and AUD; and ii) a non-randomly assigned comparison group cohort study to assess changes in symptom severity and functioning among screen positive patients identified and referred for further care by the PHC nurses compared to those not identified.
+Facility detection survey (FDS)
+Study procedure, sample and measures
+The primary objective of the Facility Detection Study was to estimate the change in detection of depression and of AUD by clinicians serving the adult chronic care population in intervention clinics. The design and power
+calculations for the sample sizes are described in greater detail elsewhere [16]. The FDS was conducted in three of the four facilities where the mental health APC module had not yet been implemented. The study procedures used at baseline and follow-up FDS were the same, with independent samples recruited in each study round. The baseline FDS was conducted from February-April 2014, before the implementation of the intervention began in April 2014. Training and embedding continued to September 2014. The follow-up FDS was 12 months after completion of the embedding period (October-December 2015) (see Fig. 1).
+Adult patients were recruited from the chronic care waiting areas of the three PHC facilities by trained recruiters who gave short oral presentations on the study. No reference to mental health was made to minimize sampling bias and limit stigma. Eligibility of patients who volunteered was confirmed in a private space and study objectives discussed. Eligibility criteria were: 18 years or older; attending the clinic for treatment for a chronic illness (e.g., HIV, tuberculosis, hypertension, etc) and capacity to understand the questions posed in either seTswana (the dominant local language) or English. Written informed consent was obtained from literate participants and illiterate participants consented by marking the form with a cross; with a witness countersigning.
+Trained fieldworkers administered a structured questionnaire, programmed into mobile devices. Questions included items on demographic characteristics, treated chronic condition(s), and screening instruments for probable AUD and depression. The Alcohol Use Disorder Identification Test (AUDIT), validated for use in South Africa [17], was used to screen for probable AUD, with participants scoring >16 considered positive for probable AUD, given nurse guidelines to provide advice for harmful drinking and refer patients with dependent drinking. Cronbach’s alphas were 0.78 (baseline) and 0.74 (follow-up). For probable depression, we used the Patient Health Questionnaire-9 (PHQ-9), with a cut-off of >10 being previously validated on a primary care population in South Africa [18]. Cronbach’s alphas were 0.88 (baseline) and 0.86 (follow up).
+All screen positive participants on either screening tool, as well as 15% of randomly selected screen-negative participants (both AUDIT and PHQ-9) were asked to return for an exit interview immediately following their clinical consultation.
+Participants’ exit interview data were used to assess clinical detection on the day of the interview. Broad criteria were used to classify participants as detected for AUD given our experience of nurse reticence to make a diagnosis of AUD (patients who reported a diagnosis of harmful or dependent drinking, and/or a referral to specialist alcohol services, and/or who received advice
+about managing problems with drinking alcohol). Two classifications (narrow and broad) were used for detection of depression: narrow - a diagnosis of depression was reported; broad - diagnosis of depression and/or referral for psychosocial counselling reported. The latter was included given our experience that nurses did not always inform patients of a diagnosis of depression when referring patients. Anti-depressant medication prescribed was also assessed. Participants who reported being on current treatment for either condition were excluded from the analysis.
+Statistical analyses
+The socio-demographic and clinical characteristics of the participants recruited over the two study rounds were analyzed using means and standard deviations for continuous measures and counts and proportions for categorical measures. For each study round we report the number of participants who screened positive on AUDIT and PHQ-9, the number of screen positive participants who completed the exit interview, and the proportion who were classified as having been detected for AUD (among AUDIT positive), or for depression (among PHQ-9 positive) using both narrow and broad criteria. For the equity analysis, where sufficient data were available, we assessed whether change in detection over time was equitable by sex and by household food security. Both these demographic variables were shown as having the highest odds ratios for depression/AUD comorbidity in the same population in a previous study [19]. For the inequity analysis, we used binomial regression models to estimate change in detection, and included the relevant interaction term into the models, with an interaction term p-value< 0.20 suggestive of inequity. Second, for the screen negative PHQ-9 participants who were randomly selected to complete the exit interview, we tabulated the number of depression diagnoses and anti-depressant medication prescription. When it was not possible to estimate the change in detection because of zero counts in the baseline round, we reported the proportion detected and 95% confidence interval for the follow up round only. Procedures to estimate proportions, 95% confidence intervals and p-values incorporated weights to adjust for the imbalance in clinic-level sample sizes between rounds [20].
+Depression cohort study
+Study procedure, sample and measures
+Details of the sample size calculation, recruitment, questionnaire design, data collection procedures, and analysis plan for the PRIME cohort studies have been described in detail by Baron et al. [21] A cohort study for AUD patients was not conducted in South Africa because of
+the very low levels of AUD identification by the providers in the baseline round of the FDS.
+Eligible participants were at least 18 years old; receiving care for a chronic physical condition; screened positive for probable depression and/or had been identified and referred for depression care by the providers; had the capacity to provide informed consent and comprehend the interview; and did not have a diagnosis of AUD or psychosis (see Fig. 2).
+Chronic care patients were informed about the study in the waiting areas of the clinics prior to their clinic consultations and informed written consent was obtained from volunteers. The same informed consent procedure was followed for low-literacy patients as described for the FDS. Immediately after their clinical consultations, individuals who had given their consent were screened using the PHQ-9, and asked whether, during the consultation, they had been identified as having depression and/ or had been referred for care to a provider within the collaborative care model. Individuals who provided affirmative responses to these questions were enrolled into the depression treatment group, regardless of their PHQ-9 score. Individuals who did not provide affirmative responses to these post consultation questions, but who scored 10 or more on the PHQ-9 were enrolled into a comparison group. Comparison group participants who later received a depression diagnosis were re-enrolled in the treatment group, and only the treatment group data of these participants were analyzed (see Fig. 2).
+Cohort study participants completed three assessments: at enrolment (baseline); 3 months and 12-months post-enrolment. Cohort recruitment and enrolment occurred from August 2014 to July 2015, and follow-up was conducted from November 2014 to September 2016 (see Fig. 1).
+Mobile devices were used by trained seTswane/English speaking fieldworkers to administer the questionnaire in private spaces at the clinic or participants’ homes. Each assessment comprised a range of demographic, clinical, health care use, social, economic, food security, and stigma-related measures. These are described in greater detail by Baron et al. [21]. Only measures pertaining to the socio-demographic and clinical characteristics of the sample are reported here.
+Primary outcomes of the cohort study were response on the PHQ-9, defined as at least 50% reduction in score from baseline to 3 months and 12 months follow-up; and remission on the PHQ-9 at both follow-up visits, defined as a score of 5 or less - used as measures of clinically significant improvement in treatment trials using the PHQ-9, e.g. Huijbregts et al. (2013) [22]. Functional impairment, was assessed at the three time points using the 12-item WHO Disability Assessment Schedule (WHODAS 2.0) - previously used in South Africa
+amongst older and HIV populations [23]. Item response theory (IRT)-based scoring was used, with scores ranging from 0 to 100, with a higher score indicating greater functional impairment.
+Analysis
+Given the non-normal distribution of the sample’s demographic and clinical characteristics, baseline characteristics of participants in the treatment and comparison groups were compared using non-parametric tests -the Mann-Whitney U test for continuous measures, and Exact Fisher’s test for categorical variables. The mean symptom severity and functioning were compared between groups at each follow-visit. Given that neither measure was normally distributed at follow-ups, a multilevel mixed effect negative binomial regression was used, controlling for HIV status and recruitment clinic, as these were imbalanced at baseline [21]. Risk ratios for the primary outcomes (i.e. 50% reduction in scores at 3 months and 12 months follow-up, as well as remission) were assessed using a modified Poisson regression, with robust variance estimator, as binomial models failed to converge [24]. Again, the models were adjusted for demographic or other clinical differences between the comparison and treatment groups at baseline. To assess equity in primary outcomes across gender and household food security, the same negative binomial regressions were conducted, this time including either gender or household food security at baseline as an interaction term.
+Results
+Facility detection survey
+In the first round, 1322 participants were eligible and consented to participate in the study. Twelve of these were on current treatment for depression and excluded from the study, resulting in a total of 1310 participants at baseline. During the second round, 1257 were eligible and consented to participate in the study; of these 11 were found to be on current treatment for depression and excluded, resulting in a total of 1246 participants at follow-up. The demographic and clinical characteristics of the Facility Detection Survey participants are presented in Table 1.
+Across both survey rounds, the mean age of the sample was 46 years; approximately three-quarters were women; the majority were receiving care for HIV or hypertension. There were significant between-round differences in participants’ employment status, food security, clinic site, and AUDIT screening.
+For depression, as seen in Table 2(a), the pre-imple-mentation diagnosis of depressive symptoms was 5.2% using the narrow definition, and 14.2% using the broader definition. Post-implementation, using the
+narrow definition, detection of depressive symptoms reached 16.2%, an increase of 11.0% (95% CI 3.3, 18.6); using the broad definition, detection increased to 26.7%, an increase of 12.5% (95% CI 1.9, 23.0). In the inequity analysis, the change in detection for food secure participants, using the broad definition, increased to 33.8%, compared to 23.0% for food insecure participants (P = 0.080). There was no evidence of inequity (P = 0.656) for men versus women participants.
+For AUD, no AUDIT-positive participants were detected in the pre-implementation round. In the
+post-implementation round, 11.7% (95% CI 0.6, 22.8) were detected. There were insufficient data for conducting inequity analyses for AUD detection and treatment.
+As seen in Table 2(b), of the 1310 participants enrolled in the pre-implementation round, 1084 screened negative on both PHQ-9 and AUDIT. Of the 1084, 118 were selected randomly to complete the exit interview, and 110 were successfully interviewed after their consultations. Of the 110, 0.9% had been detected and treatment initiated for depression using the narrow definition; and 6.7% using the broad definition. For AUD, 1.6% were prescribed anti-depressant medication and 1.6% were
+diagnosed with AUD. These proportions did not change significantly (all P > 0.05) between the pre- and post-implementation surveys.
+Depression cohort study
+Of 2602 patients screened, a total of 453 participants were enrolled in the cohort study. An initial 205 patients were diagnosed with depression and recruited into the treatment group. Another 248 patients were not diagnosed but screened positive on the PHQ-9, and were recruited into the comparison group; of these, 12 participants were subsequently diagnosed with depression at a follow-up visit at the clinic, and re-enrolled into the treatment group. The final sample was 236 for
+the comparison group and 217 for the treatment group (see Fig. 2).
+There were 82 participants (18.1%) lost to follow-up, mostly because of relocation and refusals. In the intervention group, 88.9% (n = 193) participants were followed at 3 months and 81.5% (n = 177) completed the 12-month assessment. In the comparison group, 88.6% (n = 209) completed the midline assessment and 82.6% (n = 195) the end-line assessment (see Fig. 1).
+Of the 217 participants in the treatment group, 80 participants screened below the clinical cut-off of 10 on the PHQ-9 at baseline. These participants were excluded from the analysis, to ensure that both treatment and control groups were comparable. These participants did not differ from those included in the analyses in terms
+of demographic characteristics at recruitment, besides food insecurity (Table 3). The final sample included in the analysis comprised 373 participants: 137 and 236 in the treatment and comparison groups, respectively.
+Participants in the comparison and treatment groups who were included in the analyses differed in terms of clinic of recruitment and in HIV status (Table 3). Also, participants recruited into the treatment group had significantly higher PHQ-9 scores (mean = 14.5, SD = 3.47), compared to participants in the comparison group (mean = 12.8, SD = 3.01).
+Results of the modified Poisson regressions are presented in Table 4. The proportion of participants showing at least a 50% reduction in PHQ-9 scores from
+baseline to the 3-month follow-up was greater in the treatment group (N = 69, 55.2%) than in the comparison group (N =49, 23.4%; RR = 2.10, p <0.001). The rate of participants who showed remission on the PHQ-9 (score < 5) was also greater in the treatment group (N = 40, 32.0%) compared to the comparison group (N = 25, 12.0%; RR = 2.78, p < 0.001). The same significant trends were found at the 12-month follow-up, with 57 (47.9%) participants in the treatment group reporting at least 50% reduction in PHQ-9 score compared to baseline, and 32 (26.9%) participants scoring below 5, compared to 60 participants (30.8%; RR = 1.52, p = 0.006) and 33 participants (16.9%; RR = 1.72, p = 0.016) in the comparison group, respectively.
+Change in mean scores on the PHQ-9 and WHODAS over time are also presented in Table 4. After controlling for HIV status and clinic of recruitment, the mixed effects analyses reveal a significant difference between the two groups at the 3-month follow-up; a greater decrease in PHQ-9 scores in the treatment group (M = -5.05, 95%CI: -6.02 to - 4.08) compared to the comparison group (M = -2.63, 95%CI: -3.34 to - 1.92; fi = -2.42, p < 0.001) and a greater decrease in WHODAS scores from baseline in the treatment group (M = -11.11, 95%CI: -15.18 to - 7.05) compared to the comparison group (M = -4.65, 95%CI: -8.02 to - 1.28; fi = - 6.46, p = 0.010). A similar trend was seen for the PHQ-9 scores at the 12-month follow-up compared to baseline (treatment: M = -5.07, 95%CI: -6.05 to - 4.09; comparison: M = -3.10, 95%CI: -3.83 to - 2.37, fi = - 1.97, p <0.001). The change in WHODAS scores at the 12-month follow-up was not significantly different between the two groups.
+Inequity analyses of the impact of the intervention on outcomes by gender and household food security are presented in Table 5. There is no evidence of inequity by gender at either follow-up time points. At 3 months follow-up, however, the intervention was found to have a significantly more positive effect on reducing depressive scores among participants who reported being food secure at baseline (fi = - 3.73, 95%CI -5.30 to - 2.15) (M = - 6.62, 95% CI: -7.95—5.28) compared to those reporting being food insecure (fi = - 1.23, 95%CI -2.65 to 0.20; 0 = - 2.50, p < 0.05). This trend persisted at 12 months follow-up.
+Discussion
+The results of the FDS suggest an improvement in nurse-detection and treatment initiation of depressive and AUD symptoms following implementation of the integrated collaborative care package in the district. Notably, identification remained essentially absent for individuals who were probable non-cases. In other words, the positive predictive value of nurse identification and treatment initiation was high for both depression and AUD.
+The results of the cohort study indicate that patients correctly identified and referred for further care for their depressive symptoms had a greater chance of having a clinically significant reduction in depressive symptoms to the point of being in remission at both 3 and 12 month follow-up than those not referred. They also showed a significant reduction in functional disability at 3-month follow-up compared to the non-intervention group, although this effect was not sustained at 12 months. The intervention was significantly more successful with food secure participants, suggesting that food insecurity, may serve as a barrier to improvement in depressive symptoms; being associated with poverty in more urban areas in South Africa [25]. This adds to
+the growing body of evidence suggesting the need for accompanying income generating initiatives for people with depressive symptoms from poor socio-economic contexts [26].
+While an improvement in the correct identification and treatment initiation of patients with depressive and AUD symptoms was noted, the treatment gap post training still remained large, with 75% of probable cases of depression and 84% of probable cases of AUD still not detected at 12 months follow-up. Previous studies suggest that training alone may be insufficient to improve identification of CMDs in PHC [27]. Other factors contributing to this gap resonate with our understanding of possible reasons and include individual provider level factors - with psychiatric stigma as well as providers’ own personal unresolved problems previously shown to act as barriers to the identification of emotional problems in patients [28, 29]. The inclusion of anti-stigma interventions [28], stress management and debriefing sessions to assist PHC personnel to engage in emotional labour has been previously suggested to assist integration efforts [29]. Further, the need for change management processes to accompany organizational changes associated with integrated care has also previously been highlighted [8].
+Limitations
+There are a number of limitations of this evaluation. With regard to the FDS: i) Non-random sampling compromised the representativeness of the sample. An imbalance in the demographics between the two rounds was noted - with an upward trend in food insecurity, unemployment and alcohol screen positives in the second round - partially explained by retrenchments on the mines between the two survey rounds, with mining being a major industry and source of employment in the study site; ii) There were no control clinics involved in the FDS - with the possibility that the some of the improvements in detection were due to other factors in the health system; iii) An upwards bias in detection may have been introduced given patients’ heightened awareness of their potential symptoms through exposure to the survey interview prior to their consultation - although offset by being the case for both rounds as well as having screen-negative cases; and iv) The FDS only assessed clinical detection on the day of the interview -with diagnosis of a CMD generally taking several visits [27] - thus potentially under-representing detection rates. In relation to the cohort study, recruitment into the intervention and control arm was not randomized, thus increasing potential for bias from unknown con-founders, although partially mitigated by multi-variable analyses.
+Conclusion
+This evaluation shows that the PRIME-SA intervention package assisted PHC teams to identify and manage comorbid CMDs in chronic patients under real world conditions using a collaborative stepped care model. There was an improvement in accurate detection of CMDs by PHC nurses, and a reduction in depressive symptoms in patients identified with comorbid depression and referred for care. Further research is required to assess patient outcomes in patients with comorbid AUD. In the face of the negative impact that comorbid CMDs have on treatment adherence and overall health outcomes in chronic patients, this collaborative task shared model provides a potential model for mental physical multi-morbid disease management in South Africa and other LMICs in the context of specialist resource shortages. Additional findings will be provided by two parallel pragmatic cluster randomized control trials that are currently underway, testing effectiveness of this model on depression and physical outcomes (viral load suppression in HIV patients and reductions in blood pressure in hypertensive patients) in chronic care patients [30, 31].

Evaluation-of-the-time-to-change-programme-in-England-20082011British-Journal-of-Psychiatry.txt

@@ -0,0 +1,44 @@
+The British Journalof Psychiatry (2013)
+202, S45-S48. doi: 10.1192/bjp.bp.112.112896
+Claire Henderson (pictured) is Clinical Senior Lecturer in Psychiatry, King's College London, Institute of Psychiatry, and a consultant psychiatrist at the South London & Maudsley NHS Foundation Trust. Graham Thornicroft is a consultant psychiatrist at the South London & Maudsley NHS Foundation Trust and Professor of Community Psychiatry, King's College London, Institute of Psychiatry.
+Time to Change (TTC) is the largest-ever programme in England designed to reduce stigma and discrimination against people with mental health disorders (http://www.time-to-change.org.uk/).1 The first phase of this initiative was run by three charities: Mental Health Media, Mind and Rethink Mental Illness. It was funded in the first phase with £16 million from the Big Lottery Fund and £4.5 million from Comic Relief. The programme has also benefited from the secondment of two members of staff from the Department of Health to work on stakeholder management and policy. The Department of Health also funded the annual national Attitudes to Mental Illness survey.2 The programme went on to run two sports-related programmes: the Sport and Mental Health Project (funded by the Department of Health with £83 000) and Imagine Your Goals (funded by Sport Relief and the Premier League with £620 000).
+Evaluation of the Time To Change programme
+The outcomes set by the Time To Change programme were:
+(a)	significantly increased public awareness of mental health (an estimated 30 million English adults would be reached), a 5% positive shift in public attitudes towards mental health problems and a 5% reduction in discrimination by 2012;
+(b)	100 000 people with mental health problems to have increased knowledge, confidence and assertiveness to challenge discrimination by 2012;
+(c)	provision, through physical activity, of greater opportunities for 274 500 people with a range of mental health problems to come together, both to break down discrimination and to improve well-being, by 2012.
+Time To Change was aimed both at the general population and at specific target groups (identified by people with experience of mental health problems) as well as at people with mental health problems themselves. To maximise its reach - and thus its value for money - it engaged individuals, communities and stakeholder
+organisations such as statutory health services and professional membership groups in distributing social marketing campaign materials, collaborating in staging public relations events and holding events to promote social contact between people with and without experience of mental health problems.3-10
+Evaluation of the TTC programme was based on a conceptual framework that understands stigma as consisting of difficulties of knowledge (ignorance or misinformation), attitudes (prejudice) and behaviour (discrimination).1,11 Changes in public attitudes were measured every year from 2008 to 2012 using the Department of Health’s national Attitudes to Mental Illness general population survey in England.2,12 Since its inception the survey has used a shortened list of items from the Community Attitudes toward the Mentally Ill (CAMI) scale and the Opinions about Mental Illness Scale,13,14 providing data on attitudes from 1993. In collaboration with SHiFT, which commissioned this survey between 2008 and 2011, we also developed and from 2009 added the Mental Health Knowledge Schedule (MAKS) and the Reported and Intended Behaviour Scale (RIBS) to the pre-existing attitude questions,15,16 in line with our conceptual model.
+To assess progress towards the target of a 5% reduction in discrimination we conducted an annual survey from 2008 to 2011 of discrimination as experienced by people using mental health services across England (‘Viewpoint’),17,18 using the Discrimination and Stigma Scale.19 The results are reported by Corker et al (this supplement).20Any impact of the social marketing campaign (budget £8 311066) was likely to be influenced by concurrent reporting on mental health-related topics in the mass media.21 The nature and balance of media coverage are of concern to anti-stigma campaigns internationally,22,23 leading to increasing interest in methods of content analysis.24 Analyses of English press coverage are presented by Thornicroft et al (this supplement).25
+Employers were a specific target for stakeholder engagement, and were intended users of the Time to Challenge online resource (budget £196 049), which explained good practice in the field of employment and mental health, and the rights of employees with mental health problems. Henderson et al (this supplement) report evidence of changes in employers’ knowledge, attitudes and practice in this field,26 from the repeated survey in 2009 and 2010 of a survey originally undertaken by the Shaw Trust in 2006.27,28
+Two aspects of the social marketing campaign are reported by Evans-Lacko et al (this supplement).29 First, the national TTC
+social marketing campaign used bursts of mass media advertising and public relations exercises twice a year from 2009 to 2011. The key messages of the first two bursts addressed knowledge important in reducing stigma, i.e. that mental illnesses are common and that people with such disorders can lead meaningful lives. Bursts three and four addressed prejudicial attitudes, i.e. mental illness is our last taboo, such that the accompanying discrimination and exclusion can affect people in a way that many describe as ‘worse than the illness itself’. The last two campaign bursts addressed behaviour change; i.e. we can all do something to help people with mental illness, such as maintaining social contact. Selected knowledge, attitudes and behaviour questions from the three measures used in the Attitudes to Mental Illness survey were used to evaluate the impact of each burst on the pre-identified targeted demographic group of people aged 25-45 years in middle-income groups, and this showed a positive impact on those aware of the campaign for five of the six bursts. Second, a strikingly original component of TTC involved the attendance of large numbers of people with experience of mental health problems at a series of one-day events designed to deliver social contact, addressing the second and third TTC targets (budget £1 077 214). Although the evidence for social contact in reducing prejudice towards people with mental health problems largely concerned its short-term impact,9,30 these events also increased awareness of the social marketing campaign, and together this may have created a cumulative and more sustained effect. Our data suggest a positive relationship between the quality of social contact and a reduction in prejudice (both of improved attitudes and greater confidence to tackle stigma). Time to Change similarly delivered social contact through other programme components; 32 small-scale anti-discrimination initiatives (‘Open Up’, budget £1407 243) aimed to empower people through awareness-raising and confidence-building groups and anti- discrimination projects, many of which involved the use of the creative arts. Another set of projects comprised exercise programmes for people with mental health problems in community leisure facilities delivered by local Rethink and Mind associations (budget £4431 705).
+For specific target groups (medical students, trainee teachers, trainee head teachers and social inclusion officers), the Education Not Discrimination (END) component of TTC again used social contact (budget £1 310201).9,10,31 Friedrich et al (this supplement) described the effect of END on the knowledge, attitudes and intended behaviour of medical students at four English medical schools.32 We present the results for this target group only because it is of greatest interest to this journal’s readership and because we were able to include a control group in the design, which was not the case for the other groups. The results suggest initial positive effects that were no longer present at the 6-month follow-up assessment.
+It is vital that this investment has clear national economic benefits,33,34 and so an economic evaluation was applied to most of the TTC components. In view of the high advertising costs of social marketing, Evans-Lacko et al (this supplement) present the results of an evaluation of the TTC social marketing campaign costs in relation to outcomes.35 This applied an innovative model,36 in conjunction with social marketing campaign evaluation data, to investigate the economic impact of the campaign, including the potential effects on the wider economy.
+Strengths and limitations of the evaluation
+One wholly innovative aspect of the TTC programme is its annual measurement of discriminatory experiences on the part of those using mental health services, rather than evaluating only public
+knowledge and/or attitudes.37-39 Economic analyses have been lacking in previous campaign evaluations and analysis of changes in press coverage over time has been more limited.22 The evaluation is thus relatively comprehensive, as well as informed by the involvement of people using mental health services in the development and administration of new measures.1 -
+The main limitation of this evaluation was the inability to determine the exact contribution of TTC to the changes reported in annual survey results compared with other influences on public attitudes and behaviour, newspaper coverage and employers, owing to the lack of a control population.2,20,25,26 Nevertheless, it is possible to be fairly confident that pre-burst to post-burst changes seen for the anti-stigma campaign bursts were due to the programme per se.29 Further, the Viewpoint study suffered from low response rates.20 However, after weighted analysis of the Viewpoint samples to take account of the overrepresentation of participants of White ethnicity, female gender and older age the main findings were unaffected.
+Implications of the results
+Among our assessments of knowledge, attitudes and behaviour, the most marked change between 2008 and 2011 was the significant overall reduction in the levels of experienced discrimination reported by people using mental health services.20 This survey is the first of its kind so we cannot compare these findings with previous research. However, the results are in clear contrast to the lack of improvement in public attitudes found in England, Scotland and the USA during the previous 10-15 years.12,40
+After the positive change between 2008 and 2010 there was a negative shift both in public attitudes and in some Viewpoint items.2,20 The contemporaneous national economic problems might have exacerbated inequality in access to employment for people with mental health problems,41 despite and/or since the improvements found in the survey of employers between 2006 and 2010.26 There is also evidence that hostile and stigmatising behaviour towards groups with other disabilities has increased since 2010, for example towards people with cerebral palsy (http:// www.scope.org.uk/news/attitudes-survey). This hostility might also affect people with mental health problems.42 However, although reported discrimination in terms of safety, benefits and transport appears to have increased, these increases are not significant after allowing for multiple testing of Viewpoint items.
+The patterns of changes in the Viewpoint items, taken with the positive effect of social contact on outcomes among the campaign target group, suggest that reducing stigma and discrimination might depend increasingly on more social contact, which should be explored in future work. Newspaper coverage changes also suggest such a polarisation, in that fewer articles in 2012 were neutral compared with 2008.25 Journalists and editors may themselves have become more polarised and/or be catering for more polarised attitudes in their readership. These findings raise a key question for phase 2 of TTC: that is, whether individuals with lived experience of mental illness and those close to them can, through greater disclosure, contribute to higher levels of social contact at the population level with those with mental health problems, thus reducing public stigma. The results presented by Evans-Lacko et al (this supplement),2 concerning greater levels of reported contact among the respondents of the Attitudes to Mental Illness survey, offer some support for this view.7
+The lack of change in levels of experienced discrimination from health professionals among Viewpoint participants is of concern;20 whereas initial help-seeking for mental health problems might
+increase if public attitudes and behaviours improved, a lack of reduction in the rate of negative experiences with health professionals might deter people from seeking further help. It may be that the campaign lacked market penetration among health professionals, or that the ‘clinical fallacy’ means their attitudes and behaviour are more resistant to change, i.e. the accumulated experience of staff is that they most often see people with the worst course and outcome. Medical students are also exposed to this bias, which may mitigate the impact of END.32 In contrast with this finding, evaluation of the TTC programme components was on the whole positive, including the economic evaluation.29,35
+Stigma and discrimination against people with mental illness are global challenges,19,43 and the evidence of our evaluation of phase 1 of TTC is that they can be successfully tackled with a focused, determined and long-term approach.44 With this British Journal of Psychiatry supplement we intend to communicate the results of the first phase (2008-2011) of the TTC programme to those who need to know how to intervene most effectively for the greater social inclusion of people with mental health problems worldwide.
+Henderson & Thornicroft
+36	McCrone P, Knapp M, Henri M, McDaid D. The economic impact of initiatives to reduce stigma: demonstration of a modelling approach. Epidemiol Psichiatr Soc 2010; 19: 131-9.
+37	Crisp A, Gelder MG, Goddard E, Meltzer H. Stigmatization of people with mental illnesses: a follow-up study within the Changing Minds campaign of the Royal College of Psychiatrists. World Psychiatry 2005; 4: 106-13.
+38	Akroyd S, Wyllie A. Impacts of National Media Campaign to Counter Stigma and Discrimination Associated With Mental Illness: Survey 4. Phoenix Research, 2002.
+39	See Me. See Me So Far. A Review of the First Four Years of the Scottish Anti Stigma Campaign. Scottish Executive, 2007.
+40	Pescosolido BA, Martin JK, Long JS, Medina TR, Phelan JC, Link BG. ‘A disease like any other’? A decade of change in public reactions to
+schizophrenia, depression, and alcohol dependence. Am J Psychiatry 2010; 167: 1321-30.
+41	Warner R. Recovery from Schizophrenia: Psychiatry and Political Economy. Brunner-Routledge, 2004.
+42	Clement S, Brohan E, Sayce L, Pool J, Thornicroft G. Disability hate crime and targeted violence and hostility: a mental health and discrimination perspective. J Ment Health 2011; 20: 219-25.
+43	Ucok A, Brohan E, Rose D, Sartorius N, Leese M, Yoon CK, et al. Anticipated discrimination among people with schizophrenia. Acta Psychiatr Scand 2012; 125: 77-83.
+44	Sartorius N. Short-lived campaigns are not enough. Nature 2010; 468: 163-5.
+®OPEN
+ACCESS
+s48
+https://doi.org/10.1192/bjp.bp.112.112896 Published online by Cambridge University Press

Evidence-Based.txt

@@ -0,0 +1,62 @@
+First-episode psychosis (FEP) usually refers to the initial psychotic episode of a primary psychotic disorder, which often results in fear, confusion, and significant disruption in the individual’s life and that of the family. Over the past decade, specialized FEP programs, which combine antipsychotic treatment with psychosocial treatments, have become more widespread in the United States. Individual, group, and family psychotherapy components of comprehensive programs are critical in helping clients and families understand and process the experience of psychosis and learn strategies to promote recovery and well-being.
+Coordinated specialty care (CSC) programs (referred to as early intervention services [EISs] in Europe and Australia) provide team-based, comprehensive, evidence-based care, education, and support to engage clients and their families early in the course of illness.1 The goals of these programs are to reduce the duration of
+untreated psychosis (DUP; ie, the period between onset of symptoms and initiation of antipsychotic medication treatment), prevent further disability, and promote recovery and well-being.
+There are differences in the eligibility criteria across programs,2 but typically CSC programs treat individuals between 15 and 40 years of age (although the United Kingdom has begun offering FEP services to anyone regardless of age)3 who are within the first few years of the onset of their psychosis. CSC programs are intended for individuals with primary psychosis and are not meant for individuals whose psychotic symptoms are judged to be secondary to substance use (eg, substance-induced psychosis), a mood disorder (eg, bipolar disorder, major depression), a developmental disability (eg, autism), or posttraumatic stress disorder. The most common diagnoses of persons treated in CSC programs are schizophreniform disorder, schizophrenia, and schizoaffective disorder.
+RESEARCH SUPPORT
+CSC programs have been shown to yield better outcomes than treatment as usual, including fewer symptoms, more school/work participation, less treatment dropout, and reduced use of inpatient services.4 Studies have consistently highlighted both the importance of well-resourced EISs,5,6 and of shortening DUP to improve outcomes in persons with FEP.7-9
+United States
+Recovery After an Initial Schizophrenia Episode (RAISE) was a large-scale research initiative, funded by the National Institute of Mental Health (NIMH), which involved 2 studies in the United States, the RAISE-ETP trial7,10 and the RAISE Connection Program Implementation and Evaluation study.11 The RAISE-ETP trial enrolled 404 participants in a cluster randomized controlled trial involving 34 community mental health centers in 21 states to deliver 24 months of the NAVIGATE program (a CSC program) or usual care. At 2-year follow-up, participants who received NAVIGATE remained engaged in treatment for a longer period, and demonstrated greater reductions in symptoms, greater improvement in quality of life, better interpersonal relationships, and more involvement in work/school. Outcomes were moderated by DUP, such that those with a shorter DUP (<74 weeks) benefited more from NAVIGATE than those with longer DUP (>74 weeks).12 Client and clinician family therapy manuals (and other resources from the NAVIGATE program) are available online: https:// navigateconsultants.org/manuals/.
+The RAISE Connection Program Implementation and Evaluation study enrolled 65 persons with FEP to receive a CSC program and demonstrated the feasibility of delivering CSC, including high rates of engagement.11 Another study, the Specialized Treatment Early in Psychosis (STEP) in Connecticut,13 demonstrated that those engaged in CSC, compared with usual or community care, were less likely to be hospitalized (40.0% in CSC compared with 63.1% in usual care), had significantly fewer inpatient bed days, and showed improvements in vocational engagement.14
+Across the World
+In the United Kingdom, the standard of care requires individuals with FEP be engaged with EISs within 2 weeks of the initial referral, or offered an assessment if considered to have an “At-Risk Mental State,” shown to reduce DUP.15 EISs in the United Kingdom have been linked to reduced hospital admission rates, lower relapse rates
+and symptom severity, and overall improved access to treatment.16 Superior effects of CSC programs also have been demonstrated in other countries in Europe, Canada, Australia, and Hong Kong.17-20
+COMPONENTS AND DELIVERY OF COORDINATED SPECIALTY CARE
+CSC is composed of pharmacologic management (once a month), individual psychotherapy (weekly),21 family psychoeducation,22 supported employment and education (SEE; weekly),23,24 case management (weekly), and when available, peer support services.25 Although there is some variability, CSC programs generally offer time-limited care over a period of 2 to 3 years. However, immediately stopping care at 2 years may be detrimental to an individual’s recovery, particularly if the patient has built a good therapeutic relationship.26 If treatment is required beyond those 2 years, the individual may step down to a lower level of care, with a transition into regular adult services.27
+CSC involves a multidisciplinary team ,and each team member has a distinct role. For example, the team psychiatrist/prescriber uses a shared decision-making approach in collaboration with the individual with FEP to identify the most effective and tolerable medication(s) at the lowest possible dose.28 Using an adapted Individual Placement and Support model of supported employment for serious mental illness,29 the SEE specialist works closely with the client to identify goals related to returning to work and school and provides support across all phases of the employment or education process. Case management focuses on providing resources for basic needs (eg, transportation, insurance) ideally using assertive outreach to promote engagement, respond to crises, or provide services when necessary.30 Although the peer support role is newer and therefore less well-defined, individuals with lived experience of mental health illness provide valuable support for individuals with FEP,31 for example, by increasing hopeful attitudes about recovery through sharing their own recovery story, providing support, and facilitating the client’s personal goals around community engagement (eg, exercise, going to coffee shops, or becoming more involved in extracurricular school activities). Team meetings are also key to optimal sequencing and coordination of treatment components based on the client’s goals. The team typically meets weekly for assessment and treatment planning and communicates closely with outside organizations to provide appropriate community support.
+The following section elaborates on the content, goals, and strategies used in individual, group, and family therapies for persons with FEP.
+INDIVIDUAL, GROUP, AND FAMILY THERAPIES
+The overarching objectives of group, individual, and family therapies in CSC are to 1. Help the client and family understand and cope with the experience of psychosis 2. Promote symptomatic and functional recovery and improve quality of life 3. Support the pursuit of personally meaningful goals of the client32
+A positive alliance not only helps to engage clients and families in therapy but is also related to improved symptoms and functioning in persons with FEP.33-36 As such, therapies delivered as part of CSC share several common elements in terms of their primary objectives and focus on promoting engagement and a strong alliance. Psychoeducation about psychosis and its course serves as the backbone for many of these therapies to empower clients and their families to make informed decisions both about treatment and other important aspects of clients’ lives (eg, returning to school/work). Further, given that engagement in treatment can be challenging, it is critical for therapists to prioritize developing and maintaining a strong therapeutic
+alliance with clients and families, which involves agreement on goalsand tasks of therapy, as well as the presence of a supportive bond.37 The use of reflective statements as well as an emphasis on collaboration, shared decision-making, and autonomy can foster a supportive bond as well as improved therapy engagement.38
+Individual Therapies
+Cognitive behavioral therapy for psychosis (CBT-p) aims to help clients understand and cope with symptoms, prevent relapse, and identify and work toward meaningful goals through an improved understanding of how thoughts and beliefs shape emotional reactions and behaviors in response to events.39-41 CBT-p has garnered substantial support for its use with persons with established schizophrenia and FEP.16,42,43 Although CBT-p is often delivered as an individual therapy, there is evidence that it also can be effectively delivered as a group psychotherapy.44,45 Therapists delivering CBT-p often use Socratic questioning techniques to explore clients’ understanding of their experiences and to help them identify stressors and vulnerabilities, and the stress-vulnerability model46 is often used as a framework to discuss precipitants of the initial psychotic episode as well as to identify protective factors to prevent relapse. This therapy includes both psychoeducation about psychosis and collaborative exercises aimed to help clients generate and test out alternative methods for coping with symptoms and appraising current past and present experiences, including the experience of psychosis. CBT-p is typically delivered as 16 weekly sessions over 6 months.21,47 Persons with FEP are encouraged to complete homework between sessions to promote continued understanding and practice of CBT-p exercises.43
+Individual Resiliency Training (IRT) served as the individual therapy component of NAVIGATE in the RAISE-ETP study and has been identified as a valuable intervention for persons with FEP.7,32,48 IRT is rooted in CBT-p, training in illness selfmanagement, and psychiatric rehabilitation. IRT is a manual-based therapy that emphasizes the enhancement of resiliency and strengths to support individuals’ pursuit of meaningful goals and to improve their illness management, social functioning, quality of life, and well-being. IRT draws from the structure of the Illness Management and Recovery program49 and earlier psychotherapeutic approaches for FEP emphasizing positive psychology.50 IRT contains 7 “standard” modules that are considered foundational for all persons with FEP in CSC as they help to frame the therapy, support the person in setting goals and preventing relapse, provide psychoeducation about psychosis, offer a structure to process the episode of psychosis, and promote resiliency. The standard modules cover the following:
+1.	Orientation
+2.	Assessment and goal-setting
+3.	Education about psychosis
+4.	Relapse prevention planning
+5.	Processing the episode
+6.	Developing resiliency: part one
+7.	Building a bridge to your goals
+IRT also contains 7 “individualized” modules that cover the following:
+1.	Dealing with negative feelings
+2.	Coping with symptoms
+3.	Substance use
+4.	Having fun and developing good relationships
+5.	Making choices about smoking
+6.	Nutrition and exercise
+7.	Developing resiliency: part two
+The decision to offer the content of the individualized modules is made collaboratively between the therapist and client based on the client’s personal goals.48 IRT is typically delivered on a weekly or biweekly basis for as long as needed (eg, delivered for 2 years in the RAISE-ETP trial7,32,48).
+Group Therapies
+Group-based therapies that target social cognition and social skills are effective in promoting functioning51,52 and negative symptoms53 among those with established schizophrenia. A few studies have examined their use in FEP populations,44,54,55 and these studies demonstrate considerable promise given the social cognitive difficulties that persons with FEP experience.56 Social Skills Training (SST57) is an evidence-based intervention that focuses on helping individuals learn and practice skills involved in social interactions (eg, making requests, expressing positive feelings). SST groups often include a discussion of the rationale for a skill, specific steps of the skill, role-play exercises, feedback from the group, and homework assignments. This intervention has been shown to help individuals learn and practice social skills within the group setting and, subsequently, use them effectively in the community. Number of sessions per week and total weeks depend on the needs of the clients and the setting in which it is delivered (eg, mean number of weeks = 19.3 with a range of 2-104 weeks reported in one study).52
+Cognitive enhancement therapy (CET58) has shown benefits in schizophrenia and in early psychosis. CET is composed of computer training (focused on attention, memory, and problem-solving) group therapy (focused on perspective-taking, managing emotions, reading nonverbal cues, and interpreting social situations).59 CET typically consists of 60 hours of computer training and 45 weekly group sessions. This integrated intervention has been shown to improve social cognition and neurocognition in those with established schizophrenia60 and in those with FEP.61
+Stand-alone social cognition training interventions aim to improve individuals’ capacity to understand, interpret, and use social information effectively, often targeting one or more of the primary domains of social cognition: theory of mind, emotion perception, social perception, and attributional style.62 For example, Social Cognition Interaction Training (SCIT), has been examined extensively in established schizophrenia and has been piloted in a sample of persons with FEP.63 SCIT is delivered as a 20-session to 24-session group psychotherapy typically delivered weekly64 that includes 3 phases: emotion training (eg, identifying emotions from photos of faces, relationship between emotions and thoughts), figuring out situations (eg, distinguishing between facts and guesses in social situations), and integration (discussion of how information can be applied to salient situations). Individuals learn effective social cognitive strategies, practice them within the groups, and ultimately use them in everyday interactions.
+Family Therapies
+Historically, family interventions have been underutilized in the treatment of individuals with schizophrenia,65 despite their clear benefit in reducing relapse and rehospitalization.66-68 Over recent years, however, family interventions have occupied a more central role in the treatment of FEP.12 Family intervention, as part of CSC, typically includes education, validation of the impact of psychosis on the family, communication, problem-solving, and goal-setting skills training.
+Family education about psychosis and its optimal management serves a number of purposes:
+1.	Developing a shared language for the treatment team, individual, and the family to talk about psychosis and associated symptoms
+2.	Providing information so that individuals with psychosis and their families can make informed choices about illness management
+3.	Orienting the family to how they can support the management of their relative’s illness and pursuit of personal goals
+Educational topics include information about psychosis and associated symptoms, the stress-vulnerability model of psychosis, diagnosis and prognosis, the role of the family in treatment, early warning signs monitoring, and relapse-prevention planning. Family education provides the opportunity for family members to observe how the clinician talks to the individual with psychosis about symptoms, diagnosis, treatment, and recovery.
+Families are often put under tremendous stress due to the disruptions in the family system that result from an episode of psychosis. A key underlying aspect of family interventions is the validation of this stress for the entire family system. Before the onset of the psychosis, the young adult may have been living independently, such that the onset of an illness represents a shifting of roles and worry for the entire family system. Communication, problem-solving, and goal-setting skillstraining can be important for families during this period of adjustment and heightened stress. Communication skills are aimed at reducing stressful interaction styles characterized by strong displays of negative affect or ambiguous messages and emphasize the use of direct “I statements,” reference to specific behaviors, and specific feeling statements taught using the principles skills training (eg, modeling, role playing). Common targets for communication include medication, symptoms, and disclosure of information about the illness with the immediate, as well as the extended, family. In addition, it may be important for family members to reestablish how they will make requests of one another, which dovetails with the question of reasonable expectations of the individual with psychosis during the immediate period following illness onset and beyond.
+Fostering problem-solving and goal-setting skills in the family serves the dual purpose of minimizing strife and facilitating recovery through each family member’s identification of meaningful goals. Early in treatment, families often work toward the goal of increasing shared pleasant activities, which can increase family connection, shift the focus from illness to enjoyment and fun, and help remediate negative symptoms and demoralization that are commonly associated with the experience of psychosis. Later in treatment, families often take on more challenging goals, such as assisting the young person in returning to school or work, living independently, or traveling for educational or leisure purposes.
+Multifamily group (MFG) interventions typically include 5 to 7 families who meet with 2 clinicians on a biweekly basis,69 following “joining” sessions in which each family meets individually with the clinician to form a relationship and provide information about their family’s specific needs. Each MFG session lasts approximately 90 minutes and the content of the sessions map onto 4 treatment stages corresponding to the phases of an episode of psychosis: (1) engagement between client and their family, (2) education about the psychotic disorder, (3) development of strategies, such as stress reduction, to cope with the challenges of psychosis recovery, and (4) social and vocational rehabilitation.69 Elements of MFG considered to be particularly effective include access to a social network, reduction in perception of stigmatization, availability of mutual aid, and the opportunity to hear similar experiences and
+solutions.69 Although there can be some initial challenges with establishing a critical mass of families willing to attend a group, MFG is cost-effective70 and has been demonstrated to increase perception of ability to cope with a relative’s psychosis,71 and reduce FEP program dropout rates.72
+CLINICAL CHALLENGES
+Individuals with FEP vary tremendously from one another in terms of the severity of positive symptoms, negative symptoms, and cognitive and social functioning. The most significant challenges in therapy with individuals with FEP arise from the need to adapt the treatment to each individual’s specific symptom presentation and understanding of his or her problems. Problematic substance use73 and history of trauma or posttraumatic stress disorder74 also add complexity to the treatment of some individuals. Furthermore, significant stressors beyond coping with FEP (eg, limited income, transportation barriers, homelessness) can interfere with the feasibility of delivering treatment and, thus, should be considered when trying to engage and maintain persons in therapy. In these instances, mobile teams and/or case management supports (eg, transportation paid for by health insurance or access to disability pay-ments75) are essential ingredients to involve the individual and family in care and reduce strain on poorer families.
+Negative symptoms, cognitive deficits, and impaired social and occupational functioning tend to co-occur in primary psychotic disorders and are defining features of FEP. Clinicians may struggle to engage individuals with negative symptoms and families may blame these individuals for being “lazy” or “unmotivated,” which can amplify familial stress and impede recovery. Education about negative symptoms, spending more time getting to know the individual (eg, befriending techniques76,77), and slowing down the pace of therapy as well as breaking goals into small steps can be useful. An important part of the educational process involves dispelling the myth that negative symptoms indicate a lack of distress, because in fact, individuals with negative symptoms are often bothered by these symptoms and this is related to poor quality of life.78 Therefore, recognizing and labeling this distress can serve as a rationale to build coping skills for these symptoms. Another important discovery has been the identification of common dysfunctional beliefs expressed by individuals with negative symptoms,79,80 such as beliefs about self-efficacy (eg, “I don’t have enough energy or I don’t have anything to say”) and anticipatory pleasure (eg, “I won’t have a good time”), which are thought to impair effortful responding and can be addressed through cognitive restructuring and behavioral experiments.81,82 Further, given the variability in cognitive functioning among persons with FEP (eg, due to age, effects of medication/ electroconvulsive therapy, symptoms), psychosocial interventions should be appropriately tailored to the cognitive capacity of each individual.
+Another challenge when working with persons with FEP and their families is sensitively and effectively addressing the role of trauma in therapy. Many persons with FEP have had traumatic experiences in their lives,83 which may have been associated with the experience of psychosis and psychiatric treatment (eg, involuntary hospitalization, coercive treatment, use of restraints, and/or police involvement).84 IRT includes a module called “processing the episode,” which aims to help individuals integrate, process and understand the trauma of experiencing psychosis, and interventions designed to facilitate cognitive processing of traumatic experience in FEP have shown positive outcomes.85
+Cultural and religious factors can also impact the willingness of clients and families to engage in treatment. For example, individuals of some cultural and religious
+backgrounds may not believe that psychological or psychiatric medication approaches to treatment are appropriate and may seek out alternative options (eg, shaman, exorcism, religious practices). Therapists should try to work within the cultural context of the given client and family to best support the recovery of the person with FEP. Therapists should use a curious attitude about these alternative approaches and better understand how they fit within the cultural context of the family and, importantly, assess any potential risk for the person. However, people may also be open to alternative explanations of their experience, especially when they are less distressing or more helpful. Therapists may also offer an “open door” policy so that individuals and families know that they are welcome to reconnect in the future.
+SUMMARY
+CSC programs provide team-based, comprehensive, evidence-based care, education, and support across 2 to 3 years to individuals experiencing their first episode of psychosis and their families. A collaborative clinical approach within CSCs are important. Individual, group, and family therapies represent critical aspects of CSC as they are aimed at helping individuals with FEP and their families navigate the distressing experience of psychosis and to promote recovery and well-being. Several individual (eg, CBT-p, IRT), group (eg, SST, SCIT, CET), and family (eg, family psychoeducation) therapies have demonstrated benefits for this population and are guided by the individual’s goals and long-term vision of recovery. However, there are many clinical challenges that often accompany FEP therapy delivery that warrant significant attention. Therapists should be aware of these challenges and develop strategies to engage and maintain clients and their families in therapy. It is through awareness of challenges, prioritization of the therapeutic alliance, and effective delivery of evidence-based therapies that therapists can help clients and their families work toward recovery.

Excess-mortality-from-mental-neurological-and-substance-use-disorders-in-the-Global-Burden-of-Disease-Study-2010Epidemiology-and-Psychiatric-Sciences.txt

@@ -0,0 +1,95 @@
+Introduction
+Findings from the Global Burden of Disease 2010 (GBD 2010) study have reinforced our understanding of the significant impact that mental, neurological and substance use disorders (MNSDs) have on population health (Murray et al. 2012; Whiteford et al. 2013). One of the key findings of GBD 2010 was the global health transition from communicable to noncommunicable diseases and this is particularly rapidly in low- and middle-income countries (LMICs) (Murray
+et al. 2012). The proportion of burden attributable to non-communicable disease in LMICs has risen by more than one-third, from 36% in 1990 to 49% in 2010. In contrast, the share of non-communicable disease burden in high-income countries (HICs) has raised only 3-4% over the same time period (from 80 to 83%) (Institute of Health Metrics and Evaluation, 2013). These findings hold particular public health importance for MNSDs in LMICs for the coming decades.
+GBD 2010 estimates the majority of disease burden due to MNSDs is from non-fatal health loss; only 15% of the total burden is from mortality, in terms of years of life lost (YLLs) (Institute of Health Metrics and Evaluation, 2013). This may erroneously lead to the interpretation that premature death in people with mental and neurological disorders is inconsequential, whereas evidence shows that people with MNSDs experience a significant reduction in life
+expectancy (Chang et al. 2011; Wahlbeck et al. 2011; Crump et al. 2013; Lawrence et al. 2013). In Australia and the UK males with a mental disorder die, on average, 15 years earlier than the general population and females die on average 12 years earlier (Crump et al. 2013; Lawrence et al. 2013). It is estimated that about 80% of premature deaths in people with MNSDs are due to physical illnesses, particularly cardiovascular disease, including stroke and cancer (Crump et al. 2013; Lawrence et al. 2013). Dementia is an independent risk factor for premature death with increased risk found in those patients with physical impairment and inactivity, and medical comorbidities (Park et al. 2014). Excess mortality in people with epilepsy is reported to be two- to threefold higher compared with the general population; with an increased risk of up to sixfold higher in LMICs (Diop et al. 2005). A significant proportion of these deaths are preventable, resulting from falls, drowning, burns and status epi-lepticus (Diop et al. 2005; Jette & Trevathan, 2014). A recent review has shown the highest standardised mortality ratio (SMR) among mental and substance use disorders was 14.7 for opioid use disorders (Chesney et al. 2014). In HICs the life expectancy gap is widening with the general population now enjoying a longer life while the lifespan for those with a mental disorder has remained static (Lawrence et al. 2013).
+Mortality associated with a disease can be quantified using two different, yet complementary, methods which are employed as part of GBD analyses. First, cause-specific mortality draws upon vital registration systems and verbal autopsy studies which identify deaths attributed to a single underlying cause using the International Classification of Diseases (ICD) death-coding system. Second, GBD creates natural history models for each disease, including its distribution across age and sex. This involves estimation of a range of epidemiological parameters, including excess mortality - that is, the all-cause mortality rate in a population with the disorder compared with the all-cause mortality rates in a population without the disorder. By definition, estimates of excess deaths include causespecific deaths.
+Although often arbitrary, the ICD conventions are a necessary attempt to deal with the multi-causal nature of mortality and avoid 'double-counting' of deaths. However, despite the systems clear strengths, causespecific mortality estimated via the ICD obscures the contribution of other underlying causes of death; for example, suicide as a direct result of major depressive disorder coded as injury, and will likely underestimate the true number of deaths attributable to a particular disease. On the other hand, estimation of excess mortality using natural history models will often comprise deaths from both causal and non-causal origins and
+will likely overestimate the true number of deaths attributable to a particular disorder. The challenge is to parse out causal contributions to mortality (beyond those already identified as cause-specific) from the effects of confounders.
+Quantification of the contributions of multiple causal factors to excess mortality associated with a particular disease is challenging and requires approaches such as the comparative risk assessment (CRA), which is now an integral part of the GBD studies. The fundamental approach for the GBD CRA is to calculate the proportion of deaths or disease burden caused by specific risk factors - e.g., lung cancer caused by tobacco smoking - while holding all other independent factors unchanged. A key concept when attempting to quantify causal relationships is that of 'counterfactual burden' which compares the burden associated with an outcome with the amount that would be expected in a hypothetical situation of 'ideal' risk factor exposure (e.g., zero prevalence). This approach provides a consistent method for estimating the changes in population health as a function of decreasing or increasing the level of exposure to risk factors (Lim et al. 2012). Importantly, the flexibility within counterfactual analysis allows the sum of death counts attributed to different risk factors for a particular cause to sum to more than 100% which is not permissible by ICD registry data.
+In this paper, we explore the cause-specific and excess mortality of individual MNSDs estimated by GBD 2010. We also present the additional attributable burden that can be ascribed to disorders using GBD results for CRA's assessing MNSDs as risk factors for other health outcomes. Disorders included in the analyses are grouped by: mental disorders (schizophrenia, major depression (MDD), anxiety disorders, bipolar disorder, childhood behavioural disorders (attention-deficit/hyperactivity disorder (ADHD) and conduct disorder (CD)), autistic disorder and intellectual disability); substance use disorders (alcohol, opioid, cocaine and amphetamine use disorders); and neurological disorders (dementia, epilepsy and migraine).
+Methods
+YLLs and cause of death
+The GBD 2010 methodology uses a time-based metric, YLLs, to quantify the fatal burden by underlying cause (Lozano et al. 2012). YLLs are computed by multiplying the number of deaths attributable to a particular disease at each age by a standard life expectancy at that age. The standard life expectancy represents the normative goal for survival and for 2010 was computed based on the lowest recorded death rates in
+any age group in countries with a population greater than 5 million (Salomon et al. 2012).
+Cause-specific death estimates in GBD 2010 were produced from available cause of death data for 187 countries from 1980 to 2010. Data sources included vital registration, verbal autopsy, mortality surveillance, censuses, surveys, hospitals, police records and mortuaries (Lozano et al. 2012). Cause of death ensemble modelling (CODEm) was used for all MNSDs. In summary, CODEm uses four families of statistical models testing a large set of different models using different permutations of covariates. Model ensembles were developed from these component models and model performance was assessed with rigorous out-of-sample testing of prediction error and the validity of 95% uncertainty intervals (UIs). Details relating to CODEm and the method for how these models were used in calculating YLLs are described in detail elsewhere (Lozano et al. 2012).
+YLLs for GBD 2010 were computed from causespecific mortality estimates for seven of the 15 MNSDs: schizophrenia; opioid, amphetamine, cocaine and alcohol use disorders; dementia and epilepsy (Lozano et al. 2012). As the ICD does not permit the other mental and neurological disorders to be recorded as the 'primary' cause of death, YLLs were unable to be calculated for the remaining eight disorder groups (World Health Organization, 1993; Lim et al. 2012).
+Excess mortality from a natural history model
+Drawing on a series of systematic reviews, we collated comprehensive sets of epidemiologic data for each disorder. Data were pooled, adjusting for between-study variance, and then an internally consistent epidemiologic model derived using the relationship described in the generic disease model (see the appendix) (Vos et al. 2012). To do this we used DisMod-MR, a Bayesian meta-regression tool which estimates a natural history of disease model producing age-, sex- and region-specific estimates for prevalence, incidence, remission and excess mortality (Vos et al. 2012). Where data were scarce, DisMod-MR was able to impute information with associated uncertainty ranges based on epidemiologically and geographically similar populations. Excess mortality estimates based on this natural history model are reported for MNSDs in terms of global deaths for 2010. Further details of the GBD 2010 methods for developing a natural history model of disease using DisMod-MR have been described in detail elsewhere (Vos et al. 2012).
+Counter-factual burden and CRA
+Prince et al. (2007) have summarised the evidence where a causal relationship between mental and
+substance use disorders and other health outcomes have been proposed. In GBD 2010, a series of reviews were conducted to assess the strength of evidence for MNSDs as independent risk factors for other health outcomes (Degenhardt et al. 2009a; Rehm et al. 2010a; Charlson et al. 2011; Degenhardt & Hall, 2012). Risk factor studies were identified through systematic searches of published and unpublished data with information on effect sizes and study characteristics extracted and collated (Charlson et al. 2013; Degenhardt et al. 2013; Ferrari et al. 2014). A metasynthesis was used to calculate a relative risk (RR) for MNSDs (the exposure) as a risk factor for other health outcomes. The RR was then applied to prevalence distributions of the specific exposures by sex and age-group for each geographic region to derive population attributable fractions (PAFs). More detail on the calculation of PAFs in GBD 2010 is provided by Lim et al. (2012). In some cases, for example suicide, ceiling values were calculated and applied for joint PAFs to ensure the sum of proportional contribution for all risk factors did not exceed 100% (Ferrari et al. 2014). The additional burden (YLLs and YLDs) attributable to MNSDs is the product of the PAFs and the burden for the health outcome as estimated in GBD 2010.
+Here we compare the number of cause-specific deaths reported (YLLs) for MNSDs, calculated as part of the GBD 2010 study with the number of allcause deaths derived from natural history models. We explore differences in these estimates across the age-span for each disorder. Additional YLLs attributed to disorders as underlying causes and quantified in the CRA are reported.
+Results
+YLLs and causal mortality
+Globally, the seven disorders (dementia, epilepsy, schizophrenia, alcohol use disorders, opioid use disorders, cocaine use disorders and amphetamine use disorders) for which YLLs were estimated were directly responsible for 1.3 million deaths in 2010, equating to about 12 million YLLs (see Fig. 2 in the appendix). Epilepsy and then dementia contributed the greatest proportion of YLLs within this group.
+Age-standardised YLL rates vary considerably across the seven geographical super-regions primarily due to differences in patterns of alcohol and drug use, and mental and neurological disorder prevalence. There are several regions with substantial deviations from global YLL average rates (Fig. 1). (Details of which countries are in each sup-region can be found
+on the IHME website (Institute for Health Metrics and Evaluation, 2014).)
+In 2010, YLL rates were highest in the sub-Saharan Africa region (604 YLLs per 100 000 population) and the region comprising Central/Eastern Europe plus Central Asia (593 YLLs per 100 000); the causes for which vary considerably (Fig. 1). In sub-Saharan Africa the YLL burden was driven by epilepsy which was fourfold higher than the global average and approximately 85% of all YLLs attributed to MNSDs in sub-Saharan Africa. Although the substance use disorders YLL rates appear unremarkable for this region, their YLL burden has increased 3.0% from 1990 to 2010, almost double the average global increase and the highest of all regions (Degenhardt et al. 2013). In contrast to sub-Saharan Africa, the high fatal burden in Central/Eastern Europe and Central Asia was largely caused by deaths attributed to alcohol use. High mortality due to illicit substance use disorders also contributed to the YLL rate in Central/Eastern Europe and Central Asia with all substance use disorders together explaining 73% of YLLs in the region. Countries within East Asia and the Pacific exhibit very low YLL rates across all MNSDs with little change observed between 1990 and 2010.
+Globally, neurological disorders accounted for 58% of the all MNSD YLLs in males and 81% of YLLs in females. Substance use disorders explained 39% of YLLs in males and 16% of those in females. The contribution of schizophrenia to total MNS disorder YLLs was similar for both genders (3% each). Differences in YLL patterns between the genders were influenced in part by the differing contribution to YLLs by substance use disorders compared with neurological disorders across regions (Fig. 2).
+Excess mortality from a natural history model
+The only mental disorder for which cause-specific deaths and YLLs were estimated in GBD was schizophrenia; however, several mental disorders, such as major depression and bipolar disorder, exhibit significant and documented excess-mortality (Roshanaei-Moghaddam & Katon, 2009; Baxter et al. 2011b) (Table 1). There were four disorders for which sufficient evidence of excess all-cause mortality could not be found in the literature (anxiety, childhood behavioural disorders, cannabis dependence and migraine) and therefore excess mortality was not included in the natural history of disease for these disorders.
+Mental disorders
+Figure 3 shows the estimated number of cause-specific and excess deaths for each of the five mental disorders with estimated excess mortality by age and with uncertainty bounds. While cause-specific deaths were attributed to only one mental disorder (schizophrenia),
+excess mortality was present in natural history models for five: schizophrenia, bipolar disorder, MDD, autistic disorder and intellectual disability.
+Although schizophrenia is one of the few mental disorders with cause-specific deaths permissible by ICD,
+the numbers of cause-specific deaths globally (approximately 20 000) are noticeably lower compared with allcause deaths (approximately 700 000) ascribed by the disorder’s natural history. Around 1.3 million excess deaths are estimated in the natural history model of
+bipolar disorder but there are no cause-specific deaths attributed to the disorder. The natural history of the disease suggests, however, that bipolar disorder is associated with a higher number of excess deaths globally than schizophrenia. No deaths were coded to depressive
+disorders in GBD 2010. Natural history models of MDD suggest there were more than 2.2 million excess deaths in persons with MDD, with a particularly high rate of death in older persons that is not observed in schizophrenia or bipolar disorder.
+Intellectual disability was modelled as an ‘envelope disorder for GBD 2010, meaning that the intellectual disability ascribed to all underlying causes including meningitis, Down’s syndrome, and chromosomal defects, were captured under a single disorder category. After modelling, the contributions of each specific underlying cause were separated out and a ‘rest’ category of idiopathic intellectual disability was created. There were no deaths causally attributed to intellectual disability; however, excess deaths in people with idiopathic intellectual disability were estimated to be substantial at over 900 000 deaths globally in 2010.
+At this time there is insufficient information available to determine whether premature mortality is significantly raised across the spectrum of anxiety disorders (Baxter et al. 2014) and in childhood
+behavioural disorders (Erskine et al. 2013). In GBD 2010 there were no YLLs or excess mortality associated with the natural history of disease applied to anxiety disorders or childhood behavioural disorders.
+Substance use disorders
+GBD estimates indicate more than 110 000 deaths were causally attributed to alcohol use disorders worldwide in 2010, but indicative of the true impact of alcohol dependence as an underlying cause of death in many is that over 5 million excess deaths were estimated in the same year. Over 700 000 excess deaths occurred in dependent illicit drug users in 2010 compared with only 44 000 deaths which were coded as the cause of death. The majority of these deaths can be ascribed to opioid dependence (43 000) (Fig. 4).
+Neurological disorders
+Cause-specific death estimates are more substantial for neurological disorders (Fig. 5) resulting in a less
+dramatic gap between cause-specific and excess deaths. This is likely indicative of neurological disorders being recognised more readily as the primary cause of death.
+Similar to intellectual disability, epilepsy was modelled as an envelope disorder in GBD 2010 with idiopathic epilepsy and epilepsy secondary to a range of causes, including meningitis, neonatal tetanus, iodine deficiency and a variety of birth complications, being modelled as one disorder. Cause of death modelling estimated nearly 180 000 deaths due to epilepsy in 2010 while natural history models show us about 300 000 excess deaths in fact took place. The proportion of deaths attributable to different causes differ by region and GBD 2010 showed sub-Saharan African populations had the highest death rate due to epilepsy. Around 2.1 million excess deaths worldwide were estimated from dementia for 2010, yet less than 500 000 were attributed to dementia as the primary cause of death.
+Table 2 shows that the cause-specific deaths and excess deaths directly coded to MNSDs are relatively similar up to 4 years of age but then rise sharply: in children aged 5-9 years there were 7420 cause-specific deaths compared with more than 91 000 excess deaths in the same age group. Alcohol use disorders explained the highest number of excess deaths (5.2 million): 38% of all excess deaths due to mental and neurological disorders in 2010. Considered together, the mental disorders for which no cause-specific deaths were attributed (bipolar disorder, major depression, autism and intellectual disability) explained more
+than 4.5 million deaths, equating to one third of all excess deaths in 2010.
+Counter-factual burden and CRA
+The reviews conducted as part of GBD 2010 collectively yielded sufficient evidence for several CRAs (see Table 3). Neurological disorders were not assessed as risk factors in GBD 2010.
+For mental disorders, a number of associations were investigated but data limitations meant that only suicide and IHD were able to be included as outcomes for mental disorders (suicide) and MDD (IHD) (Baxter et al. 2011a; Charlson et al. 2011; Ferrari et al. 2014). Collectively, mental and substance use disorders are estimated to be responsible for about 22 million YLLs due to death by suicide (Ferrari et al. 2014). The CRA of major depression as a risk factor for ischaemic heart disease estimated an attributable to burden of about 3.5 million YLLs (Charlson et al. 2013).
+Injecting drug use was considered as a risk factor for a number of outcomes, including blood borne viruses and liver disease, and collectively accounted for over 7 million YLLs in attributable burden. Interestingly, and despite common preconceptions, GBD results did not show any mortality-related burden from schizophrenia that could be attributed to cannabis dependence. Alcohol use was the biggest contributor with nearly 80 million attributable YLLs estimated across a number of health outcomes.
+Figure 6 shows the considerable additional burden when MNSDs are considered as underlying
+contributors to other health outcomes. Given the large estimate of mortality-related burden attributed to alcohol dependence, it is expected that, when aggregating YLLs, the regions with the largest attributable burden will be those which have highest rates of alcohol dependence, i.e., Eastern Europe and Central Asia. Sub-Saharan Africa experiences large communicable disease YLLs attributable to alcohol dependence as a
+result of the continuing high prevalence of communicable disease in relation to other regions.
+By incorporating the additional YLLs estimated using CRAs into the overall contribution of mental, substance use and neurological disorders to all cause YLLs (Table 4) we can see a dramatically different picture to that painted in Appendix 1 where YLL contributions appeared negligible in many cases.
+Contributions across regions vary in accordance with the epidemiological profile of disorders within each region, not only of mental, substance use and neurological disorders but also the health outcomes assessed in CRAs. For example, the relatively large contribution in HICs and Central/Eastern Europe and Central Asia likely to be reflective not only of high prevalence of substance use, but also of cardiovascular disease (CVD) which was assessed as an outcome of major depression and alcohol use. In contrast, the relatively lower contribution in sub-Saharan Africa is likely reflective of comparatively lower rates of both substance use disorders (risk factors) and chronic diseases such as CVD (health outcomes). If neurological disorders were assessed as risk factors for other health outcomes using CRAs this picture may have looked different for sub-Saharan Africa where YLLs attributable to neurological disorders is higher than in other regions in the world.
+Discussion
+A relatively small YLL burden was attributable to MNSDs in GBD 2010; however, numbers of excess deaths derived from natural history of disease models clearly demonstrate the high degree of mortality associated with these disorders. Quantifying the independent contributions of mental and substance use disorders to poor health outcomes through methods such as the CRA is restricted by data availability and methodological challenges such as establishing causal relationships (Baxter et al. 2011a); nevertheless, there is a growing body of literature which can help us develop hypotheses around these contributions by observing the risks associated with excess death in individuals with mental and substance use disorders.
+The relationship between mental disorders and suicide has long been recognised (Li et al. 2011). Mental disorders have also been linked to higher rates of
+death due to coronary heart disease, stroke, type II diabetes, respiratory diseases and some cancers (Hoyer et al. 2000; Crump et al. 2013). The relationship between mental disorders and physical disease, leading to premature death, is complex. People with mental disorders have an increased risk of death in several ways, for example people with MDD are more likely to develop CVD (Charlson et al. 2011). Psychotropic medications can negatively impact on cardiovascular and metabolic health (De Hert et al. 2012). Obesity and metabolic disturbances are primary risk factors for CVD and type II diabetes, and are two- to threefold more common in people with mental disorders compared with the general population (Scott & Happell, 2011). Major modifiable risk factors for chronic disease, such as smoking (Lawrence et al. 2009), poor diet and physical inactivity (Kilbourne et al. 2007; Shatenstein et al. 2007) and substance abuse (Scott & Happell, 2011), are overrepresented in people with mental disorders and these may be consequences of symptoms of MNSDs, medication effects and poor emotional regulation (Scott et al. 2013).
+Interestingly, while schizophrenia was the condition among these mental disorder among for which YLLs were attributed, the number of YLLs were very small compared with the excess mortality associated with the disorder. Our finding of high excess mortality in people with schizophrenia is in line with that found in the previous research (Laursen, 2011; Crump et al. 2013; Lawrence et al. 2013). Data linkage studies have shown that the majority of deaths in people with schizophrenia are due to chronic disease with CVD accounting for more than one-third of all premature deaths, while unnatural causes, including suicide, homicide and accidents account for just under 15% of excess deaths (Crump et al. 2013; Lawrence et al. 2013). Despite concerns over the side-effects of antipsychotic medication, lack of antipsychotic treatment has been linked with higher all-cause mortality rates (HR 1.45, 95% confidence interval (CI) 1.20-1.76),
+132 F. J. Charlson et al.
+with highest risks attributed to cancer (HR 1.94, 95% CI 1.13-3.32) and suicide (HR 2.07, 95% CI 0.73-5.87; Crump et al. 2013). Poly-pharmacy and discontinuation of medication also appear to increase risk of allcause death (Joukamaa et al. 2006; Haukka et al. 2008).
+Research from the UK suggests that the excess mortality rate in schizophrenia and bipolar disorder are comparable (Chang et al. 2011). In a recent study, it was estimated that about 80% of premature death in people with bipolar disorder is due to physical disease, almost half of which is explained by CVD (Westman et al. 2013). Just under 20% of premature deaths were explained by unnatural causes (suicide, homicide and unintentional injuries; Westman et al. 2013).
+People with developmental disorders are at twice the risk of premature death compared with the general population (Mouridsen et al. 2008). Elevated death rates in autistic spectrum disorders (ASD) are due to several causes, including accidents, respiratory diseases and seizures (Shavelle et al. 2001; Mouridsen et al. 2008). The elevated mortality risk associated with ASD may be due more to the presence of comorbid medical conditions, particularly epilepsy, and intellectual disability rather than ASD itself (Lee et al. 2008; Bilder et al. 2013).
+Individuals with intellectual disability are expected to have, on average, a life expectancy of 7-12 years less than the general community and life expectancy is dramatically lower in those more severe disability and those with a genetic disorder (e.g., Down syndrome) (Bittles et al. 2002). Intellectual disability is associated with greater tendency towards obesity and physical inactivity compared with the general population, and enhanced predisposition to mental disorders, osteoporosis, thyroid disorders, non-ischaemic heart disease and early onset of dementia (Bittles et al. 2002). In HIC, causes of death in people with intellectual disability are generally coded under congenital abnormalities, diseases of the nervous system and sense organs, mental disorders and respiratory disease (Tyrer & McGrother, 2009). Information on causes of death in LMIC populations is sparse.
+Children with ADHD or CD are two to three times more likely to experience unintentional injuries requiring medical attention compared with children without behavioural disorders (Rowe et al. 2004; Lee et al. 2008). The injuries most commonly reported included burns, poisoning and frac(Rowe et al. 2004). Adolescents and young adults with inattention disorders are more likely to be involved in traffic accidents (Jerome et al. 2006). Adults who were identified with behavioural disorders in childhood are at higher risk of cigarette smoking, binge-drinking (ADHD) and obesity (CD) (von Stumm et al. 2011) in later life. Despite the strong evidence for an association between childhood
+behavioural disorders and poorer health outcomes, there is insufficient information available to model the natural history of disease and thus no estimates quantifying excess mortality in this group at population level.
+Another important disorder demonstrating an apparent absence of excess-mortality in GBD 2010 is the umbrella anxiety disorders group. This was a necessary choice as the information on excess mortality in anxiety disorders was found to be inconsistent with some anxiety disorders; however, severe presentations such as post-traumatic stress disorder (PTSD), have previously been associated with increased deaths caused by ischaemic heart disease (IHD), neoplasms and intentional and unintentional injuries (Ahmadi et al. 2011; Lawrence et al. 2013).
+While light-to-moderate alcohol consumption has been associated with lower rates of some disease such as diabetes mellitus and coronary heart disease, heavy consumption has been associated with increased rates of chronic disease, including cancer, MNSDs, cardiovascular disease, liver and pancreas diseases (Rehm et al. 2010a). There is evidence for alcohol as a carcinogen in humans, with particularly strong causal links established between alcoholic beverage consumption and oral cavity, pharynx, larynx, oesophagus, liver, colorectal and female breast cancers (Rehm et al. 2010a). A consistent relationship has also been found between heavy alcohol consumption and epilepsy (Rehm et al. 2010a) and it is also implicated in development of depression and personality disorders, although the direction of causality and effect of confounding factors remains uncertain (Rao et al. 2000; Rohde et al. 2001). Risk of diabetes mellitus, hypertension, stroke, sudden cardiac death and other cardiovascular outcomes is elevated in those with alcohol use disorders (Rehm et al. 2010a). The relationship between alcohol consumption and liver cirrhosis is well recognised, but alcohol use disorders appear more strongly related to cirrhosis mortality v. morbidity as it negatively affects the course of existing liver disease (Rehm et al. 2010b). Heavy alcohol use is also related to higher rates of infectious diseases, such as tuberculosis, and unintentional and intentional injury, with strong evidence for a dose-response relationship (Rehm et al. 2010a).
+Excess and premature deaths in illicit drug users occur in several ways. Most obvious is the acute toxic effects of illicit drug use which may lead to overdose - the cause-specific deaths generally captured by the ICD-coding system. In addition, a substantial number of deaths are likely due to the more indirect effects of intoxication resulting in accidental injuries and violence. There are a plethora of adverse health outcomes with elevated risks of premature mortality for which illicit drug dependence is an important contributor. These outcomes are often chronic and include
+cardiovascular disease, liver disease and a range of mental disorders including psychosis. Suicide is an important outcome, particularly for opioid users where an SMR of approximately 14 has been reported in two separate reviews (Degenhardt et al. 2011; Chesney et al. 2014). Injection of drugs, most common in opioid dependence, carries a high risk of bloodborne bacterial and viral infections, notably HIV, Hepatitis B and Hepatitis C (Mathers et al. 2010; Nelson et al. 2011).
+Epilepsy is associated with two- to threefold higher than mortality in the general community, particularly in childhood onset epilepsy, with the highest standardised mortality ratio encountered in the first year or two after diagnosis (Preux & Druet-Cabanac, 2005; Sillanpaa & Shinnar, 2010; Neligan et al. 2010; Trinka et al. 2013). Common causes of premature mortality in epilepsy include acute symptomatic disorders (e.g., brain tumour and stroke), sudden unexpected death, suicide and accidents (Hitiris et al. 2007). Roughly 85% of people with epilepsy live in LMICs and here the risk of premature mortality is highest (Carpio et al. 2005; Diop et al. 2005; Newton & Garcia, 2012; Jette & Trevathan, 2014) from status epilepticus, drowning and burns associated with poor access to and/or compliance with medical treatment, cognitive impairment and age (Jilek & Rwiza, 1992; Kamgno et al. 2003; Mu et al. 2011; Ngugi et al. 2014).
+As with mental disorders, excess mortality in dementia has been associated with functional disability leading to lifestyle factors (e.g., poor eating behaviours, physical inactivity and poor hygiene) and comorbid or underlying physical conditions, including cardiovascular disease, diabetes mellitus and neoplasms (Guehne et al. 2005; Llibre Rodriguez et al. 2008). Infections, particularly pneumonia and the complications of urinary tract infections, frequently lead to death in people with dementia (Mitchell et al. 2009).
+Strategies for reducing mortality associated with MNSDs are primarily related to preventing onset of disorders, reducing case fatality, and preventing onset of fatal sequela. There is growing evidence that excess mortality in people in mental and substance use disorders can be reduced through existing evidence-based treatments and improved screening and treatment for chronic disease. There is some evidence that collaborative care by community-based health teams has the potential to reduce overall death as well as suicide deaths (Malone et al. 2007; Dieterich et al. 2010). The use of collaborative care models to improve physical health in people with MNSDs is growing in developed countries and these have demonstrated a range of positive health outcomes including reduced cardiovascular risk profiles (Druss et al. 2010). The effectiveness of these strategies
+in preventing premature mortality in LMIC populations has yet to be tested but this may be a costeffective approach to treatment where trained mental health clinicians are scarce.
+To improve life expectancy in people with comorbid mental and physical health issues requires proactive screening and adequate care for chronic disease. Screening and prevention of metabolic risk factors is essential. Strategies for early cancer detection should be prioritised and models of care developed to ensure that people with MNSDs receive the same level of physical health care and treatment as the rest of the population.
+Psychiatric treatments, specifically pharmacotherapies, may have some protective effect against excess mortality (Weinmann et al. 2009) although evidence suggests that this depends on use of medications according to best practice guidelines (Cullen et al. 2013). In contrast, some second generation antipsychotics may actually pose an elevated risk mediated by metabolic side effects (Newcomer, 2005; Smith et al. 2008; Rummel-Kluge et al. 2010).
+Much of the disease burden due to opioid dependence and injecting drug use could be averted by scaling up needle and syringe programs (NSPs), opioid substitution treatment and HIV antiretroviral therapy (Degenhardt et al. 2010; Turner et al. 2011). Both methadone and buprenorphine (the two most commonly used medications) have been listed on the WHO's List of Essential Medicines (World Health Organization, 2005) as core medications for the treatment of opioid dependence (Mattick et al. 2008, 2009). OST reduces mortality among opioid-dependent people (Davoli et al. 1994; Caplehorn & Drummer, 1999; Brugal et al. 2005; Darke et al. 2006; Gibson et al. 2008; Degenhardt et al. 2009b), with time spent in treatment halving mortality compared with that in time spent out of treatment (Degenhardt et al. 2011). A large evaluation study in multiple countries, including LMICs, has demonstrated that OST is effective in reducing opioid use and injecting risk behaviours and improving physical and mental wellbeing (Lawrinson et al. 2008). There is increasing evidence that not only HIV (Degenhardt et al. 2010) but also HCV (Turner et al. 2011) burden can been reduced through NSPs; HCV burden can also be decreased by effectively treating chronic HCV (Turner et al. 2011). The release of more effective and less toxic HCV drugs is expected to dramatically improve what have been extremely low rates of HCV treatment uptake by people who inject drugs (Swan, 2011).
+There is also scope for reducing the risk of overdose among people who continue to use opioids. There is increasing evidence that the provision of the opioid antagonist naloxone to opioid users enables peers to effectively intervene if overdoses occur (Galea et al.
+134 F. J. Charlson et al.
+2006; Sporer & Kral, 2007). Additional strategies may include: education of users about the risks of overdose (especially high risk periods such as post-release from prison or after a period of abstinence), and motivational interviews with users who have recently overdosed (Sporer, 2003). Safe injecting rooms have been proposed as an additional strategy to reduce overdose, although their population reach is likely to be more limited (Hall & Kimber, 2005). There is evidence that psychosocial interventions including self-help programmes and cognitive behavioural therapy are effective in psychostimulant dependence (Baker et al. 2005; Knapp et al. 2007).
+In low-income regions, mortality in epilepsy patients is largely due to preventable causes (Diop et al. 2005; Jette & Trevathan, 2014). Yet, the treatment gap is more than 75% in low-income countries, and more than 50% in many lower and upper middle-income countries (Jette & Trevathan, 2014). Legislation to ensure availability of affordable and efficacious antiseizure medications, clinician education in prescribing antiepileptic medications, and patient education regarding the importance of medical adherence is critical to alleviate the epilepsy treatment gap. Cost-effective epilepsy treatments are available and accurate diagnosis can be made without costly technical equipment. Targeting epilepsy risk factors, including more common structural and metabolic causes of epilepsy will likely decrease mortality risk as well. In addition, education and information provision on safe lifestyle habits in epilepsy patients (i.e., avoiding fires, swimming and driving in those with active convulsive epilepsy) will clearly be beneficial. Education to dispel myths associated with epilepsy among employers and teachers may empower those with epilepsy to seek treatment.
+Mortality in dementia patients is commonly by preventable medical conditions, including infections. Caregiver education and support services regarding proper care of patients with cognitive decline will likely decrease infection rates and thus, mortality. Government financial support for healthcare services and caregiver support would also benefit this population. Strategies to enhance nutrition, as well as monitoring and treatment of vascular risk factors including high blood pressure, hypercholesterolemia, smoking, obesity and diabetes, are important measures as well.
+Limitations of the study
+Quantifying mortality presents several challenges. Cause of death data is affected by multiple factors, including: certification skills among physicians, diagnostic and other data available for completing the death certificate, cultural variations in choosing and
+prioritising the cause of death, and institutional parameters for governing mortality reporting (Lozano et al. 2012). In LMIC populations, where many deaths are not medically certified, different data sources and diagnostic approaches are used (e.g., from surveillance systems, psychological autopsy work and disease registries) to derive cause of death estimates (Lozano et al. 2012). The implication is that cause of death assessments are subject to uncertainty; a good illustration is the widely debated difference in maternal death estimates by GBD and by the United Nations (Byass, 2010).
+Cause of death data also provided estimates of deaths due to MNSDs that were not captured within the main GBD 2010 categories. The decision was taken to create residual categories to reflect the additional mortality not captured within specific disorders. Deaths and YLLs were calculated for: 'other' mental disorders (16 140 deaths equating to 5.4 YLLs per 100 000 persons); 'other' drug use disorders (33 ,561 deaths equating to 22.5 YLLs per 100 000 persons); and 'other' neurological disorders (481142 deaths and 231.6 YLLs per 100 000). The modelling strategies for these residual groups do not allow calculation of excess mortality for comparison as done throughout this paper however the YLL estimates for these groups are exceptionally high.
+Mortality directly related to MNSDs is particularly difficult to capture in cause of death data due to the complex web of causality which link them with other physical disorders. Thus it becomes very important to identify and quantify the not inconsiderable excess premature mortality in people with MNSDs through understanding the pathway between these disorders and fatal sequelae.
+Although valuable, the CRAs undertaken as part of GBD 2010 provide an incomplete picture. There are almost certainly deaths where we may not have enough information to parse out what is causally related or what is due to confounding. Assuming multiple risk factors are independent of each other is also a limitation as done in CRA methodology. A more accurate quantification of the joint effects of multiple risk factors, that is what explains the difference between excess and cause-specific deaths, is an important area for future research.
+Conclusion
+Despite the challenges in quantifying causal mortality in MNSDs it is abundantly clear that the mortality-associated disease burden of mental, substance use and neurological disorders is significant. The continuing life expectancy gap in persons with these disorders
+represents a lack of parity between this portion of the population and the community in general (Thornicroft, 2013). People with MNSDs face additional barriers to physical health care because of stigma within the healthcare system, the 'silo' effect between mental and physical health care caused by overspecialisation, and diagnostic overshadowing of physical health issues by presence of mental disorders (Bailey et al. 2013). Differential access to 'usual' care for this group leads to poorer outcomes in terms of health loss and mortality and incurs high costs in health care provision (Centre for Mental Health, 2010). Further research and development of new strategies for reducing mortality associated with MNSDs is needed.

Excess-mortality-in-severe-mental-illness-10Year-populationbased-cohort-study-rural-EthiopiaBritish-Journal-of-Psychiatry.txt

@@ -0,0 +1,48 @@
+Premature mortality is a well-established adverse outcome of severe mental illness (SMI), most notably for schizophrenia, bipolar disorder and depressive disorder.1 Nonetheless, investigating mortality patterns remains important for: (a) for monitoring the profile of changing risk factors over time;2 (b) for evaluating the impact of sociocultural and geographic settings on mortality; (c) for reviewing the contribution of SMI to the global burden of disease; (d) for advocating for the inclusion of SMI in the global health agenda; and (e) to establish the mortality profile in population samples, which has not been demonstrated adequately to date. Moreover, the contribution of SMI to the global burden of disease has increased in the most recent analyses3,4 but is still likely to be underestimated substantially. For example, although mental disorders, particularly SMI, are the strongest predictors of mortality from self-harm,5 self-harm is calculated separately in the global burden of disease estimations.6 In addition, the indirect yet substantial contributions of mental disorders to mortality related to physical conditions are underrecognised in calculations of the global burden of disease.6,7 This underestimation has the potential to perpetuate the low prioritisation of mental disorders and the underinvestment in research and services related to SMI, particularly in low-income country settings, where policy-makers have to prioritise disorders with the highest burden and best outcome returns for their investment. Furthermore, our knowledge about mortality associated with SMI derives from clinical samples recruited in the context of service receipt or hospital admission, mostly from high-income countries, although over 80% of the world’s population lives in low- and middleincome countries with limited access to treatment. The little knowledge we have about the mortality of people with SMI from low- and middle-income countries comes from anecdotal accounts8 and the pioneering studies of the World Health Organization, which are now over three decades old.9-11 For example the International Pilot Study of Schizophrenia was initiated over 45 years ago, in 1966.11 Many changes have occurred
+in our understanding of mental disorders since: more refined methods of illness classification, case identification and monitoring have evolved; new methods for ascertainment of causes of mortality applicable in low-income settings have enabled researchers to define causes of death in more precise ways. Additionally, virtually no data exist on the mortality outcomes in bipolar disorder and severe depressive disorders in low-income country settings, which are also important contributors to premature mortality alongside schizophrenia. There is, therefore, a pressing need for up-to-date, methodologically rigorous, population-based studies from low-income countries. The aim of this report is to present the mortality outcomes of people with SMI from the Butajira-Ethiopia study on SMI, a recently completed large-scale, population-based cohort study.
+Method
+The cohort
+The study cohort and the methods for follow-up have been described in detail previously.12-14 A summary description of the cohort is presented below with the focus on the conduct of the mortality assessment. The Butajira cohort on SMI was established between March 1998 and May 2001 in the Butajira district, located 132 km from Addis Ababa, the capital of Ethiopia. At the start of the study, the area was administratively divided into 46 subdistricts and all except one inaccessible subdistrict were included in the study. Demographically, the district contained predominantly rural sites with only four of the subdistricts, representing 10.9% of the population, being urban. This is similar to the urban-rural balance seen in the southern region of Ethiopia.15
+The cohort was identified through a two-stage sampling design. In the first stage, potential participants with SMI (schizophrenia, bipolar disorder and severe depression) were identified through a supervised door-to-door survey and the key
+Fekadu et al
+informant method.16 The door-to-door survey consisted of lay interview with the Composite International Diagnostic Interview (CIDI), version 2.117 targeting adults aged 15-49 years, estimated at the beginning of the study to be 83 282.18 The CIDI was administered to 68 378 individuals (82.1% of the target population) and the CIDI case identification was augmented by trained key informants, community leaders and elders selected from each village.16 In the second stage, people with potential disorder were assessed using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN),19 a structured instrument to confirm diagnosis. The SCAN was administered by trained clinicians and the ratings were validated against clinical diagnosis by psychiatrists,2 with excellent agreement between the SCAN diagnosis and clinical diagnosis. The first cohort identification lasted 3 years (1998-2001) and 844 participants were recruited. Using similar methodology, a leakage study identified 75 additional participants in the next 3 years from the same catchment area (Fig. 1). Thus, the total cohort consisted of 919 participants with SMI, composed mostly of men (n = 572, 62.2%). This gender difference was primarily because the majority of individuals identified with schizophrenia were men (n = 296, 82.7%). There was no gender difference in bipolar disorder and the difference in severe depression was in the expected direction; women were overrepresented (n = 132; 61.4%). Details of the baseline demographic characteristics are presented in Table 1.
+Monitoring of cohort and confirmation of survival status
+The cohort was monitored serially over a median of 11.3 years (interquartile range 10.7-11.9), ending in February 2012. Three
+monitoring mechanisms were established. First, each subdistrict was allocated to trained field workers who were involved in the initial data collection. These field workers developed intimate knowledge of the patients and the family and provided a report on the status of patients about once per month. Second, patients had a monthly clinical follow-up at which time their clinical status, medication changes and any other relevant indicators were recorded. Third, all participants were assessed annually for diagnostic stability and symptom and functional status. Through this continuous monitoring the survival status of participants was ascertained prospectively for all entrants into the cohort.
+Death was confirmed within 1 month of death by field workers who completed the verbal autopsy using the WHO verbal autopsy questionnaire that was previously adapted for use in the Butajira area.22,23 Verbal autopsy is described as the ‘most promising’ method of ascertaining causes of death in settings where most deaths occur at home and are not attended by a doctor as is the case in our study setting.24 The verbal autopsy method has also informed the latest global burden of disease data on causes of death from low- income settings.4,25 The verbal autopsy questionnaire is prepared for completion by lay interviewers based on information from close family members who have knowledge of the terminal illness.26-29 The verbal autopsy questionnaire starts by asking about basic sociodemographic and personal habits (smoking and alcohol use). This is followed by details about the circumstances leading to death (signs and symptoms of illness, duration of the identified signs and symptoms, and explores potential external causes).
+The lay field workers of the Butajira study on SMI who administered the verbal autopsy questionnaire to close relatives of the deceased were trained in the administration of the questionnaire. Once the questionnaire was completed, decision on diagnosis of the cause of death was made by consensus of two physicians. The verbal autopsy data were supplemented by the Broad Rating Schedule (BRS).30 The BRS was originally developed to summarise the findings during follow-up of patients with a diagnosis of schizophrenia but the principles of the BRS are applicable to all disabling disorders. The symptomatic and functional status of the participants is rated for the last month. The rating is made based on all available information, including information from the participant, other informants and records. The BRS also assesses symptoms and disabilities using a modified version of the Global Assessment of Functioning scale (GAF). The score ranged from 1 (persistent inability to function in almost all areas) to 90 (good function in all areas). Generally a score of 60 and above is considered a reasonable level of functioning. The BRS contains sections on participants lost to follow-up and deceased. The causes of death were then grouped into broad ICD-10 disease classes32 as presented in Table 3.
+Analytic methods
+Data analysis was conducted with SPSS version 21 and Stata version 11 for Windows. Mortality was standardised using the latest (2007) census of the Southern Nations Nationalities Peoples Region of Ethiopia,15 which includes Butajira. In the absence of appropriate dates of birth, we followed the method proposed by Breslow & Day33 to calculate person-years of follow-up, and relied on age at entry, date of entry and date of exit. We estimated the amount of time contributed by each individual to each 5-year age category and summed up all those contributions for all cohort members and obtained the total number of person-years of observation in that category. Each participant was assumed to contribute 0.5 years to the age category of the participant at commencement in the study, but the precise follow-up duration
+was calculated for the exit year. A full 1 year was given to each intervening year. For participants exiting the study within a year of entry into the cohort the exact length of contributed time was calculated. We estimated the potential years of life lost (YLL) using the 2009 national data on life expectancy.34 As was the case in the global burden of disease estimation,4,35 YLL were computed by multiplying the number of deaths at each age (x years) by the standard life expectancy of the reference population (the Ethiopia population) at age x. Put slightly differently, YLL was computed by estimating the difference between the actual age at death of an individual in the cohort who died from any cause, and the expected age at death. This may be represented with the following formula:3 YLL = £di(E — i'), where i = actual age at death; d = number of deaths at age i; and E = Expected age at death estimated according to the 2009 life tables for Ethiopia, based on age and gender. The sum of the YLL was then divided by the number of deceased individuals to derive the mean YLL. Standardised mortality ratios were calculated as the ratio of the number of observed deaths in the sample with SMI to the number expected if the sample with SMI had the same mortality rate as the population within Southern region. Factors associated with premature mortality were computed using the Cox proportional hazards model. We also estimated the life expectancy at birth of the cohort based on the life expectancy of the population of the Southern region using Chiang’s method of abridged life tables in 5-year groups.37-39 This method has been recommended for its application to relatively small populations.38 Since those under 15 years of age would be unlikely to receive a diagnosis of SMI, the mortality rates of the Southern Region for the under-15-years age groups was applied in substitution.40 Additionally, we substituted the mortality rates of the Southern region for mortality above 60 because only a small number of participants had been above 60 (n = 6) and using the SMI cohort would lead to unstable estimation. Life expectancy was estimated for the whole cohort and by gender as well as by the three disorders. The differences in life expectancy at birth between those with SMI and that of the Southern region were calculated.
+Ethical considerations
+The study was initially approved by the ethics committee of the Department of Community Health and then by the Institutional Review Board of the Faculty of Medicine and the College of Health Sciences of Addis Ababa University. Treatment was made available free of charge for all patients needing treatment.
+Results
+In total 121 participants (13.2% of the initial cohort) died during the follow-up period. Details of the demographic characteristics are presented in Table 1. Nearly twice as many patients with schizophrenia died (n = 65, 18.2%) compared with those with bipolar disorder (n = 33, 9.5%, w2(1) = 10.9, P =0.001) or severe depression (n = 23, 10.7%; w2(1) = 5.7, P = 0.016). When all diagnostic categories were considered together, more deaths occurred among men (n = 88, 15.4%) than women (n = 33, 9.5%, w2(1)=6.5, P = 0.011). However, the difference was not statistically significant when comparison was stratified by diagnostic groups even though comparatively more men died in all categories: 19.3% v. 12.9% for schizophrenia, 10.9% v. 7.8% for bipolar disorder and 12.0% v. 9.8% for severe depression.
+Overall, mortality was twice that of the standard population (SMR = 213.9, 95% CI 177-256) (Table 2). The SMR was highest for schizophrenia (302.7, 95% CI 234-386), which was significantly higher than that of bipolar disorder (SMR =150.1,
+95% CI 103- 211) although not that of severe depression (SMR = 169.9, 95% CI 108- 255).
+The commonest cause of death, based on categories of the ICD-10,3 was related to infectious conditions (49.6%, n = 60) prevalent in the study area (Table 3). Unnatural causes accounted for a quarter of all causes of death (24.8%, n = 30), most arising from suicide (n = 19, 15.7%); the rest from other unnatural causes such as road traffic accidents and homicide (n =11). Proportionately, those with bipolar disorder had the highest mortality from suicide (24.2%, n = 8/33) although the difference
+was not statistically significant compared with both those with schizophrenia (13.8%, n = 9/65) and severe depression (8.6%, n = 2/23). Deaths from the various causes peaked in the first 5 years of follow-up and were more or less constant afterwards (Fig. 2).
+On average the YLL per person for all patients with SMI was 28.4 years. The YLL was slightly higher among women (30.0 years) compared with men (26.9 years). The YLL was also slightly higher for those with severe depression (29.4 years) compared with that for those with schizophrenia (27.7 years) and bipolar disorder (29.0 years). However, because of the larger number of patients who died, schizophrenia made the largest contribution to the overall YLL (52.4%), followed by bipolar disorder (27.9%) and severe depression (19.7%). Whereas deaths related to injuries accounted for 22.4% of YLL, suicide alone contributed to 8.7% of all YLL. Overall, those with SMI had a life expectancy gap of about 6 years and it was nearly 10 years for those with schizophrenia and depression (Table 4). The life expectancy gap for men and women was 5 and 6 years, respectively.
+The two factors associated independently with mortality were male gender, and shorter time in remission (Table 5). Thus, those who were in remission for less than 50% of the follow-up period had double the risk of dying (Hazard Ratio (HR) = 2.02, 95% CI 1.31-3.12, P = 0.002). Men also had increased mortality (HR= 1.67, 95% CI 1.04-2.66, P =0.032). Older age at enrolment into the study did not significantly increase the risk although there may be a trend (HR= 1.85, 95% CI 0.98-3.12, P = 0.060). For all
+disorder subtypes, longer periods in a symptomatic state were associated with shorter survival time (Table 6). For each of the diagnostic groups, spending a higher percentage of follow-up time in episode was associated with an increased risk of premature mortality. For those with depressive disorder (but not the other disorders), subthreshold symptoms were also associated with risk of premature mortality.
+In the month prior to death, most of the deceased were symptomatic (70.2%, n = 85/121) and functionally impaired (75.2%, n = 91/121) because of their mental illness (Table 7). Impairment was consistently high across diagnostic groups: schizophrenia (78.5%), bipolar disorder (75.8%) and severe depression (65.3%). Similarly, high percentages of the deceased were symptomatic across the diagnostic groups: schizophrenia (n = 44, 68.8%); bipolar disorder (n = 26, 78.8%) and severe depression
+(n = 15, 64.2%). Moreover, mental disorders were considered to be relevant to the death in over half of the deaths either because they were presumed to have led directly to death (n = 34, 28.8%) or were linked to the death (n = 29, 24.6%).
+Discussion
+To our knowledge this is the largest single-site study describing mortality in SMI from a low-income country setting. It is also one of the very few studies in the world describing mortality experience of community-ascertained cases with minimum treatment exposure. The study has several novel features that contribute to addressing the gap in our knowledge about mortality associated with SMI. First, the study employed rigorous community-based case identification and diagnostic methods, screening a population of over 68 000 people, and has monitored the individuals identified continuously over an average of 10 years of follow-up. Second, the cohort had very low treatment exposure at recruitment. Third, the report provides data on mortality associated with bipolar disorder and severe depression, where we have virtually no data from low-income countries. Finally, the causes of mortality were established close to the time of death and used validated methodologies. Moreover, by using prospective clinical and functional measures, our study also contributes to the evidence on the potential role of psychopathology in causing premature mortality.
+Main findings
+The study confirms that people with SMI, irrespective of setting, are at increased risk of mortality. Specifically, in this particular setting, the risk of mortality among those with SMI was double that of the general population, which is consistent with previous reports of mortality and SMI in high-income countries,41-49 However, the mortality figures were lower than what we had anticipated. This may be partly because of the increased mortality in the general population related to the high burden of preventable causes other than injuries.2 ’ 3 The SMR is also lower than the figure we previously reported based on a 5-year follow-up study of schizophrenia.50 This is likely to be because of higher mortality in the early parts of the follow-up as demonstrated in Fig. 2. Moreover, increased risk of early mortality has been
+demonstrated in service contact samples of both schizo-phrenia51,52 and mood disorders,45 particularly for unnatural causes.53 Thus, the previous report may have suffered from the short-term nature of the study given the higher burden of mortality in the early phases of the follow-up period. This is of interest from both a research and service perspective. In terms of research, longer term studies of mortality are more likely to provide more stable and more accurate estimation and should be encouraged. From a service perspective, given the mixed nature of our cohort, which was composed of both individuals with chronic and recent-onset disorders, the finding underscores the need for vigilance in the early periods of service provision irrespective of the duration of illness at the time of first-service contact. However, we cannot rule out the possibility that the reduction in mortality might have been as a result of better care provision related to recent restructurings in the healthcare system, for example, expansion of primary healthcare, although this would be expected to have a greater impact on the population without SMI. A distinct methodology would be required to detect the impact of health system changes on mortality. Based on the current findings, however, we recommend that provision of enhanced care in the early phases of the illness may be essential to improving the mortality outcomes of people with SMI.
+Deaths from infectious diseases
+Most patients in the study died from infectious causes prevalent within Butajira.23 This is partly a reflection of the limited care available for the population. Therefore, improving the general health of the public is likely to have an impact on the survival of people with SMI. Patients with SMI may also be at a particular disadvantage because they may not complain when they have symptoms, and family support may have been already overstretched when patients develop infections. Further exploration of the mechanisms underlying death from infectious conditions is warranted.
+Deaths from unnatural causes
+A large proportion of individuals have also died from unnatural causes, primarily suicide. Studies of suicide in Ethiopia have reported the rate to be between 6 and 8 per 100 000 per year.54,55 The rate among patients with SMI in the Butajira cohort was
+about 200 per 100000 per year, a substantially higher rate. The overall mortality rate from all unnatural causes was also high, accounting for 24.8% of all causes (n = 30/121) compared with the rate in the Butajira area which has been estimated at 1.7%.2 In a study from Addis Ababa, the capital of Ethiopia, unnatural causes of death other than suicide accounted for a higher rate (11.2%)55 although still around 50% lower than that of the Butajira sample with SMI.
+Contribution of psychopathology to mortality
+Given the high rate of deaths from unnatural causes and the high level of psychopathology in the deceased group, it is likely that psychopathology contributed to the death of many patients. This is supported by the finding that SMI was likely to have contributed directly to the deaths of at least a third of all deaths and contribute substantially to the other deaths. Additional evidence for the potential contribution of psychopathology to mortality comes from the survival analyses in which better survival was predicted by longer periods in remission, while longer duration in a symptomatic state was associated with increased mortality in all disorder categories. Two recommendations follow from these findings. First, it is reasonable to suggest that improving mental healthcare provision may improve mortality outcomes. Second, the contribution of psychopathology to mortality has to be accounted for in the estimation of the global disease burden although this may call for a new technology.
+preventable causes. The study adds new data not only on mortality in people with SMI in low-income countries, but also to the worldwide database on mortality among those with SMI identified and living in the community. The study highlights the need to improve the mental healthcare as well as the physical healthcare of people with SMI. The study also makes a case for the inclusion of mortality directly attributable to mental disorders in the estimation of the global disease burden. This is important to increase the visibility of mental health in the public health agenda, reflecting its rightful place.
+Life expectancy
+People with SMI lost about three decades of their life to premature death in this rural Ethiopian setting. Extrapolating this nationally, people with SMI lose a substantial amount of potential life years annually. Despite overlap in confidence intervals with the Southern region, the findings regarding the life expectancy gaps are important. A life expectancy at birth for people with schizophrenia and depression (46 years) is very low. Given the low life expectancy of the population in general, the life expectancy figures are worth paying attention to and have to be taken as evidence of the substantial neglect of people with SMI in such settings. Moreover, this study will serve as a baseline for future studies of the life expectancy gap in people with SMI in Ethiopia, which is likely to grow with improvement in the life expectancy of the general population.
+Limitations
+There are a few limitations worth mentioning. First, because of the lack of data on national cause-specific mortality, we were unable to present cause-specific mortality ratios for people with SMI. Second, cause of mortality confirmed by a physician would have been the best approach to determine causes of mortality; however, short of a physician determination, we have used the next best approach, the verbal autopsy method. The historical nature of the verbal autopsy assessment may have led to random misclassification of cause of mortality. Finally, information related to self-harm is not always volunteered. Therefore, the figure for suicide may be an underestimate.
+Implications
+This relatively large-scale cohort study from rural Ethiopia confirms the increased risk of mortality associated with SMI irrespective of setting. The study re-establishes mortality as a very important outcome for people with SMI in low-income settings despite the high burden of premature mortality from diverse

Exploration-of-morbidity-suicide-and-allcause-mortality-in-a-Scottish-forensic-cohort-over-20-yearsBJPsych-Open.txt

@@ -0,0 +1,83 @@
+BJPsych Open (2020)
+6, e62, 1-10. doi: 10.1192/bjO.2020.40
+Background
+The elevated mortality risk associated with a diversity of mental disorders when contrasted with the general population is an accepted finding.1-3 The number of years of life lost in relation to all-cause mortality varies from 7 to 24 depending on the nature of the condi-tion.2 Substance use disorder (SUD) conveys the highest potential number of years lost (9-24); however, this is closely followed by the ranges for personality disorders (13-22 years), schizophrenia (1020 years) and bipolar disorder (9-20 years) demonstrating a psychophysiological influence (which we define as a physiological response mediated by biochemical pathways to psychological distress) upon morbidity/mortality beyond the physical impact of illicit substances.
+In relation to dual diagnosis the risk of death among individuals with severe mental illness (SMI) and concurrent SUD is in excess of those with SMI alone.4 Evidence suggests all-cause mortality rates for individuals with schizophrenia may also be increasing,1,5,6 but high rates of suicide, particularly after discharge from psychiatric hospital admissions7 and unnatural deaths8 do not entirely account for the observed disparity in mortality compared with the general population.1
+All-cause mortality in forensic settings
+Although research primarily focuses on mainstream services, higher rates of all-cause mortality have also been reported within prison settings where the burden of mental health is also in excess of general population comparisons,9 with mortality risk further compounded by the high prevalence of SUD reported among prison-ers.10 Similar high rates of all-cause mortality are apparent among the forensic psychiatry literature.1 - In 2011, Clarke et al12 noted the deaths of 9.6% of their cohort (n = 595) over a maximum 20-year follow-up, Fazel et al11 reported 29.9% of their population (n = 6520) dying over a mean of 15.6 years and Coid
+et al13 found 4.9% of their cohort (n = 409) died over a mean of 6.2 years. Although they report differing mortality rates they all exhibit high rates of suicide, 32%, 22.7% and 50% of the reported deaths, respectively. In addition, Fazel et al11 noted the death of 14.2% of their cohort and Clarke et al12 22.8% of deaths from acci-dental/unnatural causes.
+There is limited literature regarding morbidity/mortality among forensic patients who represent a specific subset of individuals experiencing SMI.14 Findings from mainstream populations cannot be safely generalised because of differing recovery and treatment pathways which, in the case of forensic patients, require to be balanced against the risk to the public.15 To address this and from within the context of recovery, a cohort from The State Hospital, Carstairs, the high secure hospital for Scotland and Northern Ireland, has been explored 20 years from their involvement in The State Hospital Survey.16 As part of this follow-up morbidity and mortality has been examined to explore influencing factors at a local level.
+Aims
+The aims of this paper are to:
+(a)	explore morbidity among this group;
+(b)	delineate which patients are at greatest risk of premature mortality;
+(c)	assess the extent of suicide and unnatural deaths;
+(d)	establish which factors, if any, appear protective to cohort members.
+Method
+Cohort
+The State Hospital Survey16 identified a whole population cohort of 241 patients (male n = 213, mean age 36 years, female n = 28, mean
+https://doi.org/10.1192/bjo.2020.40 Published online by Cambridge University Press
+age 32 years) resident in the high secure State Hospital, Carstairs, at some point between 25 August 1992 and 13 August 1993. Detention was under civil (n = 92, 38.2%) and criminal procedures (n = 149, 61.8%). Baseline data were collected from case notes and clinical interview. At baseline, the mean lifetime psychiatric stay was 9.3 years (range: 0.08-45). Following the baseline study the mean high security admission was 6.8 years (median 4.3, range: 0.0922.4). Individuals who were subject to restrictions on discharge, excluding prison transfers (n = 81), experienced a mean 9.4 years (median 6.1, range: 0.15-22.4) in high security. Non-restricted individuals and individuals who were prison transfers (n = 160) spent a mean of 5.4 years (median 3.5, range: 0.09-22.4) in high security. Additional detailed cohort characteristics have been reported elsewhere.16
+In 2014 follow-up was initiated with clinical and morbidity data collected for a mean of 19.2 years and mortality data only, for a mean of 21.1 years (median 25.1, range: 0.61-25.4) with 5093.61 person-years at risk (PYAR).
+The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. The study was approved by South East Scotland Regional Ethical Committee 01, reference 15/SS/0015. Supplemental approvals and the wider study protocol are described elsewhere.1 Written informed consent was obtained from all living participants.
+Data sources
+The Electronic Data Research and Innovation Service (eDRIS) utilised the cohort’s unique Scottish Community Health Index (CHI) numbers (CHI is a population register used for healthcare purposes) to search the data-sets of National Services Scotland (NSS), the Scottish Health data controller. NSS provided Scottish morbidity/ mortality information and emigration data. The National Health Service Central Register (NHSCR), using full name and date of birth, supplied UK mortality information and indicated general practitioner registration out with Scotland. Robust mortality status information was unavailable for seven individuals traced to Northern Ireland and one individual who was overseas therefore these eight male participants are classed as mortality status unknown.
+NSS provided date of death and ICD-9/10 codes for causes of death. NHSCR corroborated that data and supplied all death certificate information, identified deaths undetected by NSS and deaths within the rest of the UK. NSS provided ICD-9/10 codes recorded during general hospital in-patient and day-case admissions (Scottish Morbidity Record 01), event duration, partial date (month/year) and admission type with data requested for all deceased and consented individuals.
+Years of birth for the deceased cohort were applied to the expectation of life, by gender and selected age, Scotland, 1861 to 2011 table.18 Deaths were categorised as premature; defined as dying before the age listed based on year of birth/closest period or as post-expected, where death occurred at any point beyond the specified age. Using indirect standardisation19 we calculated standardised mortality ratios (SMR) and 95% CI comparing the risk of death among the cohort with the risk of death within the Scottish population.
+Data analysis
+We used Cox regression analysis to examine variation in risk of premature death over time according to baseline characteristics. The time in the study ran from 1 September 1992 (or recruitment to baseline study) until 31 December 2017 or date of death. Cases of
+individuals experiencing death post-expected age were filtered from analysis. The eight individuals with mortality status unknown were right censored20 with time in study as recruitment until discharge from The State Hospital when they were lost to follow-up. Independent variables were based upon previous literature and where preliminary analysis yielded statistically significant results.
+Kaplan-Meier plots were constructed for each covariate to assess the assumption of proportional hazard. Where the assumption was violated a time-dependent covariate was included. Cox regression models were constructed for each covariate while controlling for age at baseline. All analysis were conducted using SPSS version 22.21
+Results
+Mortality
+Eighty-nine individuals (36.9%) were deceased as of 31 December 2017 providing an all-cause crude death rate (CDR) of 1747/100 000 PYAR (95% CI 1403-2150). At point of death: 51.7% (n = 46) were resident in the community, 30.3% (n = 27) within low secure/open wards, 15.7% (n = 14) in high security and 2.2% (n = 2) were detained in prison.
+Mean age of death was 55.6 years (range: 30.5-84.7). Seventyseven (36.2%) men died and 12 (42.9%) women at average ages of 56.6 years (range: 30.5-84.7) and 48.9 years (range: 36.2-66.1) respectively.
+Table 1 outlines mortality status by primary diagnosis. As detailed in Table 2, 52 (67.5%) men died prematurely, and 11 (91.7%) women died prematurely. The mean years of potential life lost were 14.9 years (range: 0.09-35.7) for the men and 24.1 years (range: 5.2-35.8) for the women (10 women; one woman was born, lived primarily and died overseas). Excluding prison transfers, 22.2% (n = 18/81) prematurely deceased individuals were subject to restrictions at baseline (specific mortality rate 1052/100 000 PYAR, 95% CI 623-1663) with 28.1% (n = 45/160) non-restricted individuals having prematurely died (specific mortality rate 1330/100 000 PYAR, 95% CI 970-1780).
+The SMRs by gender are detailed in Table 3. The SMR for allcause deaths (SMR = 397, 95% CI 321-487) indicates a mortality rate almost four times that observed within the Scottish population. The ratio for the males (SMR = 297, 95% CI 236-370) returns a threefold increase in mortality whereas the females exhibit an SMR (SMR= 1000, 95% CI 542-1700) ten times the population rate. A high proportion of deaths (91.0%) occurred as a result of natural causes (SMR = 401, 95% CI 321-496).
+Suicide was defined by explicit codes; E950-959 (ICD-9) and X60-X84 (ICD-10). In addition, National Records Scotland include ‘event of undetermined intent’ among probable suicides. There were five suicides representing 5.6% of the 89 deaths (CDR 98/100 000 PYAR, 95% CI 32-229). Suicide mortality (SMR 625, 95% CI 229-1385) was six times the population rate. Accidental deaths occurred in three individuals (3.4%) (SMR = 375, 95% CI 95-1021) at almost four times the population rate. Re-categorising substance misuse deaths (n = 4, 4.5%) as accidental raises unnatural mortality to almost nine times (SMR= 875, 95% CI 382-1731) the population rate.
+Cause of death
+Underlying cause of death was assessed using ICD-9/10 classifications and categorised as premature/post-expected age as detailed in Table 4. Premature death occurred in 63 (70.8%) people. The primary cause of premature death (n = 20, 31.7%) was respiratory
+disease/cancer, with circulatory disease/events (n = 12, 19.0%) forming the next largest group with other cancers responsible for 11.1% (n = 7). There were no significant differences between premature and post-expected age deaths by cause.
+As noted there were n = 5 suicides. For these people baseline primary diagnoses were mixed: three had schizophrenia, one had antisocial personality disorder (ASPD) and one alcoholism. One had comorbid ASPD. At baseline four reported a history of heavy/abusive alcohol misuse and all confirmed polysubstance drug use. At death, two were under mental health legislation, four lived in the community and one in a low secure ward. All accidental
+deaths occurred in the community, two because of assault and one asphyxiation with food.
+Statistical analysis
+Cox regression analysis (Table 5) suggests that men detained under civil provisions at baseline are more likely to die prematurely than those detained under criminal provisions. Males experience a lower risk of premature death and higher survival than females. Diagnosis of intellectual disability at baseline is associated with lower hazard and increased survival. Being female and having a
+comorbid diagnosis of ASPD significantly increased the likelihood of premature death, a finding not replicated in the males. In general, and for the males, receiving antipsychotic medication by depot injection was associated with higher hazard and shorter survival. Substance misuse at baseline (alcohol misuse or illicit drug use) was significantly associated with an increased hazard of premature death both overall and among males.
+accidental harm. Prematurely deceased individuals received significantly more diagnoses (11.59, P = 0.022) compared with consented participants (5.8 diagnoses).
+Those who died prematurely spent on average significantly more days as general hospital in-patients in relation to total trauma (14.1 days, P = 0.007), urgent (12.4 days, P = 0.011) and routine admissions (15, P < 0.001) than living participants.
+Morbidity
+Data relating to Scottish general hospital in-patient admissions for the 89 deceased and 66 consented participants were obtained from NSS (n = 154 as one participant died following consent). A nil return was obtained if the individual had never had a general hospital in-patient admission in Scotland between baseline and 31 December 2014/date of death. Data were acquired for 115, with a further 39 individuals receiving a nil return. Table 6 reports the mean number of unique ICD-10 codes allocated for each ICD-10 block. Table 7 outlines the number of individuals in receipt of an ICD-9/10 classification by ICD-10 block description and Table 8 the number of days spent as a general hospital in-patient.
+Examination of the tables presented provides an overview of the physical health of the alive, prematurely deceased and post-expected age deceased groups. No significant differences were observed in terms of mean endocrine, nutritional and metabolic disease diagnoses between the living participants and those who prematurely died. Similarly, no significant differences were observed for diseases of the circulatory system. In relation to respiratory system diseases, a significant difference was observed in mean diagnoses between prematurely deceased (1.44 diagnoses, P = 0.002) and living participants (0.48 diagnoses). In terms of ICD-10 blocks S-Y ‘Injury, poisoning & certain other consequences of external causes’, a significant difference was observed between prematurely deceased (4.05 diagnoses, P = 0.005) and living participants (1.31 diagnoses). Unfortunately, injuries resulting from self-harm cannot be distinguished from
+Discussion
+This study investigated morbidity and mortality findings from a 20year follow-up of a cohort of high secure forensic patients first recruited as a whole population survey and followed through the context of recovery.
+Unnatural and suicide deaths
+In stark contrast to previously reported forensic6,22 and mainstream psychiatric findings2,23 we did not observe exceptional rates of suicide and accidental/unnatural deaths. We described a sixfold (SMR = 625) increase in suicide and an almost fourfold (SMR = 375) increase in unnatural deaths against the general population. In comparison, Clarke et al12 reported a SMR for suicide of 3231, and for all unnatural deaths of 1898 in relation to a cohort followed for a maximum of 20 years (n = 595, 5593 PYAR) after first admission to a medium secure unit. Including the few drug/alcohol deaths within accidental (unnatural) deaths we only returned an almost ninefold increase. No suicide/accidental deaths occurred within high secure care, and only one of our eight suicide/accidental deaths happened within the psychiatric hospital environment. Of the n = 31 suicide/open verdicts and accidental deaths reported by Clarke et al12 almost half (n = 15) the preventable deaths occurred within high/medium security and the wider hospital environment.
+We considered a number of factors in exploring this finding of a low rate of suicide and accidental deaths within our cohort. First, was the Scottish suicide rate lower than elsewhere in the UK? The 2018 Scottish suicide rate (16.1/100 000 persons) was greater than the comparably reported English rate (10.3/100 000).24 Therefore Scotland does not have a naturally low rate of suicide.
+Second, was our low rate of suicide reflected in the general adult psychiatry population in Scotland? This was refuted by a historical Scottish population study of discharged, long-term (>1 year) psychiatric in-patients that reported an increased suicide risk 13 times the population rate.25
+Third, was the diagnostic profile of this Scottish cohort impactful? Unlike other UK regions,13 within Scotland offenders with a primary diagnosis of personality disorder generally remain within the criminal justice system. At baseline only 5% received a primary diagnosis of personality disorder whereas 70% attracted a schizophrenia diagnosis. This compares with 25.8% of individuals with personality disorder in a Swedish forensic cohort,11 26.6% in an English medium secure cohort12 and 13.5% in an English community cohort.1 Only the Clarke et al12 study splits suicide between mental illness and psychopathy at 72.2% and 16.6%, respectively, which is reduced in the personality disorder group given that it
+was 26.6% of their cohort. The different diagnostic profiles may be a factor in explaining our findings but this is not supported by the Clarke et al12 study and would not account for the extent of the difference. Suicide rates are generally high among individuals experiencing personality or psychotic disorder26 with risk of accidental death higher for personality disorder compared with schizophrenia.8 Contextualising our suicide deaths within our diagnostic profile we observe a similar pattern; 60% of people who died by suicide had schizophrenia at baseline, one person had ASPD and another alcoholism. A total of 80% were male and the same proportion had a history of heavy/abusive alcohol use, all displayed polysubstance drug misuse, known risk factors for non-natural death.26 That description differs from the overall cohort, 47% of whom reported substance misuse at baseline and 29% had comorbid ASPD.
+Fourth, does the year the cohort was established or the length of follow-up influence findings? Compared with other studies11-13 our cohort was followed for the longest time therefore increasing rather than decreasing suicide likelihood. Alternatively, have more recently established cohorts observed behavioural change, making suicide/accidental in-patient death more likely? Only one suicide has been noted in our cohort source hospital (The State Hospital)
+since 1996 (email from Health Records, The State Hospital, tsh. Health_Records@nhs.net, November 2019).
+Finally, is engagement with the Scottish forensic mental health system protective against suicide or accidental death? An examination of suicides (n = 14) occurring within high secure care in Scotland (The State Hospital, 1972-1996)27 noted treatment-resistant schizophrenia and having committed violent offences as suicide risk factors. Greater liberalisation of the hospital, increased activities and reintroduction of clozapine were suggested for the suicide reduction observed over time. This continuing developmental process alongside estate renewal may influence our observed low suicide/accidental death rate while patients resided in high secure care. At death two individuals were under mental health legisla-tion/restrictions. Examining CDR figures, restrictions made little difference to the premature death rate. Also considered was if a longer period of contact and mandated relationship with services may protect against suicide; alternatively, lack of autonomy may be a negative influence. Neither was apparent within our findings.
+We hypothesise that factors within the Scottish forensic inpatient environment; physical, procedural and/or relational are protective in terms of suicide prevention or in deterring behaviour leading to accidental death and that these may have an ongoing effect on patients’ relationships with services in the community. There may also be organisational factors that reduce avoidable deaths. Scottish forensic services have advantages in terms of size and cohesion, with few independent secure beds, and the strategic lead of the Forensic Network.
+Risk of premature mortality
+A large proportion of our cohort died (36.9%) demonstrating an almost fourfold increase (SMR=397) in all-cause mortality. This represents a slightly higher CDR than reported for forensic services in England/Wales,22 mirrors an English general adult community-recruited cohort experiencing psychosis28 and is lower than reported internationally.22 In contrast, 91% of our cohort died of natural causes four times (SMR = 401) the rate of the general population. Natural deaths within the discharged Scottish general adult psychiatry in-patient population has been reported at SMR = 16929 indicating a difference in mortality between Scottish forensic and general adult in-patients.
+Overall, 70.8% of deaths were premature and naturally occurring at an average age 55.6 years. Respiratory disease/cancer was the underlying cause of almost a third of premature and 36% of all deaths, with cardiovascular disease related to 19% of premature and 21% of all mortality. Study recruitment occurred during 1992/3 when almost all patients smoked or were passive smokers. Since December 2011, our cohort source hospital, The State Hospital, Carstairs has been entirely smoke free.
+Rates of respiratory related deaths, in excess of fourfold the expected level have been noted in an early population study of discharged Scottish general adult psychiatry long-term (>1 year) in-patients25 with rates of death related to cardiovascular disease being slightly raised at SMR = 160. A later examination29 capturing all discharged general adult psychiatric in-patients reported that cardiovascular disease, despite displaying only a small rise in relative risk (SMR =170) was responsible for 67% of total mortality and 54% of the total years of potential life lost. A similar English study examining death within a year of discharge of patients with schizophrenia reported SMR of 470, and 250 for respiratory and circulatory disease deaths, respectively, with increased rates from 1999 to 2006.6
+The cohort of 28 females represents around 50% of Scottish female forensic patients, reported as n = 56 in 200430 and n = 60 in 2017.31 There is an obvious disparity between the male and female groups. Almost all women died prematurely aged under 48
+years with mean loss of 24 years of potential life. Their SMR for natural death was three- to fourfold the male rate and that held for females with schizophrenia, rising to fivefold for a primary diagnosis of ASPD. Receiving a comorbid diagnosis of ASPD was also associated with premature death. This suggests that ASPD may have a unique impact on the physical health of women in a manner not reflected in risky behaviour or suicide completion.
+There is a lack of published literature regarding gender differences in mortality among patients located within forensic services with papers and reviews generally commenting on the percentage of males within the cohort.1 , Where data does exist,12 our results replicate those findings, namely that SMR for death because of natural causes, unnatural causes and suicide are higher for the female than the equivalent male cohort. However, as with those presented here, findings must be interpreted with caution because of the extreme gender imbalance evident within forensic psychiatric services and reflected in small numbers of females in research cohorts.
+A similar picture has, however, been observed within mainstream services, specifically in relation to individuals with schizophrenia. A population follow-up of all in-patients admitted with schizophrenia diagnosis from 1980 to 2006 indicated that up to 1992 the SMR for males in relation to all-cause, unnatural and natural deaths exceeded that of females; however, for individuals admitted post 1992 the opposite was observed: females displayed greater all-cause, unnatural and natural SMR.3 More recently an American population study reported that female SMR for allcause mortality exceeded the male equivalent; however, the SMR for unnatural death was higher among males.5
+Almost 68% of males died prematurely at mean age 50 years. Males detained under civil measures were significantly more likely to die prematurely than those males under criminal procedures. Similarly, males in receipt of depot antipsychotics at baseline were significantly more likely to have died prematurely. It is conceivable that as civil patients transfer to high secure care because of acute morbid positive symptoms and/or locally unmanageable levels of violence or aggression16 and by utilising depot medication as a proxy for illness severity, we may be observing those who experienced the severest symptoms and psychological distress, dying prematurely. Specifically, that is, individuals who may be more greatly influenced by psychophysiological factors.
+Scottish forensic services acknowledge that female patients represent a heterogeneous group, often more chaotic and challenging with differing needs to male patients,30 with higher rates of mortality33 and generally poorer outcomes.3 Although Scotland lacks female high secure care and exclusively single gender medium secure provision (with some services provided within England), because of the general complexity of female patients, providing appropriate relational security and support can be more important than physical security.31 Regardless of shortcomings in the Scottish female forensic estate, the high levels of premature mortality of both genders cannot be overlooked.
+Despite spending on average longer as high secure in-patients, with exposure to the health benefits that offers, there was little difference in CDR for those subject to restrictions on discharge and those not. Fazel et al22 pointed out that mortality rates among forensic inpatients are high but more closely reflect the general adult psychiatry population than prisoners. They postulate that greater than any factors negatively influencing mortality within forensic environments are the poor lifestyle choices evident among general psychiatric populations that can compound psychotropic medication side-effects. Given our respiratory findings, however, the impact of a heavy smoking environment must be acknowledged. Although poor lifestyle choices have an impact on morbidity/mortality among individuals experiencing SMI again the increased morbidity risks and resultant mortality are not fully explained by behavioural patterns.35,36
+Protective factors
+Within this cohort some protective factors appeared evident. Although both genders died at early ages the males lived on average almost 8 years longer compared with the females. Males with a primary diagnosis of intellectual disability appeared less likely to die (SMR= 147, 95% CI 53.9-326) and it could be that those patients with intellectual disability received less psychotropic medication and accompanying side-effects. One reason suggested for high rates of morbidity among individuals with mental ill health is the propensity for diagnostic overshadowing to occur. This is when disparities occur in the treatment and diagnosis of physical disorders as a result of misattribution of physical symptoms to mental illness.37 Although we interpreted experiencing primary intellectual disability as a protective factor, diagnostic overshadowing has been presented as a particular problem within the intellectual disability population with symptoms related to physical or mental ill health being misattributed to their intellectual disabilitiy.38 We suggest that in line with our assertion that engagement with the Scottish forensic mental health system may be protective against unnatural death and death by suicide, for individuals with intellectual disability, location within services almost exclusively under National Health Service operational control may confer advantages to forensic patients with intellectual disability in terms of staffing, their training and support and patient services offered. Higher levels of in-patient and community support to avoid offending and foster appropriate behaviour may also reduce stress and encourage patients with intellectual disability to adopt heathier lifestyles, with a potential reduction in drug and alcohol use. More attention may be paid to their physical health by staff, and general practitioner referrals encouraged and supported, leading to earlier physical health intervention and a reduction in diagnostic overshadowing.
+It is also acknowledged that individuals with intellectual disability represent a particularly vulnerable subpopulation within the prison environment, being subject to high rates of mental disorder39 and possibly greater risk of suicidal ideation than the general population.40 We propose that a contributing factor to this poor outlook is a lack of equivalence, equality and equity for people with intellectual disability within the UK prison system.41 Again we propose that our intellectual disability cohort were protected from premature mortality precisely because they were, where appropriate, diverted from the prison environment and supported by specialist forensic psychiatric services within hospital and community settings, designed to promote and provide equality of life experience. Further research is required to address the paucity of literature robustly identifying and exploring the journey of individuals with intellectual disability through services.
+Morbidity
+Antipsychotic medications, a reliable mechanism for easing symptoms, reducing distress and therefore enhancing recovery carry with them a diversity of side-effects: activating, sedating and metabolic. Unsurprisingly antipsychotics have been targeted as a possible reason for the premature mortality associated with schizophre-nia.1,42 Population studies43 indicate that individuals with schizophrenia treated with antipsychotics or antidepressants have a lower risk of death compared with individuals not receiving such medications. A primary diagnosis of schizophrenia was applied to 70% of this cohort, and receiving depot neuroleptics was significantly associated with premature death; however, there were no significant differences in mean diagnoses of cardiovascular or endocrine, nutritional and metabolic disease applied to the living participants and the prematurely deceased. What is apparent is the significant difference in terms of respiratory disease, with those dying prematurely receiving more diagnoses. Although
+undoubtedly the legacy of an era, smoking remains an issue for forensic patients after discharge from controlled in-patient environments.
+Those who died prematurely attracted significantly more diagnoses related to injury, poisoning and other external causes; however, these did not translate into high numbers of traumatic deaths or completed suicides. Overall, those dying prematurely received more physical health diagnoses across all listed ICD-10 blocks and spent significantly more days as general hospital inpatients than those who remained alive evidencing the poorer physical state of that group.
+Scotland does not have the healthiest national population, indeed the ‘Scottish effect’ is much examined with 17 hypotheses proposed to account for excessive premature mortality.44 We suggest that something akin to the ‘Scottish effect’, specifically in terms of biopsychosocial stress and the resultant physiological stress response, is being observed among mainstream and forensic psychiatric populations. High levels of adverse life events observed among Scottish forensic patients45 together with the dose-response association observed between adverse life events/psychological distress and a negative impact upon subjective and objective physical health46 may be evidenced within our reported cohort. There is increasing physiological evidence of heightened levels of oxidative stress and inflammation within anxiety, depressive, bipolar disorder and schizophrenia.36,47 These processes provide mechanistic pathways by which leucocyte telomere length can be shortened. Metaanalysis has evidenced shortened telomere length across a range of psychiatric disorders36 with length being an indicator of cell ageing and short length associated with age-related morbidities, for example immune dysregulation, cancer, diabetes and cardiovascular disease.48 Interpreting the morbidity and excess mortality observed in cohorts such as this through the lens of the ‘Scottish effect’ may lead to better targeted local interventions.
+Limitations
+This study has several limitations. Follow-up physical health information could only be requested for deceased or consented participants. The wider study adopts a gatekeeper approach therefore all cohort members whose gatekeeper denied access because of their inability to provide ‘fair’ consent49 or they were not physically/mentally well enough to be approached were excluded. This removed individuals with the greatest physical health needs and/or the severest psychiatric symptomatology. The mortality status of eight individuals remained unknown although of these seven resided in Northern Ireland. We could not locate or engage with appropriate Northern Ireland services to confirm their status or locate their current care team. It proved impossible to access information without consent. Accessing cohort member gatekeepers within England was equally difficult but mortality information moved between NHSCR and their English counterparts. Some individuals with mortality status unknown could have died because of suicide or accidental causes, raising our mortality profile, but it would remain lower than the reported studies. We were also as reliant on the general hospital clinicians accurately recording and/or applying the most appropriate ICD-10 codes as we were of the cohort member and their accompanying staff providing a precise physical health history.

Exposure to parental mortality and markers of morbidit.txt

@@ -0,0 +1,74 @@
+INTRODUCTION
+Suicide is one of the most important causes of death among young people in Europe.1 Parental risk factors, including parental suicidal behaviour, 2—4 parental psychiatric morbidity2 4 and parental non-suicidal death,4 contribute to suicidal behaviour.
+A life-course approach may add to our understanding of the aetiology of suicidal behaviour.5—8 This approach is based on the assumption that the impact of biological and social risk factors as well as protective factors may vary with age at exposure due to a differential impact at different developmental stages.8 Any variations in risk for suicidal behaviour
+in the offspring with age at first exposure to parental factors may be due to the accumulation of exposures during the life span, due to a latency in outcome manifestation, or due to specific developmental mechanisms that act exclusively (‘critical period hypothesis’) or predominantly (‘sensitive period hypothesis’) during a specific age window.8 In recent years, the life course paradigm has reached a considerable resonance in chronic disease epidemiology.8 In spite of this development, applications related to suicidal behaviour in the offspring remain sparse.7 In order to explore the effects of age at exposure to parental mortality and markers of morbidity on the risks of suicide and attempted suicide in offspring, we conducted the present study.
+METHODS
+Study design
+We conducted a matched case—control study through record linkages between Swedish national registers. The study base consisted of all individuals, born in Sweden between January 1973 and December 1983, who were singletons and for whom information on both biological parents was available. The cases comprised all individuals recorded in the National Patient Register (NPR) or in the Causes of Death Register (CDR) due to attempted or completed suicide (E950—E959 in the International Classification of Diseases ICD-8 and ICD-9, X60—X84 in ICD-10). The National Board of Health and Welfare provided us with information on the equivalence of ICD-8/ ICD-9 and ICD-10 codes.
+Cases included suicides and attempted suicides with uncertainty about intention (E980—E989 in ICD-8 and ICD-9, Y10—Y34 in ICD-10). Uncertain and certain diagnoses were combined to limit temporal and regional variation in ascertainment routines.2 A sensitivity analysis of certain and uncertain suicidal behaviours established the comparability of the estimates.
+Attempted suicide in offspring was also analysed with regard to the method used. We distinguished violent methods, including hanging, use of firearms and knives, jumping from heights and in front of moving objects, and drowning (ICD-8 and ICD-9: E953-957 and E983-987; ICD-10: X70-X82 and Y20-Y32) from non-violent methods, which included all forms of poisoning (ICD-8 and ICD-9: E950-952 and E980-982; ICD-10: X60-X69 and Y10-Y19). This classification strategy was in accordance with
+related literature investigating violent and non-violent suicide methods.9
+Cases comprised 1407 individuals with suicide completion and 17159 individuals with attempted suicide. They were each matched by sex, month, year and county of birth to up to 10 randomly selected controls. Only individuals who were alive and living in Sweden at the time of the index event were eligible to serve as controls. Events coded as attempted suicides and suicides before the age of 10 years might be misclassified and were excluded from the analyses. Suicide victims were assessed from 1 January 1983 until 31 December 2004, and were up to 31 years of age at the end of follow-up. Suicide attempters were assessed from the same date up until 31 December 2006.
+Data sources
+In Sweden, all residents are identified by a unique identification number. This enables merging of individual information from different national registers.
+Children were linked to their biological parents using the Multi-Generation Register (MGR). Of the individuals born after 1950, less than 2% could not be linked to their parents.10
+We derived data on completed suicides in parents and offspring and parental deaths due to other causes from the Causes of Death Register (CDR). The National Patient Register (NPR) provided information on attempted suicides in parents and offspring, and also data on the dates and diagnoses of hospital care based on clinical assessments. The Register of the Social Insurance Agency (RSIA) provided information on diagnosis-specific disability pension. All diagnoses in the NPR, CDR and RSIA were classified in accordance with ICD-8, ICD-9 and ICD-10.
+The Population and Housing Censuses (PHC) provided data on parental socioeconomic status, maternal marital status, and parental education in 1970. Data on parental education in 1990 were retrieved from the Longitudinal Integration Database for Health Insurance and Labour Market Studies (LISA). Information on emigration and immigration was extracted from the Register of the Total Population. Table 1 summarises details of the variables and registers.
+Exposure variables
+The exposure assessment for each case and control refers to the time up until the event involving the index subject. Parental markers of morbidity, namely diagnosis-specific disability pension, attempted suicide, and inpatient care due to mental disorder, were categorised into age windows corresponding to offspring’s age at first exposure. The categories were exposure before the age of 3 years, between 3 and 10, and over the age of 10. We considered only the main diagnosis, and the first date of
+hospital inpatient care or receipt of disability pension. Also, diagnoses before the birth of the child were allocated to the youngest exposure window, since it was deemed likely that chronic illness underlying the markers of morbidity would result in early exposure of the child. Parental markers of mortality, namely suicide completion and death due to other causes, were categorised into exposures before and after the age of 10 years due to relatively small numbers of cases. Whenever the child was exposed to a marker that was positive for both parents, the age window corresponding to the earlier exposure was used. Absence of exposure was employed as the reference category.
+Covariates
+Maternal and paternal age at offspring’s exposure were used as continuous variables. Parental education was measured as education up to 10 years (primary education), 10 to 13 years (secondary education), and >13 years (university education, reference category). We used either maternal or paternal education, whichever was the higher, in accordance with the dominance principle which has been shown to perform well in classifying families.15 Whenever both parents were born in 1940 or earlier we used educational data from 1970. These parents were at least 30 years old at time of measurement. For all other parents, we used educational data from 1990. Categories for parental socioeconomic status were unskilled workers, skilled workers, low level salaried employees, intermediate or high level salaried employees (reference category), and others. We used either maternal or paternal socioeconomic status, whichever was the higher. Data from the 1980 census were used for events that occurred between 1983 and 1990, 1985 census data for events between 1991 and 1998, and 1990 data for events from 1999 to 2006. In the case of internal missing data, we took information from the preceding census. This approach ensured that all measurements were taken at a point in time preceding offspring’s suicidal behaviour.
+Maternal marital status was dichotomised into married and cohabiting versus other status (unmarried, divorced, widowed, non-cohabiting), using the same approach as for socioeconomic status with regard to the choice of census year.
+There were missing data on covariates in 4.0% of cases of attempted suicide, and in 4.2% of cases of completed suicide. Missing data were coded as a separate category. A sensitivity analysis showed similar patterns of suicide and attempted suicide risks in cases with complete information and in cases with missing data.
+Statistical analysis
+The outcome variables were completed suicide and attempted suicide in offspring. Univariate and multivariate ORs for the exposure variables were estimated using conditional logistic
+234
+J Epidemiol Community Health 2012;66:233-239. doi:10.1136/jech.2010.109595
+regression. Interactions between offspring’s sex, offspring’s age at onset of suicidal behaviour as well as maternal marital status and the exposure variables were subjected to partial likelihood ratio tests by introducing corresponding product terms into the adjusted models.
+Attributable proportions for the fully adjusted model were calculated using the formula AP=S(1—(1/OR)), where S is the number of cases with exposure divided by the total number of cases.
+Further, in order to test for a possible linear trend across offspring’s age at exposure, we included offspring’s age at exposure as a continuous variable in the fully adjusted models. These analyses were restricted to exposed individuals only.
+Data processing was performed using SPSS V15.0 for Windows.
+RESULTS
+The numbers of suicide victims and suicide attempters, respectively, were 1019 (72.4%) and 6003 (35.0%) boys/men, and 388 (27.6%) and 11 156 (65.0%) girls/women. Mean age was 22.3 years (SD 3.7) at suicide completion, and 21.1 years (SD 4.4) at first registered attempted suicide. With regard to suicide methods, for certain suicides (N—1077), the main method was hanging (41.0%), followed by poisoning (22.1%) and jumping (19.0%). For uncertain suicides (N—330), poisoning (67.3%) was followed by other methods (10.4%) and jumping (10.0%). For certain attempted suicides (N—13 976), poisoning (88.9%) was the most frequent method, followed by cutting (4.2%). For uncertain attempted suicides (N—3183), the main method was poisoning (76.2%), followed by other methods (17.2%). Of attempted suicides, 14 853 (86.6%) were coded as non-violent attempts, and 2306 (13.4%) as violent.
+A majority of all exposures to disability pensions attributed to parental mental disorder (N—9145) were coded under the headings neurotic, stress-related and somatoform disorders (48.0%), and affective disorders (22.9%).
+In exposures to disability pensions granted due to somatic disorders (N—25432), disorders of the musculoskeletal system and connective tissue were the most frequent diagnoses (69.2%).
+Tables 2—5 show the results of the univariate and multivariate analyses.
+Parental suicidal behaviour
+Exposure to parental attempted suicide was associated with a 3.4-fold increase in suicide risk (table 2), and a 3.5-fold increase
+in risk of attempted suicide (table 3) in offspring. The ratios decreased to 2.6 and 2.6, respectively, in the fully adjusted models.
+Exposure to parental suicide was associated with a 3.5-fold increase in suicide risk (table 2), and a 2.6-fold increase in risk of attempted suicide (table 3). The ratios decreased to 2.5 and 1.8, respectively, in the fully adjusted models. Adjustment for parental age alone had little effect on the estimates (tables 2 and 3). At the second adjustment step, parental education and marital status had the largest effects.
+The risk of completed suicide was most pronounced among offspring exposed to parental completed suicide in the earliest age window (table 4). There was a constant and an increasing risk of attempted suicide among offspring across exposure windows in relation to parental attempted suicide and parental completed suicide, respectively (table 5). Separate analyses of violent and non-violent attempted suicides in offspring revealed that there was one slight alteration to the general pattern: the risk of violent attempted suicide after parental suicide was 1.6 (95% CI 1.0 to 2.7) when exposed up to the age of 10 years, and 1.4 (95% CI 0.9 to 2.3) when exposed over the age of 10, indicating a more pronounced risk of violent attempted suicide in offspring in the case of early exposure to parental suicide.
+Parental diagnosis-specific disability pension
+Parental disability pension due to a psychiatric disorder was associated with a 2.9- and 2.7-fold risk of completed suicide (table 2) and attempted suicide (table 3) in offspring, respectively. The estimates fell to 1.9 and 1.7, respectively, after full adjustment. Exposure to parental somatic disability pension had a smaller effect than exposure to psychiatric disability pension, and was associated with a 1.5-fold risk of suicide completion (table 2), and a 1.7-fold risk of attempted suicide (table 3). The estimates decreased to 1.3 and 1.5, respectively, in the adjusted models. Offspring’s risk of completed and attempted suicide showed a similar pattern with regard to the timing of first exposure to parental diagnosis-specific disability pension. The earlier the exposure, the greater was the risk of suicidal behaviour (tables 4 and 5). There was a significant increase in suicide risk with decreasing age at exposure to psychiatric disability pension (p—0.027) (table 4). Further, there was a significant general increase in attempted suicide risk with decreasing age at exposure to somatic disability pension (p—0.001, table 5).
+Parental inpatient care due to mental disorder
+Parental inpatient care due to mental disorder was associated with a 2.6-fold risk of completed suicide (table 2) and a 3.0-fold risk of attempted suicide (table 3). The estimates decreased to 1.8 and 2.1, respectively, in the fully adjusted models, largely due to the effects of parental education and maternal marital status. The risks of both completed and attempted suicide were highest in
+the case of exposure in the youngest age window (tables 4 and 5). There was an increase in attempted suicide risks with decreasing age at first exposure (p<0.001, table 5).
+Parental death due to causes other than suicide
+Parental death due to causes other than suicide was associated with a 1.7- and 1.6-fold risk for suicide and attempted suicide
+among offspring, respectively (tables 2 and 3). ORs decreased to 1.3 and 1.3, respectively, in the adjusted models. In contrast to other markers of parental morbidity, exposure to parental non-suicidal death after the age of 10 years was associated with an increased risk of attempted and completed suicide only in offspring exposed over the age of 10 (tables 4 and 5). There was a significant increase in attempted suicide risk with increasing age at exposure to parental non-suicidal death (p=0.018) (table 5).
+Interaction effects of offspring’s sex, offspring’s age at onset of suicidal behaviour and maternal marital status with the exposure variables
+Partial likelihood ratio tests indicated that exposure to parental death due to other causes than suicide increased the risk of attempted suicide more in boys/men than in girls/women (p=0.006). In comparison to non-exposed girls/women, the estimates for exposed boys/men and exposed girls/women were 1.40 (1.26 to 1.55) and 1.15 (1.05 to 1.26), respectively.
+Furthermore, exposure to parental somatic disability pension increased the risk of attempted suicide more when onset of suicidal behaviour was before the age of 20 years, than in the case of later onset (p=0.012). Compared with unexposed offspring with onset after the age of 20 years, the respective estimates were 1.66 (1.54 to 1.80) for exposed offspring with onset before 20 years and 1.46 (1.39 to 1.55) for exposed offspring with onset after the age of 20 years.
+With regard to suicide attempt in offspring, there were several interaction effects between parental markers of morbidity and
+maternal marital status. Specifically, exposure to parental attempted suicide (p<0.001), parental inpatient care due to mental disorders (p=0.002), parental psychiatric disability pension (p=0.05) and parental somatic disability pension (p=0.001) interacted with maternal marital status. In comparison with unexposed offspring of married and cohabiting mothers, the suicide attempt risk for offspring of unmarried, divorced, widowed and non-cohabiting mothers exposed to parental attempted suicide were 2.51 (2.11 to 2.99), and the estimates for exposed offspring of married and cohabiting mothers were 1.87 (1.56 to 2.24). The respective estimates for exposure to parental inpatient care due to mental disorders, parental psychiatric and somatic disability pension were 3.17 (2.99 to 3.35), 2.69 (2.49 to 2.91) and 2.33 (2.18 to 2.49) for offspring of unmarried, divorced, widowed and non-cohabiting mothers, respectively, and 2.22 (2.10 to 2.36), 1.83 (1.67 to 2.02) and 1.58 (1.49 to 1.68) for exposed offspring of married/cohab-iting mothers, respectively.
+DISCUSSION
+Main findings
+A general pattern of increasing risk of suicide and attempted suicide in offspring with decreasing age at exposure to parental risk factors emerged. This pattern was present for parental somatic disability pension, parental inpatient care due to mental disorders and parental psychiatric disability pension. With regard to offspring’s risk of completed suicide, the pattern was
+also present for parental suicide and attempted suicide. For parental non-suicidal deaths, the pattern was the opposite.
+Strengths and limitations
+The main strengths of the present register study were the large number of participants, the population-based design, the full coverage of cases, the opportunity to control for potential confounders, the high quality of the data10-14 and the very low dropout. Difficulties, such as recall bias, which is often present in studies based on data from clinical settings, could be avoided due to the use of national registers.
+The study also has some limitations. For the Register of the Social Insurance Agency, which provided data on disability pension, misclassifications within the group of psychiatric diagnoses, entailing an under-reporting of psychotic disorders, have been discussed previously.16 This register has not yet been evaluated in detail. Furthermore, we could not address the effect of cumulative exposure to morbidity and mortality in both parents due to the relatively small number of cases in several exposure windows. A recent study identified a clear increase in attempted suicide risk in the offspring when suicidality or psychiatric disorders occurred in several family members.2 Further, the present data on attempted suicides and mental disorders only covered individuals who were hospitalised. It is estimated that approximately 25% of suicide attempters receive inpatient care.17 The findings, therefore, only apply to attempts that resulted in hospitalisation.
+Comparison with previous studies
+Parental suicidal behaviour2 4 7 and parental inpatient care due to mental disorders,2 4 as well as parental non-suicidal death,4 have previously been identified as risk factors of suicidal behaviour in offspring. Knowledge of possible consequences of being on disability pension is limited.18 Receiving a disability pension seems to increase an individual’s own risk of suicide beyond the effects of his or her psychiatric hospitalisation and socioeconomic status.18 19 A recent study identified an increased suicide risk in offspring exposed to parental disability pension, but did not consider the nature of the parental diagnoses involved.6 In the present study, exposures to both parental psychiatric and somatic disability pension were associated with an increased risk of suicide and attempted suicide.
+Concerning possible impacts of age at exposure to parental risk factors on the risk of suicidal behaviour in offspring, earlier studies found that the risk of bipolar disorder, which is an important risk factor of suicidal behaviour,20 was most pronounced when exposure to maternal suicide21 or parental loss22 occurred before the age of 10 years. A recent study analysed the impact of offspring’s age at exposure to parental suicide on the risk of completed suicide and identified an increased risk for offspring exposed under the age of 17 years compared to offspring exposed later in life.7 The present findings add that children exposed at the very youngest ages show an important vulnerability that may lead to suicidal behaviour later in life.
+Interpretation
+The present study was the first to identify an association between parental somatic disability pension and offspring’s risk of suicidal behaviour. The association was present for both sexes and remained after controlling for parental suicidal behaviour, parental socioeconomic conditions and marital status. This finding may be related to the genetic transmission of parental somatic disorders, or to uncontrolled co-morbidity of mental
+disorders in this group, or to the psychosocial consequences of having a parent on disability pension.
+The finding of more pronounced risks of attempted suicide in offspring of unmarried, non-cohabiting, divorced and widowed mothers stresses the importance of a specific awareness and support from the environment and healthcare system.23 Other socioeconomic indicators, specifically the parental socioeconomic index and parental education, did not significantly alter the effects of exposure to parental morbidity and mortality.
+In accordance with the hypothesis of sensitive life periods,5 8 exposure to parental somatic disability pension, parental inpatient care due to mental disorders and psychiatric disability pension, seem to increase the susceptibility to attempted suicide and suicide in offspring most when exposure appears early in life.
+Similarly, exposure to parental suicide may increase the susceptibility to completed suicide. Of note, the present risk estimates were only slightly attenuated by controlling for parental age at offspring’s exposure, which is a crude marker of parental age at onset of morbidity. In previous research, early-onset major depression in offspring, which is an important risk factor of suicidal behaviour, has been discussed to be caused by a stronger genetic disposition that may be reflected in early onset of major depression in parents.24
+Compared to the risk pattern in completed suicide, the risk of attempted suicide in offspring exposed to parental suicidal behaviour was more pronounced for later exposure windows. This stronger effect in adolescence and young adulthood may underscore the importance of parental suicidal behaviours as triggering events for non-fatal suicidal behaviour in offspring. Of note, imitation has been discussed as an important factor in suicidal behaviour, particularly among young people.25 A triggering effect may also be reflected in the association of parental non-suicidal death and suicidal behaviour in offspring exposed after the age of 10 years.
+The risk patterns of attempted suicides varied somewhat with suicide method. The risk patterns for violent attempt in offspring exposed to parental suicide resembled more the patterns estimated for completed suicide than for attempted suicide in general. Related research suggests that violent attempts constitute a further step in the suicidal process, and that medically serious attempted suicides and suicides are two overlapping populations that share common psychiatric diagnostic and history features.9 26 Future research is warranted to further scrutinise the present findings in this context.
+CONCLUSION
+The present findings comply with the ‘sensitive period’ hypothesis in life course epidemiology8 and add to our understanding of suicidal behaviour in the familial context during the life span. Parental disability pension, parental psychiatric inpatient care and parental suicide seem to increase the offspring’s susceptibility to suicide, particularly when they occur early in life. Parental non-suicidal death has most detrimental effects when occurring in adolescence or young adulthood. Early interventions in families with parental morbidity or suicidal behaviour seem necessary to prevent suicide in offspring. Evaluations of such interventions suggest possible preventive effects. An early intervention for families at psychosocial risk resulted in short-term improvements in motherechild relations.27 28 The six-week intervention focused on strengthening the mothers in their care giving skills and improving the motherechild relationship and interaction. Another study, which targeted early intervention through a five-year family counselling programme

Factors Influencing Professional Help-Seeking for Suicidality.txt

@@ -0,0 +1,67 @@
+Suicide is a global public health issue that causes around 800,000 deaths every year (World Health Organization, 2014). Many suicides could be prevented if individuals with suicidal thoughts or behavior sought help from appropriate health services that met their needs (Trueland, 2014). However, a significant proportion of individuals with suicidal thoughts or behavior do not seek help (Bruf-faerts et al., 2011; Luoma, Martin, & Pearson, 2002; Pitman & Osborn, 2011). Two published reviews examined the factors that influence help-seeking for suicidal behavior and self-harm, but focusing exclusively on adolescents (aged from 11 to 19) and young adults (aged up to 26; Michelmore & Hindley, 2012; Rowe et al., 2014). Several factors such as stigmatization, fear of confidentiality being breached, and high self-reliance have been highlighted as barriers to help-seeking. However, it is not clear if the factors identified among young people are applicable to all age groups. Exploring a wider age cohort is important since low rates of help-seeking for suicidal ideation and behaviors (suicidality) are not specific to young people, but are also evident in older adults (Lee, Lin, Liu, & Lin, 2008).
+Previous reviews exploring help-seeking factors for suicidality have also examined the factors that influence both
+© 2017 Hogrefe Publishing
+professional and nonprofessional help-seeking sources. The current review focuses specifically on professional help-seeking to refine the findings, as professionally trained health workers are more likely to provide evidence-based treatments and accurate information than family or friends are, who may find it difficult to recognize potentially unhealthy thinking states that may lead to suicide (Owens, Lambert, Donovan, & Lloyd, 2005). In line with the World Health Organization, the current review defines professional help-seeking as use of professional support such as health services, formal social institutions, or professional care providers, either in the public or private sector (Gary, 2007). In consideration of the benefits and low rates of professional help-seeking for suicidality internationally, a review to summarize the current findings may help inform researchers and health professionals of the knowledge gaps and barriers to professional help-seeking, and help develop high-quality suicide prevention campaigns.
+This review aims to explore the existing literature of factors that influence professional help-seeking along the progression of suicide, from suicidal ideation to suicidal behavior and death by suicide. For this purpose, this review presents three distinct research foci: (a) professional
+Crisis (2018), 39(3), 175-196
+https://doi.org/10.1027/0227-5910/a000485
+help-seeking intentions for perceived suicidal ideation, among people with or without suicidality; (b) professional help-seeking behavior among people with suicidality; and (c) suicidal decedents’ health services use. To our knowledge, this is the first review to summarize the factors influencing professional help-seeking for suicidality across these multiple study designs.
+Method
+Search Strategy
+Relevant studies in English were identified using Medline and PsycInfo, accessed through the Ovid interface. Databases were searched between January 1, 1995, and December 31, 2015. Search terms used were: suicid* AND (‘seek*’ OR ‘service use’ OR ‘service util*’ OR ‘getting’ OR ‘care util*’) AND (‘help*’ OR ‘treat*’ OR ‘service*’ OR ‘care*’) within abstract or title. References from identified studies were also reviewed for completeness.
+Selection of Studies
+A total of3,183 research papers were identified through the search strategy. Of these, 1,912 abstracts were
+screened, after removing non-human studies, duplicate abstracts, and non-English language articles. The first author (JH) examined all titles and abstracts, and 109 articles were retained based on the following inclusion criteria: (a) published in a peer-reviewed journal; (b) not an editorial or a review; (c) examined the factors that influence professional help-seeking intentions for suicidality, or examined the factors that influence professional help-seeking behavior among individuals with suicidality or suicide decedents. Studies containing both professional and nonprofessional help-seeking sources such as family, friends, and intimate partners were included, but only the results from professional help-seeking sources were coded.
+The full text of the 109 articles was obtained and double-coded by the first author and one of the coauthors independently (either PJB, ALC or RR). Disagreements between reviewers were resolved through discussion with a third coder. This process resulted in 54 additional research papers being excluded, and 55 total studies being included in the review. Figure 1 depicts the PRISMA flow diagram for inclusion.
+Data Extraction
+All papers were coded using a pro-forma coding sheet to identify a range of study characteristics. Factors that influence professional help-seeking were coded based on a
+narrative synthesis instead of a meta-analysis due to the diverse nature of the study designs and outcomes included in this review. Factors were marked as adjusted if potential confounding variables were accounted for in statistical analysis. Otherwise, factors were marked as unadjusted. To evaluate the quality of the included studies, the assessments of quantitative studies (Glasziou, Irwig, Bain, & Colditz, 2001) and qualitative studies (Mills, Jadad, Ross, & Wilson, 2005) were used. Studies that used both quantitative and qualitative methods were coded with respect to the relevant assessments. The first author and one of the coauthors independently (either PJB, ALC or RR) coded all the studies. Any disagreement was resolved by discussion. The results are presented in Table A1 and A2 in the appendix. The majority of the included studies were of sound quality.
+Results
+Study Characteristics
+In total, 55 publications met the inclusion criteria for the review. Studies were classified based on three distinct research foci; specifically 15 (27%) studies examined professional help-seeking intentions for perceived suicidal ideation, among people with or without suicidality (Type I), 21 (38%) examined professional help-seeking behavior among people with suicidality (Type II), and 19 (35%) studies examined suicidal decedents’ health services use (Type III).
+The sample size ranged from 107 to 1,896 participants (Mdn = 302) in the Type I studies, from 15 to 5,100 (Mdn = 543) in the Type II studies, and from 49 to 19,426 (Mdn = 370) in the Type III studies. The proportion of females ranged from 0% to 78% (Mdn = 65%) in the Type I studies, from 0% to 87% (Mdn = 63%) in the Type II studies, and from 3% to 40% (Mdn = 25%) in the Type III studies. The characteristics of the included studies are presented in Table B1 in the appendix. Only the factors influencing professional help-seeking were coded and analyzed. Table 1 presents a summary of the factors being reported five times or more across the three types of studies.
+Professional Help-Seeking Intentions for Perceived Suicidal Ideation Among People With or Without Suicidality
+Of the studies, 15 examined professional help-seeking intentions for perceived suicidal ideation (Table B1); 13 (87%) were conducted in Australasia, one in Japan, and one
+in the United States. Twelve studies (80%) were among a student population. All the studies used self-reported measurements; 11 studies (73%) used the General Help-Seeking Questionnaire (GHSQ; Wilson, Deane, Ciarrochi, & Rick-wood, 2005) to measure help-seeking intentions for perceived suicidal ideation. Another four studies used similar questions asking participants’ willingness to seek help for perceived suicide ideation. It is noticeable that the definition of professional help varies across studies, from psychologist only, to mental health professionals, or to a broader range of professionals including family doctor, physician, phone helpline, religious/spiritual leader, and social workers.
+Age was identified as a factor associated with professional help-seeking for perceived suicidal ideation with mixed results. Older age was associated with lower help-seeking intentions among elderly Japanese adults (Sakamoto, Tanaka, Neichi, & Ono, 2004), while younger Australians had lower help-seeking intentions (Calear, Batterham, & Christensen, 2014). In a study of 527 New Zealand prisoners with a wider age range from 16 to 72 years, older age was positively associated with greater intentions to seek psychological help in prison (Skogstad, Deane, & Spicer, 2006). Male subjects were usually found to have lower help-seeking intentions than their female counterparts (Calear et al., 2014; Ciarrochi & Deane, 2001).
+The severity ofmental health issues including depression (Calear et al., 2014; Wilson & Deane, 2010) and psychological stress (Wilson, Deane, Marshall, & Dalley, 2010) was negatively associated with professional help-seeking intentions, while the effect of anxiety symptoms was in the opposite direction - it facilitated help-seeking intentions (Calear et al., 2014). Seven studies (47%) supported the existence of a help-negation effect of suicidal ideation on help-seeking intentions - the presence of suicidal ideation is associated with lower help-seeking intentions (Calear et al., 2014; Carlton & Deane, 2000; Deane, Wilson, & Ciarrochi, 2001; Wilson & Deane, 2010; Wilson, Deane, & Ciarrochi, 2005; Wilson et al., 2010; Yakunina, Rogers, Waehler, & Werth, 2010).
+Increased stigma toward people who die by suicide measured by the Stigma of Suicide Scale (SOSS) (Batterham, Calear, & Christensen, 2013) was significantly associated with lower help-seeking intentions toward mental health professionals among Australian adults in the community (Calear et al., 2014); however, the effect of stigma was not significant in a study of 321 US university students measured by a different scale (the Stigma of Suicide Scale, SSS; Yakunina et al., 2010), nor in a study of Australian university students (measured by SOSS; Chan, Batterham, Christensen, & Galletly, 2014). In addition, stigma toward mental illness was addressed as an important barrier to help-seeking in a qualitative study among eight New Zealand university students (Curtis, 2010).
+Positive attitudes toward psychological help measured by the Attitudes Towards Seeking Professional Psychological Help Scale (Fischer & Farina, 1995) were found to facilitate help-seeking intentions (Carlton & Deane, 2000; Deane & Todd, 1996; Skogstad et al., 2006; Yakunina et al., 2010), while high self-stigma measured by the Self-Stigma of Seeking Help Scale (SSOSH; Vogel, Wade, & Haake, 2006) and perceived stigma of seeking help measured by the Stigma Scale for Receiving Psychological Help (SSRPH; Komiya, Good, & Sherrod, 2000) were significantly associated with low help-seeking intentions (Yakunina et al., 2010). Prior treatment was reported as a facilitator to help-seeking intentions (Carlton & Deane, 2000; Ciarrochi & Deane, 2001). Skogstad et al. (2006) suggested, however, that the characteristics of prior treatment may influence whether it has a positive, negative, or null relationship with help-seeking.
+Although willingness to seek help from nonprofessionals such as a partner and family is associated with higher intentions to seek help from professionals (Wilson, Rick-wood, Bushnell, Caputi, & Thomas, 2011), a quantitative study conducted by Yakunina et al. (2010) among US college students suggests that social support only facilitates help-seeking from nonprofessional sources, but not from professional sources.
+Self-reliance was found to be a barrier ofhelp-seeking intentions for suicidal thoughts among a sample of New Zealand university students (Curtis, 2010). A quantitative study among Australian university students found a significant but weak negative relationship between autonomy and formal help-seeking intentions (Wilson et al., 2011). Additional potential barriers of help-seeking intentions include hopelessness (Ciarrochi & Deane, 2001) and difficulty in identifying and describing emotions (Ciarrochi, Wilson, Deane, & Rickwood, 2003; Ciarrochi & Deane, 2001).
+Professional Help-Seeking Behavior Among People With Suicidality
+In all, 21 studies examined professional help-seeking behavior among people with suicidality (Table B1). Of these, 16 (76%) were conducted in the United States or Canada, three (14%) in Australia, and two (10%) in Europe. Only eight studies (38%) specified that the purpose of service use was relevant to suicidality. The other studies did not give information on whether people had disclosed their suicidality to professionals or not. Only one study (5%) used actual service use records, while the majority (95%) of the findings were based on self-reported service use.
+In the identified studies, older age was generally associated with higher self-reported service use among people
+with suicidality (Ahmedani et al., 2012; De Leo, Cerin, Spathonis, & Burgis, 2005; Encrenaz et al., 2012). Females were more likely to talk about suicidal ideation to health professionals (Encrenaz et al., 2012) and use health services (Ahmedani et al., 2012; De Leo et al., 2005; Routhier, Leduc, Lesage, & Benigeri, 2012; Wong, Brownson, Rutkowski, Nguyen, & Becker, 2014) than males were. Individuals of ethnic minorities including Latino (Ahmedani et al., 2012; Downs & Eisenberg, 2012; Freedenthal, 2007; Meyer, Teylan, & Schwartz, 2015; Wu, Katic, Liu, Fan, & Fuller, 2010), Black (Ahmedani et al., 2012; Freedenthal, 2007; Meyer et al., 2015; Wu et al., 2010), and Asian (Ahmedani et al., 2012; Downs & Eisenberg, 2012; Wong et al., 2014) tended to have less help-seeking behavior compared with Caucasian.
+The presence of mental health issues such as depression (Ahmedani et al., 2012; Arria et al., 2011; Encrenaz et al., 2012; Husky et al., 2012; Vasiliadis, Gagne, Jozwi-ak, & Preville, 2013; Wu et al., 2010), anxiety (Ahmedani et al., 2012; Arria et al., 2011; Encrenaz et al., 2012; Wu et al., 2010), and substance use (Encrenaz et al., 2012; Freedenthal, 2007; Routhier et al., 2012) was associated with greater help-seeking behavior among people with suicidality. In addition, severe suicidal intent (De Leo et al., 2005; Wong et al., 2014), history of suicidal ideation (Husky et al., 2012; McKibben et al., 2014; Pagura, Fotti, Katz, Sareen, & Tea, 2009), suicidal plans (Encrenaz et al., 2012; Husky et al., 2012), and suicidal attempts (Ballard et al., 2014; De Leo et al., 2005; Encrenaz et al., 2012; Freedenthal, 2007; McKibben et al., 2014; Milner & De Leo, 2010; Pagura et al., 2009) also facilitated help-seeking behavior.
+Four qualitative studies highlighted the negative influence of stigma toward mental health issues or suicide on help-seeking behavior (Czyz, Horwitz, Eisenberg, Kramer, & King, 2013; De Leo et al., 2005; Freedenthal & Stiffman, 2007; Strike, Rhodes, Bergmans, & Links, 2006). In addition, stigmatizing attitudes toward mental health services also impeded people’s help-seeking behavior (Czyz et al., 2013; De Leo et al., 2005; Downs & Eisenberg, 2012). Two quantitative studies reported prior treatment as a facilitator to help-seeking behavior (Arria et al., 2011; Milner & De Leo, 2010), although other studies suggested that the relationship between these factors depended on the specific setting of treatments such as hospital emergency department (Routhier et al., 2012), level of satisfaction with prior treatment (Czyz et al., 2013; Strike et al., 2006), and whether the relationship between clients and medical helpers was consistent (Osvath, Michel, & Fekete, 2003).
+Family and friends’ influence on help-seeking behavior was complicated. On one hand, disclosure of suicidal intentions to surrounding people (De Leo et al., 2005; En-crenaz et al., 2012; Wong et al., 2014) and being referred
+to professionals by family and friends (De Leo et al., 2005; Wong et al., 2014) are associated with higher likelihood of service use. On the other hand, findings also suggest that warm and trusting relationships may impede therapy or medication as they may decrease the subjective perception of distress (Downs & Eisenberg, 2012) and the need for professional help (Czyz et al., 2013).
+In addition, no perceived need for treatment was identified as the most frequently perceived reason for not seeking professional help in two studies ofUS students: 66% (Czyz et al., 2013) and 29% (Freedenthal & Stiffman, 2007) of students cited this reason as the primary motivation for not seeking help. Furthermore, in a third US study suggested that perceived need for help significantly facilitated treatment use among students with serious previous suicidal thoughts (Downs & Eisenberg, 2012).
+Self-reliance was reported to be a barrier for help-seeking behavior in two of the US qualitative studies (Czyz et al., 2013; Freedenthal & Stiffman, 2007) identified in the current review. Religion was also found to influence youth suicide attempters’ mental health service use in a qualitative study of Canadians, although it is unclear whether it facilitates or impedes health service use (Bullock, Nadeau, & Renaud, 2012). Pragmatic barriers including lack of time, long waitlists, financial difficulties, and lack of transportation or practicing GPs (Ahmedani et al., 2012; Czyz et al., 2013; Freedenthal, 2007; Osvath et al., 2003) were also identified as factors impacting help-seeking among the identified studies.
+Suicidal Decedents’ Health Services Use
+Professional help-seeking behavior among suicidal decedents (Table B1) was examined in 19 studies. Four studies (21%) were conducted in Canada, four (21%) in the UK, two (11%) in the United States, two (11%) in Australia, two (11%) in Hong Kong, two (11%) in Taiwan, one (5%) in Sweden, one (5%) in Korea, and one (5%) in Singapore. Six studies (31%) collected survey or interview data from decedents’ family, friends, carers, or community members. Other studies used the data from administrative data or patient records. All the studies did not specify whether decedents had disclosed their suicidal intentions to professionals during service use.
+Females were generally reported to have greater service use compared with males (Chang et al., 2012; Cho et al., 2013; Hamdi, Price, Qassem, Amin, & Jones, 2008; Law, Wong, & Yip, 2010; Lee et al., 2008; O’Neill, Corry, Murphy, Brady, & Bunting, 2014; Rhodes et al., 2013; Vasili-adis, Ngamini-Ngui, & Lesage, 2015; Vassilas & Morgan, 1997). Being single (Basham et al., 2011), employed (Hamdi et al., 2008; Law et al., 2010; Loh, Tai, Ng, Chia,
+& Chia, 2012), having a low income (Law et al., 2010), problem gambling (Seguin et al., 2010), and being in a rural area (Vasiliadis et al., 2015) were associated with less service use before suicide.
+Previous diagnosis of mental disorders (Hamdi et al., 2008; Law et al., 2010; Law, Wong, & Yip, 2015; Vasiliadis et al., 2015), anxiety (De Leo, Draper, Snowdon, & Kolves, 2013), mood disorders (De Leo et al., 2013), or problematic alcohol use (Sveticic, Milner, & De Leo, 2012) was significantly related to increased contact with health services among suicidal decedents. A family history of suicide facilitated mental health service use among young suicide decedents in Singapore (Loh et al., 2012). A low level of suicide intent (Law et al., 2010), and history of suicidal attempts (Hamdi et al., 2008; Kisely, Campbell, Cartwright, Bowes, & Jackson, 2011; Loh et al., 2012; Sveticic et al., 2012) also increased help-seeking behavior among suicidal decedents. Previous contact with primary care services and a GP was found to facilitate mental health service use among suicidal decedents in a UK study (Hamdi et al., 2008). In addition, shame and stigma surrounding mental health issues (Moskos, Olson, Halbern, & Gray, 2007; Tornblom, Werbart, & Rydelius, 2015) were described as barriers to seeking professional help by the informants in two of the qualitative studies identified.
+In another of the qualitative studies, 36% of suicidal decedents had been persuaded to seek medical help by a close friend or relative (Owens et al., 2005), suggesting that members of the family and immediate social network may play a key role in determining whether or not suicidal individuals would seek help from a medical practitioner. Furthermore, suicidal decedents who did not seek help were depicted as self-reliant and resourceful characters who were expected to be able to solve their own problems by informants (Owens et al., 2005).
+Discussion
+To our knowledge, this is the first systematic review delineating the factors that influence professional help-seeking for suicidality across three stages: professional help-seeking intentions for perceived suicidal ideation, among people with or without suicidality, help-seeking behavior among people with suicidality, and service use of suicidal decedents. Several potentially important barriers were identified. To sum up, low perceived need for treatment, high self-reliance, and stigmatizing attitudes toward suicide and/or mental health issues were identified as barriers to professional help-seeking across the three subgroups of studies, although much of the evidence was from the studies using qualitative design. The presence of suicid
+ality and other mental health issues, such as depression and psychological stress, generally reduced help-seeking intentions for perceived suicidal ideation, while facilitating help-seeking behavior among people with suicidality or suicidal decedents. The influence of social support on professional help-seeking was more complex. People with good social support seemed to favor informal sources, which might reduce professional help-seeking. However, there was also evidence that referrals to professionals by family or friends facilitated service use among individuals with suicidal thoughts and behavior. Other factors such as treatment history, exposure to suicide, and knowledge of suicide were found to influence help-seeking intentions and behavior; however, evidence was scant or too diverse to draw robust conclusions. The findings on demographic factors were somewhat inconsistent, suggesting local factors such as culture, ethnicity, and religion may need to be taken into consideration.
+Suicidality and Mental Health Issues
+Suicidality and a number of mental health-related factors were significantly associated with professional help-seeking. Higher levels of suicidal ideation, depression, and psychological stress were associated with lower help-seeking intentions for perceived suicidal ideation, while a history of suicidality and mental health issues tended to facilitate help-seeking behavior among people with suicidality and suicidal decedents. The underlying basis has not been identified. One possible explanation is that people with a history of suicidality and mental health issues may have to receive physical or mental health treatments in medical settings regardless of whether they are willing to seek care or not.
+Attitudes and Knowledge of Suicide and Mental Health Issues
+Stigmatizing attitudes and a limited knowledge of suicide and mental health issues were identified as barriers to seeking professional help, although much of the research was qualitative in nature. The lack of quantitative studies exploring these factors may reflect the limited number of validated scales that are available to measure suicide attitudes and knowledge (Batterham et al., 2013). Further investigation of these relationships in quantitative studies is warranted, and in particular there is a need for research to distinguish between attitudes and knowledge, and to differentiate between self-stigma and personal and perceived stigma.
+Attitudes Toward Professional Treatments, Perceived Need for Treatment, and Treatment History
+Attitudes toward health professional treatments were associated with both help-seeking intentions and behavior across the three subgroups of studies. Positive attitudes toward professionals facilitated professional help-seeking. Meanwhile, a lack of perceived need for treatment was the most frequently self-reported barrier to professional help-seeking among people with suicidality. However, the relationship between treatment history and help-seeking was not consistent across all of the identified studies. The quality and satisfaction level of previous treatment may therefore need to be taken into consideration when studying the effects of previous treatment on help-seeking. All of these results highlight the important role that professional treatments play in the help-seeking process.
+Self-Reliance and Social Support
+Self-reliance was a frequently reported barrier in the studies identified in the current review for both help-seeking intentions for perceived suicidal ideation and actual help-seeking behavior for suicidality. Extreme views on the value of self-reliance may contribute to self-stigma, whereby people’s negative attitudes about help-seeking combined with self-reliance prevent them from disclosing their symptoms or engaging in professional treatment even in the face of dangerous mental health symptoms (Labouliere, Kleinman, & Gould, 2015). However, only one quantitative study suggested a significant but weak negative effect of self-reliance on predicting intentions to seek help from a mental health service among a student sample (Wilson et al., 2011). All other studies reporting on self-reliance are qualitative.
+The influence of social support on professional helpseeking was complex. In some studies, seeking support from informal sources and having better social support facilitated professional help-seeking for suicidality, while in others people with good social support could rely on their social network instead of professional services (Downs & Eisenberg, 2012). It is also of note that encouragement from others was suggested to be an important motivation for seeking professional help: 80% of people with suicidality used health service because others thought it was important for them to (De Leo et al., 2005), while 36% of suicidal decedents had been persuaded to seek medical help by a close friend or relative (Owens et al., 2005).
+Limitations
+There are potential limitations of this review. First, different types of research, such as quantitative and qualitative studies, were included in the current review in order to report a greater coverage of findings. Although the majority of the included studies were of sound quality, the variability of the findings may relate to the inclusion of a variety of research designs in the review. Second, only results from published English-language studies were coded in this review, which may limit the generalizability of the findings. Third, many studies included in this review were based on students or selected samples such as prisoners and soldiers. Further investigation in broader community samples, particularly on perceived help-seeking intentions for suicidal ideation, may allow for greater generalization of the findings. Fourth, as noted throughout, some of the factors identified in the review were tested in relatively few studies, and most of the included studies were conducted in industrialized countries. As culture may have a considerable influence on help-seeking, further studies among developing countries and comparative studies are needed.
+amine whether participants had disclosed their suicidality to professionals or not. Future studies are therefore recommended to better delineate the purposes of help-seeking.
+Acknowledgments
+PJB and ALC are supported by National Health and Medical Research Council (NHMRC) Fellowships 1083311 and 1122544. This research is also supported by the NHMRC Centre of Research Excellence (CRE) 1042580 in Suicide Prevention, and PJB and ALC are Chief Investigators on this CRE.
+The authors declare no competing interests.
+Conclusion
+This review highlights several potentially important barriers to professional help-seeking across all types of studies. They include high self-reliance, lack of perceived need for treatment, and stigmatizing attitudes toward suicide, toward mental health problems, and toward seeking professional treatment, which warrant additional high-quality studies using quantitative and qualitative methods, especially in developing countries. The positive influence of suicidality and mental health issues on help-seeking behavior but not on intentions suggests people with suicidal thoughts or behavior may be passively engaged with health services, where stigmatizing attitudes and unpleasant previous treatment experience could discourage subsequent help-seeking behavior. Efforts to make mental health services more approachable, less coercive, and less stigmatizing are strongly encouraged. In addition, the findings of this review also suggest that enabling factors might be lacking among those without a history of suicidality or mental health issues. Future suicide prevention initiatives may need to target the broader community, connecting with individuals who have never used mental health services. The complex influence of social support on professional help-seeking also needs further investigation, especially in relation to the gatekeeper effects of family and friends. It is also notable that most of the studies among people with suicidality and suicidal decedents failed to ex

Financial-incentives-improve-recognition-but-not-treatment-of-cardiovascular-risk-factors-in-severe-mental-illnessPLoS-ONE.txt

@@ -0,0 +1,41 @@
+Introduction
+People with severe mental illnesses (SMI), including schizophrenia and bipolar affective disorder, are known to be at significant risk of premature morbidity and mortality. In the UK, individuals with SMI have a life expectancy around 12 years less than the general UK population, [1] and similar disparities are seen in the US.[2] Cardiovascular disease is a major contributor to this health inequality.[3]
+The UK Quality and Outcomes Framework (QOF) aims to improve quality in primary care through linking financial incentives to performance against indicators.[4] From its inception in 2004, the QOF has incentivised annual physical health review for people with SMI (S1 Appendix). The nature of this review was unspecified until 2011, when more explicit cardiovascular risk factor indicators were introduced; the majority of indicators were withdrawn in 2014.
+To date there has been little evaluation of the impact of the QOF on the recognition and management of cardiovascular risk factors in people with SMI. One study found the QOF incentives reduced inequalities in cardiovascular risk factor testing between those with and without SMI,[5] although the nature and time period of the analysis was relatively limited and did not evaluate risk factor detection. Elsewhere the QOF has been shown to have increased consultation rates[6] in people with SMI and also to have coincided temporally with an increase in recording of comorbidities[6,7].
+The current study builds on the above findings by exploring whether the QOF indicators have been associated with improvements in the identification and management of cardiovascular risk factors in people with SMI.
+Methods
+Anonymised data detailing diagnoses, prescribing, test results and demographics were extracted from the Clinical Practice Research Datalink (CPRD). CPRD captures data recorded by the general practitioner (GP) as part of routine care, and covers a 6.9% representative sample of the UK population.[8] Patients were included in the analysis during continuous periods of registration at those participating practices which met CPRD's internal quality standards. The study was approved by the CPRD Independent Scientific Advisory Committee (protocol reference 15_110RMn).
+Study design
+A retrospective open cohort design was used with cases having a lifetime SMI diagnosis and an unmatched population comparison group without SMI. The study period included the consecutive 'financial years' from 1st April 1995 to 31st March 2014. Two interventions were considered: intervention 1 from April 2004 (introduction of the QOF SMI annual review indicator), and intervention 2 from April 2011 (change to specific cardiovascular indicators).
+Data from April 1995 to March 2003 were used to ascertain trends in the outcome before the introduction of QOF. The 2003/04 year was excluded as per previous studies[6], as some practices were preparing for the introduction of the QOF in the year prior to its introduction.
+Within each year of analysis case and comparison group members could be eligible for the full financial year or for a proportion of the year. The first day of inclusion in the analysis was
+the latest of the patient reaching age 35, first day of continuous registration, the patient's GP practice meeting CPRD's data quality criteria, or (for cases) the date of SMI diagnosis. Eligibility for inclusion in the analysis ended at the first of having the outcome of interest, leaving the practice, death, or the practice ceasing to submit data to CPRD.
+Case and comparison group
+Cases included all available patients aged >35 years with a life-time diagnosis of SMI. SMI was defined as schizophrenia, bipolar affective disorder, psychotic depression and other non-tran-sient, non-organic psychoses. We identified Read codes corresponding to these conditions. Code lists were based on previously published lists from the clinicalcodes.org repository[9], supplemented by free-text searches for the aforementioned clinical terms. A snowballing approach was then employed to identify additional terms similarly categorised in the Read code hierarchy. Additional QOF codes were included to capture any change in coding practice post-QOF. The resulting list of codes was reviewed by two clinicians to confirm the appropriateness or otherwise of included codes, as well as identifying excluded diagnoses. As way of validation, the prevalence of the resulting outcomes was checked against existing literature and alternative published code lists. Patients with codes related to prodromal schizophrenia, to a “single episode” or to a “reactive” episode were excluded, unless they also had a valid diagnostic SMI code recorded. Code lists are available from the University of Bristol Research Data Repository.[1Q]
+The comparison group consisted of unmatched, randomly selected patients aged >35 years without SMI who had a period of continuous registration during the study period, aiming for a minimum ratio of controls to cases of 5:1. Patients were excluded if they had ever been prescribed medication used in the treatment of SMI, with the exception of those who had ever had an epilepsy Read code recorded (i.e. when there is a likelihood of medication being used as an anticonvulsant rather than a mood stabiliser).
+We placed a restriction on the lower age limit of our population as outcomes are rare in those under 35 years, and this improved our power to detect differences.
+Clinical outcomes
+We assessed four outcomes related to diagnosis and two related to treatment, which we considered of clinical importance and straightforward to measure using routine health-record data: first ever recording of elevated serum cholesterol >5.0mmol/L, first ever diagnosis of diabetes mellitus, first ever diagnosis/recording of obesity, first ever diagnosis of hypertension, first ever prescription of anti-diabetic medication, and first ever prescription of lipid-modify-ing medication.
+Hypertension and diabetes mellitus outcomes were identified by diagnostic and administrative Read codes, elevated cholesterol by test results, obesity by Read codes or body mass index values >30.0kg/m2, and medications from product code lists identified from the CPRD dictionary. A similar approach to code list development was employed as for the identification of SMI.
+Statistical analyses
+Annual incidence rates were calculated for all six outcomes for the case and comparison groups. The proportion of the financial year spent eligible for inclusion in the analysis by each patient was combined to create a denominator of 'person-years' active for each financial year for both groups.
+For all outcomes an interrupted time series analysis (ITSA) was performed. This approach is recognised to be amongst the strongest quasi-experimental approaches to intervention anal-ysis.[11,12] ITSA utilises data collected over equally spaced time intervals before and after an intervention. It assumes that, in the absence of the intervention, trends prior to the intervention could have been extrapolated to predict future trends.
+A mixed effect segmented logistic regression model was used for the ITSA, with the inclusion of a random intercept to allow for variation between general practices. The ITSA model has been described elsewhere[13| and can be extended to incorporate a control group, as shown in S2 Appendix. Additional terms were added to allow analysis of the second intervention (change to QOF SMI indicators in 2011) and potential confounding by age (categorised using 40, 50 and 60 year cut-offs) and gender. Data manipulation was undertaken using Stata13.0[14] and analyses were conducted in R (version 3.2.4) using the package lme4.[15] Step changes occurring immediately after the intervention's introduction are reflected in the model intercept, with subsequent effects on the temporal trend reflected by the model slope. Model fit was assessed by plotting the predicted probability of the outcome from the fitted model against the observed probability of the outcome for the SMI and non-SMI groups (S3 Appendix).
+We hypothesised that outcome measurements closer together in time may be more similar than outcomes further apart. We used the Cumby-Huizinga test to explore the data for the presence of this issue (autocorrelation), which may have resulted in part due to the grouping of patients by general practice. The autocorrelation was almost zero and appeared to be negligible (the test for lag orders 1 to 5 strongly accepted the null hypothesis of independence in the series, as did the test at the individual lag). On account of the size of the dataset and complexity of the statistical models, we therefore decided to take a parsimonious approach and not account for potential autocorrelation in subsequent analyses.
+Results
+The number of patients and practices included varied with the growth of the CPRD dataset, with 232, 595 and 530 practices contributing data in 1995/96, 2004/05 and 2013/14 respectively. Numbers of patients and demographic characteristics are summarised in Table 1. A total of 67,239 people with SMI and 359,951 people without SMI were included in the analysis, with the former group typically providing fewer days of continuous data (1936 vs. 2703 days) in part due to increased mortality.
+Fig 1 shows the ITSA results. Numerical results are reported in Table 2 (detection of risk factors) and Table 3 (treatment of risk factors). Following the incentivisation of annual reviews for people with SMI in April 2004, there is strong evidence of an immediate increase (i.e. intercept change) in the recording of elevated cholesterol.
+Immediately after the intervention in 2004, the odds of an SMI patient having elevated cholesterol test results are 1.21 (95% CI: 1.10±1.33) times higher, after the intervention in 2004. These odds are 37% (95% CI: 24%-51%) higher than for a non-SMI patient. The results similarly show increased recognition of diabetes (OR 1.21, 95% CI: 0.99±1.49), obesity (OR 1.21, 95% CI: 1.06±1.38) and hypertension (OR 1.19, 95% CI: 1.04±1.38), in the SMI compared to the non-SMI group (Table 2).
+A relative change over time coinciding with the 2004 QOF incentives was seen only for elevated cholesterol (OR 1.03, 95% CI: 1.00±1.05), whereas the remainder saw no relative change in gradient from 2004 to 2010, suggesting the immediate effects observed were sustained.
+The introduction of cardiovascular specific SMI indicators in 2011 was associated with further immediate increases (i.e. changes in intercept) in recognition of cases of elevated
+cholesterol (OR 1.84, 95% CI: 1.72-1.97) and obesity (OR 1.39, 95% CI: 1.26-1.53) relative to the non-SMI group. However, this increase was not sustained over the following two years that the incentives remained in place, with significant reducing trends relative to the non-SMI group. No added improvements in case-finding of diabetes or hypertension were found in association with the 2011 incentives.
+Prescribing of both anti-diabetic and lipid-modifying medications increased significantly over the study period, and to a greater extent in the SMI group. There was no evidence that these changes were related to the QOF indicators introduced in 2004 (Table 3). The SMI indicator introduced in 2011 was also not found to be associated with an increase in prescribing of anti-diabetic medications in the SMI group relative to the non-SMI group.
+Following the 2011 QOF SMI incentives, there was also no strong evidence of an immediate change in prescribing of lipid modifying medications. There was some indication that the change in the effect of time attributable to the 2011 QOF indicator differed for the SMI group compared to the non-SMI group (OR 0.94, 95% CI: 0.89-0.99).
+In the early years of the study there was a second peak in age of diagnosis of SMI in older patients (>60 years) which was absent following the introduction of QOF, raising concerns that a change in coding practice had influenced the findings. As a sensitivity analysis, we removed patients whose age at first diagnosis of SMI was at least 60 years; this had little impact on our findings (S4 Appendix). In addition, we conducted a falsification analysis, using fabricated policy indicator years of 2000 and 2008 (S5 Appendix). There were no significant changes in outcome observed in relation to these dates. The only exception was an apparent change in recording of elevated cholesterol in 2008 but this was in the opposite direction to that observed in the main 2011 analysis. These findings provide reassurance that the observed changes in outcomes in the main analysis are indeed related to the QOF interventions.
+Discussion
+We have found strong evidence that primary care incentives promoting physical health reviews in patients with SMI can result in improvements in the identification but not treatment of cardiovascular risk factors.
+This study benefits from the use of robust methodology and external generalizability. Interrupted time series analysis with a comparison group is a powerful, pseudo-experimental methodology that allows both the temporal trends prior to the intervention and any change coinciding with the intervention to be taken into account. However, the study's limitations also require consideration. First, it is not possible to distinguish improved case finding from
+genuine changes in incidence, although the latter are likely to happen gradually. It is also possible that the increases in diagnoses reflect improved recording as opposed to detection per se. Nevertheless, we believe that recording of clinical information in electronic health records is still likely to facilitate delivery of care, such as through improved case finding, and thus failure to record a risk factor is for all practical purposes equivalent to failing to identify it in the first place. In addition, laboratory values (e.g. cholesterol) are automatically populated from the laboratory systems, and as such reflect genuine measurement and not simply recording. Secondly, we found differences in the age distribution of our SMI population in the earlier years of the study, potentially reflecting changes in coding practice; a fall over time in the age at diagnosis of SMI has also been described elsewhere.[6] Nevertheless, we adjusted for age in our analyses, and also found no notable differences when age was not accounted for by the models, reassuring us that our findings are robust to this issue. Additional supplementary analyses, excluding those with a first diagnosis of SMI at age 60 years or older, supports these findings. Thirdly, ITSA may be rendered invalid by the presence of a confounder that changed in a time period coinciding with the intervention. This is pertinent to the current study as the QOF targeted other broad key health outcomes in the general population, some of which may also have been relevant to people with SMI. This is accounted for by the inclusion of a comparison group in the study design, although the assumption must be made that other health service changes would impact on both groups equally. We are unaware of other major interventions (e.g. clinical guidelines) that coincided temporally with the QOF changes. Although National Service Frameworks were introduced in the early 2000s to set minimum standards for care in areas including diabetes, we expect that the impact of this is likely to be small, as firstly our study focuses on case identification rather than quality of care, and secondly there is no reason to suspect it to have a significant differential impact on the two study groups. Fourthly, the QOF SMI indicators introduced in 2011 relating to lipids and blood glucose/HbA1c testing applied only to those aged 40 and over whilst cases aged 35 and over are included in this study. This could have resulted in a slight underestimate of the true effect. Finally, the study is limited by the small number of data points following the 2011 changes to the QOF SMI indicators. As these indicators were in place for only 3 years, a more complete analysis including longer term trends is impossible, although we believe our analysis probably still provides valid and useful insights into the effect of the later intervention.
+Under-recording of cardiovascular risk factors in the SMI population has been previously reported[16] and it is thus reassuring that incentives appear to help address this. When specific cardiovascular SMI indicators were introduced in 2011, people with SMI had already been subject to annual reviews for seven years, meaning those most willing to attend reviews who were at highest cardiovascular risk may already have been identified. This contextual difference makes it difficult to know whether there are differences in the effectiveness of the 2004 and 2011 indicators. Of note, the sharp increases in recording of both obesity and elevated cholesterol in 2011 despite the seven preceding years of annual reviews perhaps suggest the later cardiovascular-specific indicator resulted in a “catch-up” of previously undetected risk factors; the subsequent drop in recording may reflect patients being tested in earlier years or being otherwise harder to reach. Despite this, the levels of detection do not fall below the pre-2011 trend for either elevated cholesterol or obesity. One must therefore be cautious about both interpreting the initial improvement in 2011 as evidence of the superiority of a more specific incentive, and of interpreting the post-2011 decrease as meaning this indicator was either not sustained or less effective than the indicators introduced in 2004. The majority of indicators were withdrawn in 2014 meaning a more complete analysis of the 2011 indicators will not be possible: it is not clear whether they were in place for sufficiently long to become usual practice.
+The analysis suggests that the QOF SMI indicators have not affected prescribing of either anti-diabetic or lipid-modifying medications. This is despite longstanding (i.e. pre-QOF), readily available UK clinical guidance on pharmacotherapy for the associated risk factors in the general population.[17] National guidance[18] advising clinicians to be aware of and treat increased cardiovascular risk in SMI (specifically schizophrenia) patients has only been available since 2009, but further work is nonetheless required to explore the reasons why the increase in detection of risk factors is not matched by increased treatment. The findings are consistent with studies describing under-treatment in this population in the presence of dysli-pidaemia and hyperglycaemia[19] and in other areas such as stroke[2Q] and arthritis[21]. A number of potential explanations for this have been proposed, including at the patient-level (e.g. cognitive impairment, poor adherence), physician-level (e.g. stigmatization, complexity of care), and service-level (e.g. fragmentation of care, lack of resources).[22] It is unclear whether the increase in medication use due to the QOF prescribing incentives that has been observed in the general population[23] would translate to an increase in prescribing if targeted specifically at the SMI population.
+A number of issues raised by this study merit further investigation. These include whether or not more specific cardiovascular indicators offer additional value, patients' and clinicians' views on the role of SMI indicators, the reason that case detection was not followed by improvements in treatment, and whether there is value in specific incentives to encourage treatment of physical problems in this population. However, what remains clear is that financial incentives for GPs improve the detection of cardiovascular risk factors in a challenging patient group in which identification of physical health problems is known to be poor. Incentives may well have a broader role in reducing health inequalities and improving the care and treatment of patients with severe mental illness.

From ideation to action.txt

@@ -0,0 +1,91 @@
+1.	Introduction
+Suicide is a global health problem, and although suicide affects people across the lifespan, it is the second leading cause of death of 16-29 year olds worldwide (World Health Organisation, 2014), as well as being the leading cause of death among people under 50 in the UK (Snowcroft, 2017). Recent research has identified a wide range of social, psychological and biological factors that act to increase suicide risk (O’ Connor and Nock, 2014), although these factors often do not distinguish between those who will think about suicide and those who will go on to act on suicidal thoughts (Klonsky and May, 2014). With around 60% of transitions from suicidal ideation to a first attempt occurring
+within a year of ideation onset (Nock et al., 2008), it is crucial that we identify factors that distinguish those whose suicidal thoughts may transition into suicidal behaviours (Kessler et al., 2005).
+In light of this, recent models of suicidal behaviour have adopted an ideation-to-action framework, where the development of suicidal ideation and the transition to a suicide attempt are viewed as distinct processes (Klonsky et al., 2017). The first theoretical model to emphasise this distinction was the interpersonal-psychological theory of suicide (IPT; Joiner, 2005), proposing that suicidal desire (comprised of perceived burdensomeness and thwarted belongingness) alone was insufficient to lead to a serious suicide attempt/death by suicide. A suicidal individual must also have the capability to act upon that desire
+characterised by a lowered physical pain sensitivity and high fearlessness about death that overrides the instinct towards self-preservation (Joiner, 2005). Although there has been considerable evidence for the key premises underpinning the IPT (Chu et al., 2017), a recent systematic review of IPT studies found limited evidence for an interaction between perceived burdensomeness, thwarted belongingness and acquired capability in association with suicide attempts, with the authors concluding that the relationships between the variables may be less straightforward than originally presented (Ma et al., 2016). Therefore, models of suicidal behaviour may need to account for a more complex relationship between suicidal ideation and the transition to a suicide attempt.
+In this vein, the integrated motivational-volitional model of suicidal behaviour (IMV; O'Connor, 2011) was proposed in 2011 and refined in 2018 (O’Connor and Kirtley, 2018). The IMV model is a tri-partite framework (Fig. 1) that builds upon previous theories to map the context in which suicide may occur (the pre-motivational phase), the development of suicidal ideation (the motivational phase) and the transition of suicidal thoughts into suicidal behaviours (the volitional phase). Building upon the cry of pain hypothesis (Williams, 1997), the motivational phase focuses on feelings of defeat and entrapment as the key drivers of suicidal ideation. Importantly for the present study, within the final phase of the model (volitional phase), it is argued that a group of factors, labelled volitional moderators, governs the transition from thinking about suicide to attempting/dying by suicide. In addition to Joiner's concept of acquired capability, these factors include im-pulsivity, planning, exposure to the suicidal acts of others, access to means, past suicidal behaviour and mental imagery about death (O’Connor and Kirtley, 2018).
+There has been support for the main facets of the IMV model (e.g., Dhingra et al., 2016; O'Connor et al., 2013; Wetherall et al., 2018), including a growing body of evidence demonstrating that volitional
+moderators do indeed differentiate between those who think about suicide and those who engage in suicidal behaviour (O'Connor et al., 2016; O’ Connor and Kirtley, 2018). For example, in one study of adolescents, only volitional phase variables (self-harm by friends and family, thinking about peers’ self-harm, impulsivity) and stress differentiated between those with thoughts of self-harm and those who engaged in self-harm (O'Connor et al., 2012). Similarly, in a test of the IMV facets with students, within a multivariable model, only the volitional phase factors (exposure to suicide, impulsivity and fearlessness about death) distinguished between those who had experienced suicidal ideation and those who had attempted suicide (Dhingra et al., 2015). Additionally, in a recent cohort study, exposure to the self-harm of others (alongside psychiatric disorder) was key to differentiating between adolescents who had made a suicide attempt compared to those who had thought about but not attempted suicide (Mars et al., 2018).
+A final model utilising the ideati.on-to-acti.on framework is the more recent three-step theory (3ST; Klonsky and May, 2015). The initial steps tap the development and escalation of suicidal ideation with a combination of pain, hopelessness and a lack of connectedness, and in the final step ideation progresses to an attempt when the capability for suicide is present. The concept of acquired capability has been a consistent component across all three models explored, with recent evidence suggesting that when those high on capability become agitated, suicidal intensity increases, thereby facilitating suicidal behaviour by providing sufficient energy and arousal (Ribeiro et al., 2015). Therefore, this concept, along with the additional volitional factors of im-pulsivity, exposure to suicide and mental imagery about death, are key variables to be explored more fully as factors that can differentiate those who think about suicide from those who will make a suicide attempt.
+1.1.	Current study
+This study aimed to investigate a key premise of the IMV model; namely that volitional phase variables govern the transition from suicidal ideation to suicide attempts when motivational phase variables are controlled for (O’Connor & Kirtley, 2018). Although a small number of studies have investigated the psychological factors associated with behavioural enaction (e.g., Dhingra et al., 2015), to our knowledge this is the most detailed study of its kind and the first study to do so in a nationally representative sample. To this end, the Scottish Wellbeing Study (O’ Connor et al., 2018), a nationally representative interviewbased survey of young adults aged 18-34 years across Scotland, was conducted. In short, we hypothesised that (i) motivational and volitional phase factors would differentiate non-suicidal controls from those who had a history of suicidal ideation or suicide attempts, and (ii) only volitional phase factors would differentiate between those who had a history of suicidal ideation and those who had attempted suicide in a multivariable model.
+2.	Method
+2.1.	Sample and procedure
+The data are from the Scottish Wellbeing Study (O’Connor et al., 2018) which is a nationally representative sample of young people aged 18-34 years (n = 3508) from across Scotland. Recruitment was conducted by Ipsos MORI, a social research organisation, between 25th March 2013 and 12th December 2013. A quota sampling methodology was utilised; quotas were based on age (three quota groups), sex and working status (for more details, see O’ Connor et al., 2018). Following written consent, participants completed an hour-long interview, carried out face-to-face in their homes, using Computer Assisted Personal Interviewing (CAPI), with confidential completion of sensitive questions (including suicidal history) on a personal computer. Participants were compensated £25 for their time. Ethical approval was obtained from the University of Stirling (Psychology Department) ethics committee as well as from the US Department of Defense Human Research Protections Office.
+2.2.	Measures
+2.2.1.	Outcome measure: lifetime history of suicidal ideation and attempts
+This was assessed with two items drawn from the Adult Psychiatric Morbidity Survey (APMS; McManus et al., 2007): “Have you ever seriously thought of taking your life, but not actually attempted to do so?” and “Have you ever made an attempt to take your life, by taking an overdose of tablets or in some other way?”. Responses to these questions were “no”, “yes” or “would rather not say”. These items were used to create a 3 category variable indicating if participants had (i) no history of suicidal ideation/ attempt (control group), (ii) had experienced suicidal ideation but had never attempted suicide (suicidal ideation group), or (iii) had reported having attempted suicide in the past (suicidal attempt group).
+2.2.2.	Motivational phase risk factors
+2.2.2.1.	Defeat. The Defeat Scale (Gilbert & Allan, 1998) is a 16-item self-report measure of perceived failed struggle and loss of rank (e.g., “I feel that I have not made it in life”). This scale has good psychometric properties and is significantly correlated with depressive symptoms (Griffiths et al., 2014). In the present study the measure had high internal reliability (Cronbach's a = 0.96).
+2.2.2.2.	Entrapment. The 16-item Entrapment Scale (Gilbert & Allan, 1998) is a measure of the sense of being unable to escape feelings of defeat and rejection (e.g., I am in a situation I feel trapped in). This measure consists of 10 items reflecting external entrapment
+(entrapment by external situations), and 6 items tapping internal entrapment (entrapment by one's own thoughts and feelings). The scale has good psychometric properties (Griffiths et al., 2014) and demonstrated high internal consistency in the present study (Cronbach's a = 0.96).
+2.2.2.3.	Perceived burdensomeness and thwarted belongingness. These were assessed using the 12-item Interpersonal Needs Questionnaire (INQ; Van Orden et al., 2012). The INQ includes 7-items to tap burdensomeness (e.g., “I feel like a burden on the people in my life”) and 5-items to assess belongingness (e.g., “I feel disconnected from other people”). The scales have been shown to have good internal consistency and construct validity (Van Orden et al., 2012), including in this study (perceived burdensomeness Cronbach's a = 0.87, thwarted belongingness Cronbach's a = 0.84).
+2.2.2.4.	Goal disengagement and goal reengagement. The 10-item goal adjustment scale (GAS; Wrosch et al., 2003) consists of a 4-item goal disengagement (e.g., “If I have to stop pursuing an important goal in my life its easy for me to stop thinking about the goal and let it go”) subscale and a 6-item goal reengagement (e.g., “If I have to stop pursuing an important goal in my life I start working on other new goals”) subscale. Both subscales have shown good validity (Wrosch et al., 2003), and in the present study they had adequate to good internal consistency (goal disengagement Cronbach's a = 0.70, goal reengagement Cronbach's a = 0.87).
+2.2.2.5.	Social support. The 7-item ENRICHD Social Support Instrument (ESSI; Mitchell et al., 2003), taps four defining attributes of social support: emotional, instrumental, informational, and appraisal (e.g., “Is there someone available to give you good advice about a problem?”). It has been found to be a valid and reliable measure of social support (Vaglio et al., 2004), and displayed good internal reliability in the present study (Cronbach's a = 0.87).
+2.2.2.6.	Resilience. Resilience was measured using the 10-item Brief Resilience Scale (BRS; Campbell-Sills and Stein, 2007), adapted from the 25-item Connor-Davidson Resilience Scale (CD-RISC; Connor and Davidson, 2003). This 10-item version (e.g., “Coping with stress can strengthen me”) has good psychometric properties and is highly correlated with the original 25-item version (Campbell-Sills and Stein, 2007), and in the present study it displayed excellent internal consistency (Cronbach's a = 0.90).
+2.2.3.	Volitional phase risk factors
+2.2.3.1.	Acquired capability. The Acquired Capability for Suicide Scale (ACSS; Van Orden et al., 2008) is a 5-item measure designed to assess one's fearlessness about death and physical pain sensitivity (e.g., “The pain involved in dying frightens me”). The scale has demonstrated convergent and discriminant validity (Van Orden et al., 2008), and in this study the ACSS had a relatively low internal consistency of 0.63 (Cronbach's a).
+2.2.3.2.	Impulsivity. This was assessed using the 30-item Barratt Impulsiveness Scale Version 11 (BIS-11; Patton et al., 1995); a selfreport questionnaire that accounts for the multi-faceted nature of the construct (i.e., attentional, motor and non-planning impulsiveness) that provides a general impulsiveness score (e.g., “I act on the spur of the moment” ). The BIS is a commonly used scale that has been shown to correlate with behavioural measures of impulsivity (Martins et al., 2004), and it displayed good internal validity in the present study (Cronbach's a = 0.83).
+2.2.3.3.	Mental imagery. Eight questions were asked to establish the frequency with which participants imagine death related imagery when they feel down or distressed, including engaging in self-harm or suicidal
+behaviour (e.g., “...images of yourself planning/preparing to harm yourself or make a suicide attempt”). Greater presence of suicide-related imagery has been linked to higher levels of suicidal ideation (Holmes et al., 2007). The scale displayed good internal reliability (Cronbach's a = 0.84).
+2.2.3.4.	Exposure to suicide. Participants were asked three items to establish whether they had friends or family who attempted or died by suicide (e.g., “Has anyone among your family attempted suicide?”). These items have been used in previous research (O'Connor et al., 2012) and have been shown to differentiate between those who think about suicide and those who attempt suicide (Dhingra et al., 2015).
+2.2.4.	Covariates: demographic characteristics and mood
+2.2.4.1.	Demographic characteristics. We recorded the following demographic information: age, gender, marital status (married vs. not married), ethnicity (white vs. non-white) and economic activity (employed, inactive and unemployed).
+2.2.4.2.	Depressive symptoms. The Beck Depression Inventory-II (BDI-II; Beck et al., 1996) is a well-established measure tapping a range of depressive symptoms (e.g., self-dislike, loss of energy) containing 21 items. It has been shown to yield reliable, internally consistent, and valid scores in many different populations (e.g., Dozois et al., 1998), and in this study, it displayed high internal reliability (Cronbach's a = 0.95).
+2.3.	Statistical analysis
+Data analysis was conducted using SPSS version 22. The missing data included items missed by participants and participants selecting ‘would rather not say’. We used every participant's data as long as they had completed 75% or more of a psychological scale, this resulted in minimal missing data, < 1% on any variable (range 0.31-0.86%; including those who had refused). These small amounts of missing data were checked against demographic characteristics and as there were no significant associations, expectation maximisation (EM) was applied to replace missing items for each scale. The multinomial regression model included only those who completed > 75% of every measure (n = 3330; 95% of total sample), with a small proportion of the data EM replaced. More information on the EM replacement method is included in the supplementary materials.
+Additionally, the data were weighted to ensure that the attained sample based on the quota variables was in line with the population in the sample frame using rim weighting. Overall, as the quotas were almost always met (30-34 year olds, full-time students and full-time workers were slightly under-represented) the effect of the weights was small. All analyses and reporting of data were conducted with the weights on. More information on the rim weighting is included in the supplementary materials.
+To investigate the respective influence of the motivational and volitional phase variables, initial univariate multinomal regression analyses were conducted. To control for the number of comparisons the Holm-Bonferroni correction method (Holm, 1979) was applied. In order to identify which variables independently distinguished between the groups, a multivariable multinomial logistic regression was performed. Specifically, demographic and mood variables were entered as covariates (age, gender, marital status, ethnicity, economic activity and depressive symptoms), followed by the motivational phase variables (defeat, entrapment, perceived burdensomeness, thwarted belongingness, goal disengagement, goal reengagement, social support and resilience) and then the volitional phase variables (acquired capability, impulsivity, mental images, exposure to suicide death (family & friend), exposure to suicide attempt by friend, exposure to suicide attempt by family) were entered. Odds ratios (OR) indicating the likelihood of each variable's association with the higher risk group were reported (i.e., the
+ideation and attempt groups relative to the controls, and the attempt group relative to the ideation group), with those greater than one indicating increased risk and less than one decreased risk. To estimate the variance explained by the volitional variables in distinguishing between the suicide ideation and attempt groups, a binary logistic regression was conducted with only the volitional variables.
+To better understand how well the volitional phase measures distinguish between those who have thought of suicide only and those who have made a suicide attempt at an individual level, the sensitivity (i.e., proportion of the sample high on a volitional phase variable that were correctly identified as having made a suicide attempt) and specificity (i.e., the proportion of the sample that were low on a volitional phase variables and had not made a suicide attempt) of each of the volitional phase variables is reported, along with their positive predictive value (i.e., the probability that the individual high on a volitional phase variable had attempted suicide) and negative predictive value (i.e., the probability that the individual low on a volitional phase variable had not attempted suicide). A cut-off score (mean + 1SD) was created for the continuous variables to indicate those ‘high’ and ‘low’ on a particular volitional phase variable.
+3.	Results
+3.1.	Sample characteristics
+In the primary analysis (n = 3330), the majority of the sample had no suicidal history (n = 2470; 74.6%), 14.3% (n = 481) had experienced suicidal ideation in their lifetime but had never made a suicide attempt, and 11% (n = 379) had attempted suicide in their lifetime. The descriptive statistics by group membership (i.e., ideation vs. attempt vs. control) and univariate differences for those who responded to the suicidal history questions (n = 3435) are provided in Table 1. With demographics, the univariate multinomial regression analyses indicated that those with suicidal ideation were more likely to be male, not married and unemployed compared to controls, and those who had reported a suicide attempt were more likely to be female, older and unemployed than both the controls and those in the suicidal ideation group.
+Members of the control group scored significantly lower on all of the psychological risk factors compared to those in the suicide ideation and suicide attempt groups; this included depressive symptoms, defeat, entrapment, acquired capability and impulsivity. Those in the suicide attempt group reported more frequent exposure to the suicidal behaviour of others, with almost 50% having been exposed to a friend making a suicide attempt, compared to just 16% for the control group. The control group reported higher levels of protective factors such as resilience and social support. A similar pattern emerged between the two suicidal history groups; those in the suicide attempt group more strongly endorsed the motivational and volitional phase risk factors compared to those in the suicide ideation group.
+3.2.	Multivariable multinomial regression analyses
+The results of the multinomial regression analyses are presented in Table 2. The model was statistically significant (%2 (42) = 1528.60, p < 0.001; pseudo R-square (Cox and Snell) = 0.37). Those in the control group were significantly lower than both suicidal history groups on a combination of motivational (defeat and burdensomeness) and volitional phase factors (acquired capability, mental images, exposure to suicide attempt by family or friend). Additionally, those in the suicide attempt group were more likely to be female, older, and higher on impulsivity than controls. Depressive symptoms did not distinguish between any of the groups when all motivational and volitional factors were accounted for.
+Similarly, those who reported a suicide attempt were older (OR = 1.07 [95% CI = 1.03-1.10]) and more likely to be female
+(OR = 0.49 [95% CI = 0.36-0.67]) than those in the ideation group. However, consistent with the IMV model, the only psychological factors that distinguished those in the suicide attempt group from those in the suicidal ideation group were volitional phase variables; none of the mood or motivational phase variables significantly differentiated between these groups. In comparison to those in the suicidal ideation group, those who reported a suicide attempt scored significantly higher on levels of acquired capability (OR = 1.10 [95% CI = 1.06-1.14]), impulsivity (OR = 1.02 [95% CI = 1.01-1.04]), mental images about death (OR = 1.07 [95% CI = 1.03-1.10]) and they were significantly more likely to have been exposed to a suicide attempt of a friend (OR = 1.49 [95% CI = 1.09-2.06]). In a binary logistic regression, the volitional phase factors accounted for 11% of the variance in distinguishing between the suicide ideation vs. the suicide attempt groups (Nagelkerke R Square = 0.112).
+3.3.	Sensitivity and specificity of the volitional phase variables in differentiating between suicide ideation and suicide attempt groups
+The findings of the sensitivity and specificity analyses are displayed in Table 3. Being high on acquired capability, impulsivity and mental images, as well as each of the exposure variables, identified those who had made a suicide attempt over half of the time, with acquired capability being the most sensitive (56.9% correctly identified). The specificity of the individual variables was higher overall (range 57.9-62.6%), indicating that being low on the volitional phase variables was more specific at identifying those who had not made a suicide attempt. All the volitional variables, when taken together, identified around 46% of those who had made a suicide attempt, and three
+quarters of those who had not. The positive predictive values (PPV) ranged from 37.1-54.5%, with mental imagery having the highest PPV. The negative predictive values (NPV; range 61.9-77.4%) were higher; indicating being low on a volitional variable was a better predictor of who had not attempted suicide than being high was a predictor of those who had. The PPV increased when all volitional variables were taken into account, with approximately 60% of those predicted to have made a suicide attempt correct, with almost two-thirds for the NPV.
+4.	Discussion
+We tested a key premise of the integrated motivational-volitional model (IMV, O’ Connor, 2011; O’Connor & Kirtley, 2018), namely that
+volitional phase factors are key to governing the transition from suicidal ideation to a suicide attempt. We hypothesised that (i) motivational and volitional phase factors would differentiate non-suicidal controls from those who had a history of suicidal ideation or suicide attempts, and (ii) only volitional phase factors would differentiate between those who had a history of suicidal ideation and those who had attempted suicide in a multivariable analysis. Findings yielded clear evidence in support of both hypotheses. Specifically, a combination of motivational and volitional phase variables distinguished the control group from both the suicide ideation group and the suicide attempt group. Whereas, apart from some demographic differences (those in the attempt group being older and female), only volitional phase variables differentiated between those with a history of suicidal ideation and those who had reported a suicide attempt; with the latter group reporting higher levels of acquired capability, impulsivity, mental imagery about death and they were more likely to have been exposed to the suicide attempt of a friend.
+This study adds to the growing literature highlighting the importance of the volitional phase factors within the IMV model (e.g., O'Connor et al., 2012; Dhingra et al., 2015) and the ideation-to-action framework more generally (Klonsky et al., 2017). It is also unique as it is the first study of its kind to investigate the role of volitional phase factors in a large, nationally representative sample. Although motivational phase variables, including key components of the IPT (e.g., perceived burdensomeness) and the IMV model (e.g., defeat), are useful to identify who may think of suicide, they are not the key drivers of behavioural enaction. In light of the recent concerns that most risk factors do not distinguish between those suicidal individuals who are/are not at increased risk of making a suicide attempt (Klonsky and May, 2014), the present volitional phase findings are important as they address this dearth in the research literature. Crucially though, they highlight potential targets for interventions and therapies, consistent with a recent call to action to identify better markers of suicide risk (Holmes et al., 2018).
+Our study adds to the recent research on sensitivities and specificities in the context of risk assessments, showing that the latter fail to accurately predict suicidal behaviour over time (Quinlivan et al., 2017; Steeg et al., 2018). In the present study, the sensitivity of the volitional phase variables in differentiating between the suicide ideation vs. suicide attempt groups was relatively low (46% correctly identified), therefore potentially limiting their utility in assessing risk at an individual level. However, given that our study design is investigating lifetime suicidal ideation and attempts, low sensitivities are not unexpected because the measures were assessed retrospectively; in many cases individuals had thought about suicide or attempted suicide many years before taking part in the study (indeed the overwhelming majority of participants had attempted suicide more than 12 months ago). Moreover, as our measures are not diagnostic tests nor were they designed as such (they are theoretically derived constructs), the utility of reporting sensitivities and specificities is at best only informative. Nonetheless, as noted above, the associations identify key parameters that could be targeted in interventions to reduce suicide risk. One could also argue that the volitional phase variables are actually quite powerful as they still identify those who have attempted suicide compared to those who have thought about suicide years later (albeit that the effect sizes are low). Taking the findings in context, therefore, we believe that the volitional phase variables are important treatment targets which routinely should form part of a clinical formulation.
+Consistent with previous findings (e.g., Dhingra et al., 2015; Mars et al., 2018), exposure to suicide in others, in particular to the suicide attempt of a friend, was most strongly associated with belonging to the suicide attempt group. Contrary to our predictions, the other exposure variables of suicide attempt by family member or death by suicide of either a family member or a friend, did not significantly differentiate between those in the suicidal ideation and the suicide attempt groups. It would be useful to explore why these other types of exposure did not
+differentiate between the groups. Interestingly, Mars et al. (2018) found a dose response effect with adolescents, whereby exposure to self-harm in both family and friends was 5 times higher in their suicide attempt group compared to those reporting suicide ideation only. A number of mechanisms have been suggested to explain this relationship; including that exposure to suicidal peers increases risk due to suicide modelling via social learning (Insel & Gould, 2008) and cognitive accessibility (Biddle et al., 2012). Contagion may also be more likely due to assortative relating processes whereby similar individuals are more likely to associate (Joiner, 2003), and there may even be evidence for a genetic basis to imitation (Brent and Melhem, 2008). Although further research is needed to better understand the mechanisms behind this phenomenon, ultimately the present study highlights the importance of exposure to suicide as a key risk factor for a suicide attempt.
+Additionally, recent research suggests that exposure to suicidal or self-harming behaviours may act as painful and provocative life experiences which feed into acquired capability (Klonsky et al., 2017). Although measures of acquired capability were only weakly associated with suicide attempt history in a recent meta-analysis (Chu et al., 2017), the concept of having to override an innate instinct for survival appears important in understanding the transition to a suicide attempt (Klonsky and May, 2015). Specifically, having fearlessness about death and reduced pain sensitivity appear to be important mechanisms in increasing the ability to act upon one's thoughts of suicide (Smith et al., 2010). Indeed, Kirtley et al. (2016) in a systematic review found a pervasive relationship between lower pain sensitivity and self-harm more generally but highlighted the dearth of research in this area (Kirtley et al., 2016). A better understanding of how capability for suicide develops requires urgent attention, in particular whether its effects can be buffered by protective interventions such as safety planning (Stanley and Brown, 2012).
+Impulsivity could also increase acquired capability through more exposure to painful events (Anestis et al., 2014). Although impulsivity is an established risk factor, traditionally thought to facilitate suicidal behaviours by increasing the likelihood of enacting suicidal thoughts (Mann et al., 1999), more recent findings have questioned the nature of this relationship. As in this study, a meta-analysis found the relationship between trait impulsivity and suicidal behaviour was relatively small (Anestis et al., 2014). Arguably, the research fails to differentiate between state and trait impulsivity; as an individual high in trait im-pulsivity may plan a suicide attempt (and vice versa) (Gvion and Apter, 2011). Therefore, impulsivity remains a problematic concept that may be difficult to target in interventions; trait impulsivity may not accurately reflect the individual's suicidal intentions, but from a clinician's perspective it may be useful to be aware of this.
+The finding that mental imagery related to death distinguishes those who have made a suicide attempt from those who have not is important and novel. It is consistent with Holmes et al. (2007) who found that ‘flash forwards’, defined as imagined future acts of suicide or self-harm are associated with suicide risk. They may be important targets for intervention, with evidence showing that a reduction in suicidal imagery is associated with less suicidal thoughts over time (Ng et al., 2016). However, to be effective, the key mechanisms need to be explored further as there is competing evidence. For example, it has been suggested that imagery increases the cognitive availability of powerful images (Florentine and Crane, 2010), potentially leading to more distress (Holmes and Mathews, 2005); however, for some the images may also function as a deterrent for suicidal behaviour (Crane et al., 2012). In contrast, it is also suggested that habituation may occur, whereby the fear of the (suicidal) act is reduced thereby facilitating behavioural enaction (Crane et al., 2012). In short, we need to advance our understanding of how experiencing suicide ‘flash forwards’ increases suicide risk, and then how best to intervene to reduce suicide risk.
+4.1.	Limitations
+Although this study had many strengths, a number of potential limitations should be noted. First, the data were cross-sectional; therefore causality or directionality cannot be inferred. Second, as with much psychological research, the measures here are reliant on self-report, therefore they are subject to memory and reporting biases. Indeed, suicidal ideation in particular may be subject to mis-reporting (Mars et al., 2016), and as the former was assessed using a single item, we were not able to tap the intensity or severity of thoughts. Third, although the sample was representative of young people across Scotland, it may not be generalisable to other populations, in particular to clinical groups who are at increased risk of suicidal behaviour. Finally, and as noted earlier, the effect sizes of the volitional phase variables were relatively small but given the retrospective study design this is perhaps not surprising as many of the suicide attempts occurred several years ago. Therefore, future research should investigate the extent to which such factors predict suicide attempts over time. Furthermore, Prentice and Miller (1992) set out clear guidelines when small effect sizes should be considered as important. This occurs under two conditions; (1) when the intervention is minimal or (2) when the outcome is difficult to influence. Here the outcome (suicidal behaviour) is relatively hard to predict or manipulate and the predictors here are minimal (scores on a scale). This is why within medicine when a minimal intervention (e.g., aspirin) that has a small (r = 0.034, which converts to an OR of 1.13) but significant effect in reducing a difficult to influence outcome (e.g., risk of future cardiovascular events) it has important public health implications (Steering, 1988). Thus while the effect sizes are small this does not necessarily negate their importance.
+Despite these limitations, the current research is unique and represents the most robust test to date of the volitional phase of the integrated motivational-volitional model of suicidal behaviour (O’Connor and Kirtley, 2018). In the multivariable analyses, only volitional phase factors (acquired capability, exposure to a friend's suicide attempt, mental imagery and impulsivity) differentiated those who reported suicide ideation from those who reported a lifetime suicide attempt. It extends our understanding of the factors which aid the transition from suicidal thoughts to attempts and it provides strong support for the ideation-to-action framework (Klonsky et al., 2017). As highlighted, future research would benefit from more prospective studies with high-risk populations, as well as further exploration of how these particular volitional factors emerge, how best to incorporate them into risk assessment protocols and how to optimally target them in interventions.
+K. Wetherall et al
+Journal of Affective Disorders 241 (2018) 475-483
+Nock, M.K., Borges, G., Bromet, E.J., Alonso, J., Angermeyer, M., Beautrais, A., Bruffaerts, R., Chiu, W.T., de Girolamo, G., Gluzman, S., de Graaf, R., Gureje, O., Haro, J.M., Huang, Y., Karam, E., Kessler, R.C., Lepine, J.P., Levinson, D., Medina-Mora, M.E., Ono, Y., Posada-Villa, J., Williams, D., 2008. Cross-national prevalence and risk factors for suicidal ideation, plans and attempts. Br. J. Psychiatry 192, 98.
+O’Connor, R.C., 2011. The integrated motivational-volitional model of suicidal behaviour. Crisis 32, 295-298.
+O'Connor, R.C., Cleare, S., Eschle, S., Wetherall, K., Kirtley, O.J., 2016. The Integrated Motivational-Volitional Model of Suicidal Behaviour. The International Handbook of Suicide Prevention, pp. 220-240.
+O'Connor, R.C., Rasmussen, S., Hawton, K., 2012. Distinguishing adolescents who think about self-harm from those who engage in self-harm. Br. J. Psychiatry 200, 330-335.
+O'Connor, R.C., Smyth, R., Ferguson, E., Ryan, C., Williams, J.M.G., 2013. Psychological processes and repeat suicidal behavior: a four-year prospective study. J. Consult. Clin. Psychol. 81, 1137-1143.
+O'Connor, R.C., Kirtley, O.J., 2018. The integrated motivational-volitional model of suicidal behaviour. Philos. Trans. R. Soc. B.
+O'Connor, R.C., Nock, M.K., 2014. The psychology of suicidal behaviour. Lancet Psychiatry 1, 73-85.
+O'Connor, R.C., Wetherall, K., Cleare, S., Eschle, S., Drummond, J., Ferguson, E., O'Connor, D.B., O'Carroll, R., 2018. Suicide attempts and non-suicidal self-harm: a national prevalence study of young adults. B. J. Psych. Open 4, 142-148.
+Patton, J.H., Stanford, M.S., Barratt, E.S., 1995. Factor structure of the Barratt impulsiveness scale. J. Clin. Psychol. 51, 768-774.
+Prentice, D., Miller, D.T., 1992. When Small Effects Are Impressive. American Psychological Association, US.
+Quinlivan, L., Cooper, J., Meehan, D., Longson, D., Potokar, J., Hulme, T., Marsden, J., Brand, F., Lange, K., Riseborough, E., Page, L., Metcalfe, C., Davies, L., Connor, R., Hawton, K., Gunnell, D., Kapur, N., 2017. Predictive accuracy of risk scales following self-harm: multicentre, prospective cohort study. Br. J. Psychiatry 210, 429.
+Ribeiro, J.D., Bender, T.W., Buchman, J.M., Nock, M.K., Rudd, M.D., Bryan, C.J., Lim, I.C., Baker, M.T., Knight, C., Gutierrez, P.M., Joiner Jr., T.E., 2015. An investigation of the interactive effects of the capability for suicide and acute agitation on suicidality in a military sample. Depress. Anxiety 32, 25-31.
+Smith, P.N., Cukrowicz, K.C., Poindexter, E.K., Hobson, V., Cohen, L.M., 2010. The acquired capability for suicide: a comparison of suicide attempters, suicide ideators, and non-suicidal controls. Depress. Anxiety 27, 871-877.
+Snowcroft, E., 2017. Samaritans: suicide statistics report 2017 (Including data for 20132015). Samaritans.
+Stanley, B., Brown, G.K., 2012. Safety planning intervention: a brief intervention to mitigate suicide risk. Cogn. Behav. Pract. 19, 256-264.
+Steeg, S., Quinlivan, L., Nowland, R., Carroll, R., Casey, D., Clements, C., Cooper, J., Davies, L., Knipe, D., Ness, J., O'Connor, R.C., Hawton, K., Gunnell, D., Kapur, N., 2018. Accuracy of risk scales for predicting repeat self-harm and suicide: a multicentre, population-level cohort study using routine clinical data. BMC Psychiatry 18, 113.
+Steering, C., 1988. Findings from the aspirin component of the ongoing physicians' health study. N. Engl. J. Med. 318, 262-264.
+Vaglio, J., Conard, M., Poston, W.S., O'Keefe, J., Haddock, C.K., House, J., Spertus, J.A., 2004. Testing the performance of the ENRICHD social support instrument in cardiac patients. Health Qual. Life Outcomes 2, 24.
+Van Orden, K.A., Cukrowicz, K.C., Witte, T.K., Joiner, T.E., 2012. Thwarted belongingness and perceived burdensomeness: construct validity and psychometric properties of the interpersonal needs questionnaire. Psychol. Assess. 24, 197-215.
+Van Orden, K.A., Witte, T.K., Gordon, K.H., Bender, T.W., Joiner Jr, T.E., 2008. Suicidal desire and the capability for suicide: tests of the interpersonal-psychological theory of suicidal behavior among adults. J. Consult. Clin. Psychol. 76, 72-83.
+Wetherall, K., Robb, K.A., O'Connor, R.C., 2018. An examination of social comparison and suicide ideation through the lens of the integrated motivational-volitional model of suicidal behavior. Suicide Life Threat Behav.
+Williams, J.M.G., 1997. Cry of Pain: Understanding Suicide and Self Harm. Penguin, Harmondsworth, England.
+World Health Organisation, 2014. Preventing suicide: a global imperative,. http://www. who.int/mental_health/suicide-prevention/world_report_2014/en/.
+Wrosch, C., Scheier, M.F., Miller, G.E., Schulz, R., Carver, C.S., 2003. Adaptive self-regulation of unattainable goals: goal disengagement, goal reengagement, and subjective well-being. Pers. Soc. Psychol. Bull. 29, 1494-1508.
+483

Gaps-and-challenges-WHO-treatment-recommendations-for-tobacco-cessation-and-management-of-substance-use-disorders-in-people-with-severe-mental-illnessBMC-Psychiatry.txt

@@ -0,0 +1,160 @@
+Background
+The severe mental disorders (SMD), defined as schizophrenia-spectrum, psychoses and bipolar disorders as well as moderate to severe depression, are associated with markedly reduced life expectancy [1]. Worldwide, reductions in life expectancy amongst people with SMD are stark, ranging from 11 to 17 years in the UK [2], 15-20 years across Nordic countries [3], and up to 30 years reduced in low- and middle-income country (LMIC) settings such as in Ethiopia [4]. In particular, this decrement in life expectancy has been noted to be increasing over time [5].
+Although deaths from suicide and other unnatural causes may be more likely in this group compared to general populations, the majority of deaths are in fact due to preventable physical causes, such as cardiovascular disease, respiratory disorders, cancers and infectious disease [6]. In addition, lowered life expectancy may also be because comorbid substance use disorders (harmful substance use and dependence) are the most prevalent psychiatric conditions associated with SMD. Lifetime alcohol use disorders may affect up to 20% of people with schizophrenia [7] and between 24 to 35% of people with bipolar disorders [8, 9]. Comorbid substance use disorders such as cannabis use disorder [10], opioid and other drug use disorder are also known to be more prevalent in these populations compared with the general population [9]. Tobacco use has also been noted to be elevated more than five-fold in people with schizophrenia compared to reference populations [11, 12] and is a leading preventable cause of death in this group of people. Global successes in reducing tobacco use in the general population have not been mirrored by similar reductions in populations with SMD [11, 13].
+A history of substance abuse in populations with SMD has been shown to be associated with an increased risk of death from all-causes and from unnatural causes [14-16]. In addition, findings from a recent study indicated that in general, the presence of substance use disorders (across a broad spectrum of substance types) in SMD was associated with an increased risk of psychiatric admissions, psychiatric emergency department presentations and longer in-patient stays [17]. People with SMDs probably do not just use one substance in particular but are more likely to engage in
+polysubstance use [17]. Factors which make people with dual diagnoses (comorbid mental and substance use disorders) particularly vulnerable to poor health and social outcomes, include the mutually detrimental effect on the course of illness, its identification, diagnosis and treatment; double stigma and barriers to both mental and physical health care, as well as the contribution of substance use to negative health and social outcomes. For tobacco use, the prevalence of tobacco use in people with SMD is higher, and people with SMD are known to start smoking earlier and smoke more heavily [18] compared with the general population [19]. Potential aetiological pathways for premature mortality in SMD populations with these comorbidities are complex and interlinked. Some basic pathways are summarised in Table 1.
+To improve the management of comorbid conditions in adults with SMD and support the reduction of individual health behaviours constituting risk factors for these illnesses, with the aim of decreasing morbidity and premature mortality amongst people with SMD, in 2018 the World Health Organization (WHO) launched guidelines for the “Management of physical health conditions in adults with severe mental disorders” [20]. Prior to the launch of these guidelines it was recognised that whereas there are WHO guidelines addressing mental and substance use disorders as well as physical health conditions in general populations, there was an absence of guidelines specifically targeting those with SMD having comorbid conditions. The target audience for the guidelines are health care practitioners across all specialisms and levels of health care system, as well as policy makers, healthcare planners/providers, programme managers, and people living with SMD as well as their families and carers, and organisations representing the interests of people living with SMD.
+In this paper, we present the findings of a detailed comprehensive overview of existing systematic reviews on the topic areas of tobacco cessation and management of comorbid substance use disorder in SMD, which eventually led to the recommendations in the WHO guidelines on management of physical health conditions in adults with severe mental health disorders. The full guidelines and supporting materials can be accessed
+from the WHO website (https://www.who.int/mental_ health/evidence/guidelines_physical_health_and_severe_ mental_disorders/en/).
+Methods
+The methodologies used to inform the WHO recommendations for the management of tobacco and substance use disorders among people with SMD followed the GRADE (Grading of Recommendations Assessment, Development and Evaluation) process [21].
+A key outcome of the initial phase in developing the guidelines was in the identification of target areas which eventually informed the a priori research questions which followed the PICO [Population, Intervention, Comparison group, Outcomes] format. The research questions guided which physical health conditions and risk factors were to be addressed in the final disseminated guidelines [20]. This
+Table 3 Research questions- substance (drug and/ or alcohol) use disorders
+For people with SMD and substance (drug and/or alcohol) use disorder, are pharmacological and/or non-pharmacological interventions for substance use disorder effective to support reduction in substance use-related outcomes?
+Population/ Intervention / Comparison / Outcome (PICO)
+Population: people with SMD and substance (drug and/or alcohol) use disorder
+Intervention:
+pharmacologicaland/or non-pharmacological interventions for substance use disorders:
+-	Pharmacologicalinterventions
+-	Non-pharmacological interventions: e.g. motivationalinterviewing and/or CBT, psychoeducation, brief assessment interview, dual-focus interventions
+Comparison: care as usual / placebo or one treatment vs another Outcomes:
+Critical
+-	Levelof consumption
+-	Frequency of use
+-	Abstinence
+-	Relapse rates
+Important:
+-	Frequency of adverse events / side-effects
+process was informed by scoping reviews and consultation with a Guideline Development Group (GDG) of externally appointed international experts, engaged by the WHO. Selected PICO questions reflected areas of uncertainty which the GDG felt should be prioritised to inform final recommendations. The final research questions for informing systematic evidence searches were then ratified by the WHO Guideline Review Committee (GRC), which led to the formulation of specific research questions relevant to tobacco and substance use disorders among people with SMD (Tables 2 and 3).
+Figure 1 highlights the comprehensive processes which were followed, leading to the identification of relevant systematic reviews to inform the research questions relating to tobacco cessation, and treatment of substance use disorders in SMD. The retrieval, appraisal and synthesis of evidence closely followed the WHO handbook for guideline development [22]. Databases searched included: the
+Cochrane Library (including DARE), PubMed/Medline, Embase, Psychinfo, Epistemonikos and the Global Health Library. In addition, where searches had to be expanded (see step 3 in Fig. 1) the National Guideline Clearing House was also searched. Search terms employed for the research questions are displayed in supplementary material, and reflected the majority of substances listed in chapters F10-F19 of the tenth revision of the International Classification of Diseases and Related Health Problems (ICD-10) [23]. (Supplementary material: Table 1); these were informed through consultation with guideline methodologists and subject-specific experts at the WHO. Supplementary searches highlighting relevant drug-drug interactions were also employed (Supplementary material: Table 2). Searches between medicines used for tobacco cessation or treatment of substance use disorders and those used for SMDs were carried out using the drugdrug interaction software Lexi-Interact [24]. Lexi-Interact was selected for its clinical utility and the fact that it scored well on both accuracy and comprehensiveness in a review comparing drug-drug interaction software databases [25]. Searches were performed to February 2018 for the tobacco PICO question and to June 2018 for the substance use disorders PICO question.
+Systematic reviews selected for inclusion into GRADE tables conformed to the following inclusion criteria: (1) Timelines- Published within the last 5 years, preferably within the last 3 years; (2) Quality- Papers included for GRADE assessment had sufficiently high methodological quality ratings on the ‘Assessment of Multiple Systematic Reviews’ tool (AMSTAR) [26-28] (see below for further details); (3) Relevance- Retrieved papers were closely relevant to the PICO population. However, where relevant evidence could not be identified these criteria were relaxed, leading to ‘indirect evidence’ to inform recommendations (Fig. 1, step 3). Cochrane reviews or comprehensive meta-analyses and systematic reviews were given preference, wherever possible in this process.
+In order to inform the development of evidence based guidelines in a transparent manner, the GRADE approach was used [21]. An advantage of GRADE is that the certainty of the evidence can be summarised and assessment of the evidence can be separate to the strength of the recommendations which inform the final guidelines [21].
+Prior to selection for GRADE assessment, retrieved articles had to meet sufficiently high quality ratings on the AMSTAR tool [26-28]. The AMSTAR tool leads to a score across 11 domains according to which the quality of each retrieved systematic review is rated. Papers were initially assessed by a member of the team and then cross-checked by another member of the team (MS, JD, PCG). Systematic reviews fulfilling inclusion criteria with a sufficiently high AMSTAR quality rating (a positive rating on more than 6 out of 11 domains) were then assessed using the GRADE approach using
+the GRADEpro tool by a member of the team (MS), with all GRADE assessed papers subsequently rated by a second rater (JD and CB). Discordant ratings between team members on the AMSTAR and the GRADE were resolved through discussion in the team. Key attributes of studies relating to each of the PICO questions were extracted from each included study using a structured form by one member of the team and cross-checked by another. WHO guidelines for rating studies in terms of certainty of evidence, according to the GRADE were followed, to assess each study for limitations, inconsistency, indirectness, imprecision and the reporting of bias, leading to a final GRADE assessment of the certainty/ confidence of the findings reported in the review [29]. For each included study a relevant summary measure was extracted, which was either a Relative Risk (RR) or Mean Difference (MD).
+GRADE evidence profiles for each of the PICOs were presented and discussed over a series of roundtable meetings convened at the WHO in Geneva in May 2018. GDG members were selected internationally across UN member states for their expertise within the topic areas. In addition, the meetings were also attended by a guideline methodologist, the evidence review team and the WHO secretariat. The final recommendations resulted from a consideration of the background evidence for each of the PICO questions, summarised as GRADE profiles and the certainty of evidence for these, as well as taking into consideration other aspects such as whether the problem was considered a priority, how substantial desirable and undesirable anticipated effects were, whether the balance between desirable/undesirable effects favoured the intervention over the comparator, the value attached to the outcomes and the certainty of evidence relating to likely resource requirements, cost effectiveness, impact on health equity, acceptability and feasibility of the intervention. In addition the acceptability of the intervention to healthcare providers in LMICs, feasibility of the intervention and the impact of the intervention on equity and human rights were considered.
+Results
+After consultation with the GDG and WHO GRC agreed research questions specific to tobacco cessation and substance use disorders were:
+1. For people with SMD who use tobacco, are pharmacological (including nicotine replacement therapy, bupropion, varenicline) and/or non-pharmacological interventions effective to support tobacco cessation?
+2. For people with SMD and substance (drug and/or alcohol) use disorder, are pharmacological and/or
+non-pharmacological interventions for substance use disorder effective to support reduction in substance use-related outcomes?
+In total 1434 records were initially identified through the systematic searches for SMD and tobacco cessation; after screening for eligibility and removal of duplicates, 4 reviews were included in the GRADE tables for this PICO with 18 reviews in total contributing evidence through narrative synthesis. For SMD and substance use disorders, a total of 4268 records were identified. After screening and checking against eligibility criteria, 4 studies were included in the GRADE tables on this topic with a total of 16 studies included in the narrative synthesis. Figures 2 and 3 display PRISMA flow charts of relevant articles retrieved for SMD and tobacco use and with substance use disorders, respectively.
+For tobacco use in SMD, GRADE evidence profiles were compiled for: the use of Buproprion, Varenicline and Nicotine Replacement Therapies (NRT) (all versus placebo). In
+addition, GRADE profiles for non-pharmacological interventions (which included: motivational enhancement, psy-choeducational approaches, Cognitive Behavioural Therapy (CBT)), supplementing NRT were compared to standard care approaches, and the use of contingent reinforcement (using money/money plus NRT) compared to care-as-usual was assessed with GRADE [30-33] (For full recommendations with supporting evidence, including relevant drugdrug interactions for Buproprion, Varenicline and NRT see: https://www.who.int/mental_health/evidence/guidelines_ph ysical_health_and_severe_mental_disorders/en/). The GDG recommended combination pharmacological with behavioural interventions, as behavioural interventions alone have been shown to result in a relatively low abstinence rate for tobacco use in SMD.
+The certainty of evidence derived from GRADE, relating to specialised smoking cessation interventions versus standard approaches in people with SMD, was very low. There was insufficient evidence to suggest the superiority of specialised smoking interventions over standard
+smoking cessation approaches for SMD populations. In addition, the certainty of evidence relating to contingency reinforcement approaches compared with care-as-usual for tobacco cessation in SMD populations was very low.
+Pharmacological interventions identified for tobacco cessation in SMD populations were: NRT, Bupropion and Varenicline. Evidence for the efficacy of these interventions in SMD populations mostly derived from high income settings with a few exceptions (e.g. studies for Bupropion which had been conducted in China and Iran as well as in the USA). These pharmacological interventions for tobacco cessation are already recommended by the WHO in general populations, although only NRT is on the WHO essential medicines list [23]. Searches of pharmacological interactions indicated the possibility of interactions between Bupropion and psychotropic medications commonly prescribed in SMD, particularly related to lowering seizure threshold and enzyme inhibition or induction (see https://www.who.int/
+mental_health/evidence/guidelines_physical_health_and_se vere_mental_disorders/en/for full list of interactions).
+For substance use disorders and severe mental disorders, assessment of evidence using the GRADE approach included a review of evidence relating to psychological interventions such as CBT plus motivation interviewing (MI) versus care-as-usual, CBT versus care-as-usual, MI versus care-as-usual and contingency management versus care-as-usual for people with SMD and substance use disorders [34]. Brief interventions, specifically delivered in four or fewer sessions [35], were also assessed. Although these types of interventions may have a basis simply in providing education and advice [35], the brief interventions which were identified and assessed according to GRADE for these guidelines all compared motivational interviewing with CBT approaches, delivered over shorter time frames [35]. In addition, evidence relating to the efficacy of antipsychotic medications in reducing psychotic symptoms alongside other outcomes such as frequency of
+substance use, in dual diagnoses populations were also assessed [36, 37] as well the prescribing of antidepressants in depression comorbid with alcohol use disorders to improve outcomes [38].
+All of the main recommendations relating to each of the PICO questions are presented in Table 4. For dual diagnoses populations, there was a lack of evidence to support the
+superiority of any of the psychological interventions in improving outcomes related to SMD comorbid with substance use disorders. Furthermore, the review team were unable to identify any studies which had specifically assessed these populations within LMIC settings, further limiting generalisabilty. Of those studies retrieved, most were of very low certainty. The GDG reflected that the
+relative lack of evidence to support the efficacy of these interventions in people with SMD comorbid with substance use disorders may partly be due to these populations being more likely to be excluded from research [39].
+In general, the assessment of evidence using GRADE methods indicated low to very low certainty evidence from randomised controlled trials of pharmacological interventions for the management of mental disorders (whether through the use of antipsychotics or antidepressants), which did not indicate the superiority of any of the surveyed medications, when prescribed for people with SMD comorbid with substance use disorders [36-38]. Moderate side effects were noted for these interventions, which need to be taken into account when prescribing for this patient population. In addition, it was noted that medicines which may be used for the management of opioid use disorders such as Methadone and Buprenorphine have interactions with many of the commonly used psychotropic medications, including cardiac effects such as QTc prolongation, central nervous system depression and serotonergic effects (see Annex 6 of guidelines for details: https://apps.who.int/ iris/bitstream/handle/10665/275718/9789241550383-eng. pdf?ua=1).
+For both comorbid tobacco use and substance use disorders, where retrieved evidence was of very low certainty, the expertise of the international GDG was sought, who applied their expertise to the topic area. As a result of the low/very low certainty of evidence retrieved, resultant recommendations were conditional. A ‘conditional’ recommendation by the GDG indicates that GDG members concluded that beneficial effects of the intervention probably outweighed undesirable effects but with insufficient evidence for the GDG to support a ‘strong’ recommendation (with ‘strong’ recommendations indicating that the GDG felt confident that beneficial effects outweighed undesirable effects for the recommended intervention). For people with SMD and substance (drug and/or alcohol) use disorder, the low certainty of evidence led to the recommendation that the mhGAP guidelines for the management of substance use disorders should be followed (Table 4).
+The full GRADE evidence profiles are displayed in the supplementary materials (supplementary tables 1-2) and can also be accessed online. PRISMA checklist has also been provided in supplementary materials (see additional material: PRISMA checklist).
+Discussion
+These evidence-based recommendations, based on detailed and comprehensive reviews of systematic reviews, as well as consultation with an international body of experts and WHO specialists, represent a positive and important step towards tackling the 15-20 year reduction in life
+expectancy, experienced by people with SMD compared to the general population, globally. These guidelines highlight the need to adequately manage tobacco and other substance use disorders in people with SMDs, alongside optimally managing the mental disorder.
+Evidence synthesis highlighted a general lack of high-quality evidence detailing effective interventions for tobacco cessation in SMD and/or for dual diagnoses populations. This reflects a systematic exclusion of people with SMD and/or dual diagnoses from clinical trials, despite evidence indicating that mental disorders are highly comorbid with substance use. There is a need to consider and include these populations in future research [39].
+Do the guidelines go far enough? The guidelines retain a practical emphasis to inform clinicians, healthcare providers and other professional groups on best-practice recommendations and acknowledge the importance of wider multi-level interventional frameworks to address the inequalities impacting on SMD populations [40]. Within this framework, a consideration of health system factors as well as broader social determinants which include social support, stigma and attempts to reduce social exclusion play a major role [40]. In addition, although not directly addressed by the guidelines, public health actions to prevention implemented at country-level form the backdrop to recommended interventions at a whole populationlevel [41], irrespective of group-specific evidence; for example recommended interventions for tobacco cessation or harmful alcohol use could be read within the context of country-level increased taxation/pricing policies on tobacco or alcohol, restrictions on the availability of alcohol, measures to restrict drink-driving, restricted tobacco or alcohol advertising as well as population-level educational campaigns on tobacco cessation, and access to screening and brief interventions [42, 43] or other cost effective interventions [44]. In addition, the guidelines should be read in conjunction with public health/systemic interventions at country-level to address and support population-level mental health [45].
+Our searches revealed a scarcity of evidence particularly relating to dual diagnoses populations, which impacted on the ability to make strong recommendations relevant to people with SMD and comorbid substance use. The scarcity of good quality evidence to inform the recommendations reflects the experience of authors of a previous systematic review, whereby it was found that more than half of clinical randomised controlled trials on the pharmacological treatment of opioid dependence excluded people with psychiatric disorders [39]. The systematic exclusion of people with mental disorders from randomised controlled trials has also been noted in one other review in which the authors assessed the presence of psychiatric exclusion criteria in randomised controlled trials [46]. The exclusion of people with mental
+disorders from trials may in part be due to a number of factors, including trialists’ concerns that decisional capacity to take part is more likely to be impaired in people with SMDs, or concerns that the stress or unintended consequence of taking part in a trial may lead to an exacerbation of mental disorder [46]. In addition, pharmaceutical companies may stipulate extensive exclusion criteria to ensure a smoother pathway to regulation and approval for pharmaceutical products [46]. However, these practices lead to “scientific neglect” [46], and as we have highlighted in this paper, serve to perpetuate the inequalities which people with SMDs experience further. For those systematic reviews which were retrieved, there was also an absence of high-quality evidence relating to psychological interventions to address substance use disorders in dual diagnosis populations. This presents a major limitation, as there is a high co-morbidity of psychiatric and substance use disorders in clinical practice, and for practical purposes it is difficult to address one without the other. In future, research which actively includes people with SMD and comorbid substance use are needed particularly to avoid perpetuating further social exclusion and marginalisation.
+Most of the evidence which informed the development of the guidelines came from well-resourced settings. This may mean that specific issues relevant to low resource settings may impact on implementation. Issues relating to cost and capacity will need to be taken into account for some recommended interventions. The availability of certain medications- such as Varenicline (which does not currently appear in the WHO essential medicines list) may be restricted in certain contexts, although other interventions (such as NRT) are more widely available. Other factors relating to acceptability of the guidelines and longer term sustainability across countries will need to be monitored. Future guidelines may reflect feedback from people on the ground at the forefront of implementing these guidelines on tobacco use and substance use disorders in SMDs- for example following feedback from health care practitioners, policy makers and public health practitioners.
+Conclusions
+Tobacco use and substance use disorders play an important role in heightening the risk of premature mortality in people with SMDs. Our search of the evidence highlighted gaps in the evidence base, which may in part be due to the systematic exclusion of people with SMDs from clinical trials. Despite the challenges described in this paper, these guidelines may mark an important step towards addressing premature mortality in people with SMD. The recommendations may help to inform policy and decision makers globally and in LMIC settings in ensuring more equitable access to tobacco cessation and substance
+use disorder services for these populations. However the dearth of high-quality evidence and evidence from LMIC settings must inform the future research agenda.
+Das-Munshi et al. BMC Psychiatry (2020) 20:237
+Page 12 of 13
+Columbia in Canada; John Saunders, The University of Sydney, Australia; Najma Siddiqi, University of York, UK; Isolde Sommers, Danube University Krems, Austria; Charlene Sunkel, Central Gauteng Mental Health Society, South Africa; Hedinn Unnsteinsson, Prime Minister's Office, Iceland; Pieter Ventevogel, UNHCR, Switzerland Lakshmi Vijaykumar, Voluntary Health Services, Chennai, India; Inka Weissbecker, International Medical Corps, Washington DC, USA.
+Members of the WHO STEERING GROUP: Tarun Dua, Programme Manager, Department of Mental Health and Substance Abuse; Neerja Chowdhary, Technical Officer, Department of Mental Health and Substance Abuse. WHO headquarters members: Bernadette Cappello, Department of Essential Medicines and Health Products; Meg Doherty, Department of HIV/AIDS; Alexandra Fleischmann, Department of Mental Health and Substance Abuse; Dongbo Fu, Department of Prevention of Noncommunicable Diseases;
+Ernesto Jaramillo, Global TB Programme; Dzmitry Krupchanka; Department of Mental Health and Substance Abuse; Shanthi Pal, Department of Essential Medicines and Health Products; Vladimir Poznyak, Department of Mental Health and Substance Abuse; Shekhar Saxena, Department of Mental Health and Substance Abuse; Mark van Ommeren, Department of Mental Health and Substance Abuse; Cherian Varghese, Department of NCDs, Disability, Violence and Injury Prevention. We also acknowledge the contribution of other colleagues: Marco Antonio De Avila Vitoria, Department of HIV/AIDS and Chantal Mignone, Department of HIV/AIDS. WHO regional office advisors: Nazneen Anwar, WHO Regional Office for South-East Asia; Dan Chisholm, WHO Regional Office for Europe; Devora Kestel, WHO Regional Office for the Americas; Sebastiana Da Gama Nkomo, WHO Regional Office for Africa; Khalid Saeed, WHO Regional Office for the Eastern Mediterranean; Martin Vandendyck, WHO Regional Office for the Western Pacific.
+Parental consent for participation
+Not applicable.
+Disclaimer
+The authors alone are responsible for the views expressed in this article and they do not necessarily represent the views, decisions or policies of the institutions with which they are affiliated. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
+Authors’ contributions
+JD wrote the first draft of the manuscript, using materials prepared by MS. All authors (JD, MS, CB, NC, PG, KK, DK, TD, GT) contributed to the writing of the manuscript. All authors (JD, MS, CB, NC, PG, KK, DK, TD, GT) participated in the consensus meeting or reviewed the evidence and its interpretation for the development of the final recommendations (or a combination of these) and contributed to the interpretation. All authors (JD, MS, CB, NC, PG, KK, DK, TD, GT) agreed with the final version of the paper. All authors (JD, MS, CB, NC, PG, KK, DK, TD, GT) read and approved the final version of the manuscript.
+Funding
+JD is funded by the Health Foundation working together with the Academy of Medical Sciences and by the ESRC in relation to the SEP-MD study (ES/ S002715/1) and part supported by the ESRC Centre for Society and Mental Health at King's College London (ESRC Reference: ES/S012567/1). MS is supported by the NIHR Global Health Research Unit for Neglected Tropical Diseases at BSMS. PCG is supported by the UK Medical Research Council in relation the Indigo Partnership (MR/R023697/1) award. GT is supported by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care South London at King's College London NHS Foundation Trust, and the NIHR Asset Global Health Unit award. GT receives support from the National Institute of Mental Health of the National Institutes of Health under award number R01MH100470 (Cobalt study). GT is supported by the UK Medical Research Council in relation the Emilia (MR/ S001255/1) and Indigo Partnership (MR/R023697/1) awards.
+The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, the MRC, the Department of Health, the ESRC or King's College London.
+Availability of data and materials
+All supporting documents which informed the development of this manuscript and the guidelines are freely available through the web links provided in this manuscript or through contacting the authors.
+Ethics approval and consent to participate
+Ethical approvals and consent to participate were not required for this study/ not applicable.
+Consent for publication
+Not applicable.
+Competing interests
+The authors have no competing interests to declare.
+Author details
+department of Psychological Medicine, Institute of Psychiatry Psychology & Neurosciences, King's College London, South London & Maudsley NHS-Trust, De Crespigny Park, London SE5 8AF, UK. 2Centre for Global Health Research, Brighton and Sussex Medical School, Brighton, UK. 3WHO Collaborating Centre for Research and Training in Mental Health and Service Evaluation, Department of Neuroscience, Biomedicine and Movement Sciences, Section of Psychiatry, University of Verona, Verona, Italy. department of Mental Health and Substance Abuse, World Health Organization, Geneva, Switzerland. 5Centre for Global Mental Health, Institute of Psychiatry, Psychology & Neuroscience, King's College London, London, UK. 6The Chester M. Pierce, MD Division of Global Psychiatry, Massachusetts General Hospital, Boston, USA.
+Received: 28 January 2020 Accepted: 26 April 2020 Published online: 14 May 2020
+References
+1.	Hjorthoj C, et al. Years of potential life lost and life expectancy in schizophrenia: a systematic review and meta-analysis. Lancet Psychiatry. 2017;4(4):295-301.
+2.	Chang C-K, et al. Life expectancy at birth for people with serious mental illness and other major disorders from a secondary mental health care case register in London. PLoS One. 2011;6(5):e19590.
+3.	Nordentoft M, et al. Excess mortality, causes of death and life expectancy in 270,770 patients with recent onset of mental disorders in Denmark, Finland and Sweden. PloS one. 2013;8(1):e55176.
+4.	Fekadu A, et al. Excess mortality in severe mental illness: 10-year population-based cohort study in rural Ethiopia. Br J Psychiatry. 2018;206(4): 289-96.
+5.	Wahlbeck K, et al. Outcomes of Nordic mental health systems: life expectancy of patients with mental disorders. Br J Psychiatry. 2018;199(6): 453-8.
+6.	Lawrence D, Hancock KJ, Kisely S. The gap in life expectancy from preventable physical illness in psychiatric patients in Western Australia: retrospective analysis of population based registers. BMJ. 2013;346:f2539.
+7.	Koskinen J, et al. Prevalence of alcohol use disorders in schizophrenia - a systematic review and meta-analysis. Acta Psychiatr Scand. 2009;120(2):85-96.
+8.	Di Florio A, Craddock N, van den Bree M. Alcohol misuse in bipolar disorder. A systematic review and meta-analysis of comorbidity rates. Eur Psychiatry. 2014;29(3):117-24.
+9.	Hunt GE, et al. Comorbidity of bipolar and substance use disorders in national surveys of general populations, 1990-2015: systematic review and meta-analysis. J Affect Disord. 2016;206:321-30.
+10.	Koskinen J, et al. Rate of cannabis use disorders in clinical samples of patients with schizophrenia: a meta-analysis. Schizophr Bull. 2010;36(6): 1115-30.
+11.	de Leon J, Diaz FJ. A meta-analysis of worldwide studies demonstrates an association between schizophrenia and tobacco smoking behaviors. Schizophr Res. 2005;76(2):135-57.
+12.	Toftdahl NG, Nordentoft M, HjorthojC. Prevalence of substance use disorders in psychiatric patients: a nationwide Danish population-based study. Soc Psychiatry Psychiatr Epidemiol. 2016;51(1):129-40.
+13.	Hartz SM, et al. Comorbidity of severe psychotic disorders with measures of substance use. JAMA Psychiatry. 2014;71(3):248-54.
+Das-Munshi et al. BMC Psychiatry (2020) 20:237
+Page 13 of 13
+14.	Das-Munshi J, et al. Ethnicity and excess mortality in severe mental illness: a cohort study. Lancet Psychiatry. 2017;4(5):389-99.
+15.	Reininghaus U, et al. Mortality in schizophrenia and other psychoses: a 10year follow-up of the ^SOP first-episode cohort. Schizophr Bull. 2015;41(3): 664-73.
+16.	Das-Munshi J, et al. Ethnic density and other neighbourhood associations for mortality in severe mental illness: a retrospective cohort study with multi-level analysis from an urbanised and ethnically diverse location in the UK. Lancet Psychiatry. 2019;6(6):506-17.
+17.	Jorgensen KB, Nordentoft M, Hjorthoj C. Association between alcohol and substance use disorders and psychiatric service use in patients with severe mental illness: a nationwide Danish register-based cohort study. Psychol Med. 2018;48(15):2592-600.
+18.	Williams JM, et al. Increased nicotine and cotinine levels in smokers with schizophrenia and schizoaffective disorder is not a metabolic effect. Schizophr Res. 2005;79(2):323-35.
+19.	Mark Weiser MD, et al. Higher rates of cigarette smoking in male adolescents before the onset of schizophrenia: a historical-prospective cohort study. Am J Psychiatry. 2004;161(7):1219-23.
+20.	World Health Organization. Guidelines for the management of physical health conditions in adults with severe mental disorders. Geneva: World Health Organization: Licence: CC BY-NC-SA 3.0 IGO; 2018.
+21.	Barbui C, et al. Challenges in developing evidence-based recommendations using the GRADE approach: the case of mental, neurological, and substance use disorders. PLoS Med. 2010;7(8):e1000322.
+22.	Guyatt GH, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924.
+23.	World Health Organization. World Health Organization Model List of Essential Medicines: 21st List. Geneva: World Health Organization (WHO); 2019.
+24.	Lexicomp. Lexi-Interact. https://online.lexi.com. 2018.
+25.	Kheshti R, Aalipour M, Namazi S. A comparison of five common drug-drug interaction software programs regarding accuracy and comprehensiveness. J Res Pharm Pract. 2016;5(4):257-63.
+26.	Shea BJ, et al. Development of AMSTAR: a measurement tool to assess the methodological quality of systematic reviews. BMC Med Res Methodol. 2007;7(1):10.
+27.	Kung J, et al. From systematic reviews to clinical recommendations for evidence-based health care: validation of revised assessment of multiple systematic reviews (R-AMSTAR) for grading of clinical relevance. Open Dent J. 2010;4:84-91.
+28.	Shea BJ, et al. AMSTAR is a reliable and valid measurement tool to assess the methodological quality of systematic reviews. J Clin Epidemiol. 2009; 62(10):1013-20.
+29.	Balshem H, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011;64(4):401-6.
+30.	Peckham E, et al. Smoking cessation in severe mental ill health: what works? An updated systematic review and meta-analysis. BMC Psychiatry. 2017; 17(1):252.
+31.	Tsoi DT, Porwal M, Webster AC. Interventions for smoking cessation and reduction in individuals with schizophrenia. Cochrane Database Syst Rev. 2013;2013(2):CD007253.
+32.	Secades-Villa R, et al. Psychological, pharmacological, and combined smoking cessation interventions for smokers with current depression: a systematic review and meta-analysis. PLoS One. 2017;12(12):e0188849.
+33.	Roberts E, et al. Efficacy and tolerability of pharmacotherapy for smoking cessation in adults with serious mental illness: a systematic review and network meta-analysis. Addiction. 2016;111(4):599-612.
+34.	Hunt GE, et al. Psychosocial interventions for people with both severe mental illness and substance misuse. Cochrane Database Syst Rev. 2013;10: CD001088.
+35.	Boniface S, et al. The effect of brief interventions for alcohol among people with comorbid mental health conditions: a systematic review of randomized trials and narrative synthesis. Alcohol Alcohol. 2017;53(3):282-93.
+36.	Temmingh HS, et al. Risperidone versus other antipsychotics for people with severe mental illness and co-occurring substance misuse. Cochrane Database Syst Rev. 2018;1(1):CD011057.
+37.	Wilson RP, Bhattacharyya S. Antipsychotic efficacy in psychosis with comorbid cannabis misuse: a systematic review. J Psychopharmacol. 2015; 30(2):99-111.
+38.	Agabio R, Trogu E, Pani P. Antidepressants for the treatment of people with co-occurring depression and alcohol dependence. Cochrane Database Syst Rev. 2018;4:CD008581. https://doi.org/10.1002/14651858.CD008581.pub2.
+39.	Dennis BB, et al. Opioid substitution and antagonist therapy trials exclude the common addiction patient: a systematic review and analysis of eligibility criteria. Trials. 2015;16:475.
+40.	Liu NH, et al. Excess mortality in persons with severe mental disorders: a multilevel intervention framework and priorities for clinical practice, policy and research agendas. World Psychiatry. 2017;16(1):30-40.
+41.	World Health Organization. World Health Organization. SAFER alcohol control initiative to prevent and reduce alcohol-related death and disability. 2018. https://www.who.int/substance_abuse/safer/en/. 2018 [cited 2019].
+42.	World Health Organization. 'Best buys’ and other recommended interventions for the prevention and control of noncommunicable diseases: Updated (2017) appendix 3 of the global action plan for the prevention and control of noncommunicable disease 2013-2020. Geneva: World Health Organization; 2017. https://www.who.int/ncds/management/WHO_ Appendix_BestBuys.pdf.
+43.	World Health Organization, Global strategy to reduce the harmful use of alcohol, World Health Organization, Editor. 2010, World Health Organization: Geneva.
+44.	World Health Organization, Tackling NCDs: 'Best buys’ and other recommended interventions for the prevention and control of noncommunicable disease. World Health Organization, Editor. 2017, World Health Organization: Geneva. https://apps.who.int/iris/bitstream/handle/1 0665/259232/WHO-NMH-NVI-17.9-eng.pdf?sequence=1&isAllowed=y.
+45.	World Health Organization. Mental Health Action Plan: 2013-2020. https:// apps.who.int/iris/bitstream/handle/10665/89966/9789241506021_eng. pdf?sequence=1. Geneva: World health organization; 2013.
+46.	Humphreys K, Blodgett JC, Roberts LW. The exclusion of people with psychiatric disorders from medical research. J Psychiatr Res. 2015;70:28-32.
+47.	Saxena S. Excess mortality among people with mental disorders: a public health priority. Lancet Public Health. 2018;3(6):e264-5.
+48.	Degenhardt L, et al. Estimating treatment coverage for people with substance use disorders: an analysis of data from the World Mental Health Surveys. World Psychiatry. 2017;16(3):299-307.
+49.	Saraceno B, et al. Barriers to improvement of mental health services in low-income and middle-income countries. Lancet. 2007;370(9593):1164-74.
+50.	Aldridge RW, et al. Morbidity and mortality in homeless individuals, prisoners, sex workers, and individuals with substance use disorders in high-income countries: a systematic review and meta-analysis. Lancet. 2018; 391(10117):241-50.
+51.	Gilbody S, et al. Smoking cessation for people with severe mental illness (SCIMITAR+): a pragmatic randomised controlled trial. Lancet Psychiatry. 2019;6(5):379-90.
+52.	Jochelson, K. and B. Majrowski, CLEARING THE AIR: Debating smoke-free policies in psychiatric units. Kings Fund, Editor. 2006, Kings Fund: 11-13 Cavendish Square, London. https://www.kingsfund.org.uk/sites/default/files/ field/field_publication_file/clearing-the-air-debating-smoke-free-policies-psychiatric-units-karen-jochelson-bill-majrowski-kings-fund-18-july-2006.pdf.
+53.	Livingston G, et al. Dementia prevention, intervention, and care. Lancet.
+2017;390(10113):2673-734.
+Publisher's Note
+Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
+Ready to submit your research? Choose BMC and benefit from:
+•	fast, convenient online submission
+•	thorough peer review by experienced researchers in your field
+•	rapid publication on acceptance
+•	support for research data, including large and complex data types
+•	gold Open Access which fosters wider collaboration and increased citations
+•	maximum visibility for your research: over 100M website views per year
+At BMC, research is always in progress.
+Learn more biomedcentral.com/submissions
+kBMC