Suicide is a complex topic. Death by suicide is the result of a behaviour, the origin of which is both unique to the individual who dies yet which may be influenced by factors common across many who die. For example, while every path to suicide is characterised by highly personal factors, many of those who die by suicide share the commonality of a mental illness and some, a possible genetic tendency to suffer in this way as well (Bondy et al, 2006; Turecki, 2014; Turecki et al, 2015). Suicide has also been historically stigmatized with taboos and religious connections. It has recently become a target market, with purveyors of various types of suicide prevention products hawking ready-made solutions.

Suicide is also a highly emotional issue for those who have been bereaved. Personally, I (SK) lost a beloved uncle to suicide. Professionally, I (SK) have cared for scores of patients who have attempted suicide but have escaped (so far) the burden of caring for one that was successful. We have tried to help institutions and organisations sort out what to consider when addressing suicide prevention. What we have learned through all this is that we need to know what works and what does not work to prevent suicide. We need solid research to guide policy and practice. We must demonstrate that our interventions actually do decrease suicide rates!

Recently, Zalsman and colleagues have published a paper reviewing the state of the science of suicide prevention (Zalsman et al, 2016). An impressive undertaking, it could have provided us with an opportunity to learn clearly what does or does not work in many different aspects of suicide prevention. And, while it may help, unfortunately, it does not hit the target where it matters; in the bull’s eye.

The science behind the review is one concern. For example, the authors claim they conducted a systematic review, but they restricted their literature search to PubMed and Cochrane, and thus may have missed studies listed on other key databases such as PsycINFO, Embase, CINAHL, and ERIC, (Cochrane Handbook, 2011) a significant omission given the contrary conclusions of others (Harrod et al, 2014; Wei et al, 2015; Kutcher et al, 2016) relating to a component of the review. Second, the authors rated the included studies using Oxford criteria, but aggregated together studies from different levels of evidence in their analysis and did not clarify what level of evidence they chose to use to support their conclusions. This does not provide the reader confidence that they based their conclusions on the best research. Furthermore, they did not use appropriate analytic templates for assessing systematic reviews (for example AMSTAR: see Shea et al, 2007) but instead inappropriately applied Oxford Criteria to rate them. And, in their justification for studies/reviews not included, used a vague description “because of irrelevance or very low evidence”. In other words, they may have mixed the good, the bad and the ugly.

Of additional concern, the authors did not differentiate interventions as to effectiveness on the primary outcome of interest; death by suicide, from those using proxy measures of weak (such as suicidal ideation) or even dubious (such as treatment seeking and identification of at-risk individuals) relationship to completed suicide. This indeed is unfortunate phenomenon in the suicide prevention literature. More papers are published reporting on often distal or weak proxy measures of suicide instead of reporting on the primary outcome: suicide.

In short, the authors have conducted a review of some of the recently available suicide prevention literature but how they did their work tempers reader comfort for any substantial conclusions as to the effectiveness of suicide prevention interventions in actually preventing suicide. Unfortunately, some of the conclusions that they put forward cannot be supported by the evidence they provide. Let’s analyse one component for more detailed assessment: school-based suicide awareness programs.

School-based suicide awareness programs

The reviewers conclude:

School based suicide awareness programs have been shown to reduce suicide attempts (odds ratios [OR] 0.45, 95% CI 0.24 to 0.85; p=0.014) and severe suicidal ideation (0.50, 0.27 to 0.92; p=0.025).

To the reader this conclusion implies that a robust systematic review of the literature was conducted, that only high quality studies with low risk of bias were analysed and that the statistical analysis noted has been derived from a meta-analysis of the studies assessed. Unfortunately all conclusions would be incorrect!

As stated above, an incomplete search of the literature (either of individual studies or analysis of reviews) was reported. Interestingly, a recently published focused systematic review of individual studies (Wei et al, 2015) and a systematic review of reviews using AMSTAR rating criteria (Bennett et al, 2015), which came to different conclusions were not included in the references. A careful look at the studies that were included also tells a different tale than the conclusion presents.

Of the reviews listed in the references (numbers 132, 134, 135, 136), none concluded that school based suicide awareness programs were effective in preventing youth suicide:

• The Cochrane review by Harrod et al (number 132, see Harrod et al, 2014) concluded that the quality of evidence for suicide prevention self-efficacy was low and identified risk of bias in many of the studies
• Of the individual studies listed:
  • Ciffone (number 137: see Ciffone, 2007) only assessed knowledge and attitudes, hardly a suicide prevention study.
  • The Freedenthal study (number 138: see Freedenthal, 2010) did not demonstrate effectiveness and a previous systematic review assessed it as ineffective and not ready for dissemination (Wei et al, 2015)
• The Hoven report (number 139: see Hooven et al, 2010) was not a school based suicide awareness program at all, but a targeted intervention including parents and counsellors delivered to youth previously identified as high risk and primarily assessed various down-stream proxy measures including self-reported symptoms of depression, anger and anxiety
• The review by Klimes-Dougan et al (number 135, see Klimes-Dougan et al, 2013) concluded that while there was some evidence of positive impact on attitudes and help-seeking behavior, “… there was also evidence of no effects or iatrogenic effects”
• Reviews by Cusimano et al, Katz et al, and Robinson et al (number 136, see Cusimano et al, 2011; Katz et al, 2013; Robinson et al, 2013) did not conclude that any school based suicide awareness program was effective in reducing suicide and Robinson pointed out that many studies did not address possible negative impacts
• The Wilcox paper (number 131, see Wilcox et al, 2008) was not a suicide awareness program but a complex intervention designed to improve a variety of pro-social behaviours, directed at early primary grades
• The Aseltine SOS study (number 129: see Aseltine et al, 2007) reported decreased self-reports of suicide attempt in an intervention versus control group but suffers from a fatal flaw. The study did not include baseline measures so it is not possible to determine if the differences reported between intervention and control groups was due to the intervention or not. It also suffers from a high risk of bias and has been assessed as not ready for distribution when applying the OJP framework.

Evidence not included in the review

Not included in the Zalsman review is a recent study (Schilling et al, 2014) of SOS, which showed increased suicide attempts in the intervention group compared to controls, hardly a ringing endorsement for positive impact.

Zalsman’s review did not list numerous other school based suicide related studies/reviews that were published between 2005 and 2015. These include: Wei et al, 2015; Bennett et al, 2015; Schilling, 2014; Reis and Cornell, 2008; Wyman et al, 2008; Tompkins et al, 2009; Clark et al, 2012; Wyman et al, 2010; Gould et al, 2005; Robinson et al, 2011; Miller et al, 2009; Portzky, 2006. None of these studies showed decreases in suicide rates, results on distal proxy measures were mixed and the reviews did not support effectiveness of school based programs in prevention of suicide. It is not clear why these studies and reviews were not included in the Zalsman paper. Indeed, more papers/reviews were left out than were included! Notably some with different conclusions.

Indeed, to our reading, there is only one well-designed and implemented school based suicide prevention study, SEYLE (number 130: see Wasserman et al, 2015) a complex classroom based intervention that has demonstrated a positive impact on incident suicide attempts and severe suicidal ideation. While it requires replication is does demonstrate promising results. In its abstract it notes that YAM (one of the study arms) “was associated with a significant reduction of incident suicide attempts (odds ratios [OR] 0.45, 95% CI 0.24 to 0.85; p=0.014) and severe suicidal ideation (0.50, 0.27 to 0.92; p=0.025)”. If this analysis seems familiar to the reader that is because this is the information also provided in the Zalsman abstract! This analysis refers to the SELYE study ONLY! However, the way in which the Zalsman abstract is written suggests this analysis refers to all school based suicide awareness programs the reviewers evaluated. Misleading at best. Conclusion supporting, not at all.

In our opinion, the author’s recommendation about school based suicide prevention programs is at best premature, misleading and inconsistent with what more rigorous analyses have found (Harrod et al, 2014; Wei et al, 2015; Kutcher et al, 2016; Klimes-Dougan et al, 2013; Cusimano et al, 2011; Katx et al, 2013; Robinson et al, 2013). Unfortunately, the recommendation provided by the Zalsman et al group can give support to the ongoing application of interventions that are not only not known to be effective in reducing suicide rates and are not ready for distribution (Wei et al, 2015), but which may possibly lead to harm (Schilling et al, 2014).

Conclusions

In their final conclusion the authors reminded us that few if any robust public health or clinically focused suicide reduction interventions have demonstrated persistent and positive impact on suicide rates determined through randomised controlled trials (RCTs). Predictably, and to our mind, reasonably, they call for more and better research, using “robust” RCT’s, presumably, although they do not state it, using death from suicide as the primary outcome variable and with a low risk of bias. Parenthetically, their paper suggests that we may have lost a decade of time and spent untold amounts of money on suicide research that has not, in large part, been as helpful to us as it should have been. Indeed, it has been that long since the Mann et al article that this work was meant to update was published. It is discouraging to see how little further ahead we have moved in the last 11 years.

Part of this problem is because public health based suicide prevention interventions have often not measured impact using completed suicide or at least closely linked proxy measures such as hospitalisation for a suicide attempt as outcome measures. It is not very useful to measure downstream proxy indicators such as attitudes, knowledge or even ideation instead of suicide rates as evidence for suicide prevention. Some public health based interventions have been widely applied with little or no evidence for effectiveness or safety. For example, SafeTALK is internationally distributed, based on a total peer reviewed literature of one paper describing 17 veterinary students in Glasgow Scotland who reported satisfaction with their participation (Mellanby et al, 2010). Clinically driven risk assessment and management interventions have suffered from small sample sizes, inadequate control groups and lack of replication (Kutcher, 2016).

So, based on what we know now, what could we be doing to prevent suicide? First, since restricting access to lethal means may be a useful strategy, as Zalsman et al point out, governments could take steps to apply these within their jurisdictions. If this includes controlling access to guns, this may prove problematic in some locations (Barber et al, 2014).

Second, since suicide rates in people living with major mental disorders such as Depression, Schizophrenia and Bipolar Disorder are substantially higher than in the general population, prevention efforts could be directed towards improving mental health care. For example, we know (and pointed out by Zalsman et al, that in follow-up of youth who have been hospitalised for a mental illness, recurrent suicide attempts escalate over time (Goldston et al, 2015); and that a previous suicide attempt identifies the highest level of risk for lifetime death by suicide. This should thus be a cohort for whom targeted prevention research is a priority.

Third, we should curtail our application of community based suicide awareness programs such as SafeTalk for which no evidence of effectiveness or safety exist. They are based in part on the untested idea that encouraging public discourse about suicide will prevent it. While this may be the case in clinical interactions, this observation cannot be presumptively extrapolated. Research first, apply later or at least research with rigour as we apply. Not all good ideas turn out to be good ideas.

Fourth, we must stop confusing each other, policy makers and the public in our language about suicide. The use of such words as “suicidality” or phrases such, as “suicide related behaviours” do not help us move the cursor to where it needs to be. For example, there is only a weak relationship between suicidal ideation in populations and suicide rates. Thus research that reports on reducing suicidal ideation may demonstrate results that will have no impact on suicide rates. We need to focus on suicide as the primary outcome in our research or at least use more meaningful proxy measures, such as rates of hospitalisation as a result of a suicide attempt and stop evaluating programs using distal proxy measures.

Fifth, suicide prevention intervention research in locations where rates are much higher than about 15/100,000 should address social, cultural, socio-economic or other factors that could be at play there. Interventions should be based on solid evidence of effectiveness, an expectation that is not always the case in policy (see for example Health Canada, 2008). Tailoring known effective interventions to the factors at play in specific locations may result in better outcomes than a one size fits all strategy.

In short, there is much research to be done to better understand suicide and improve our efforts to prevent it. We have not done so well to date. The work of Zalsman et al demonstrates that despite decades of work we still have not been very successful and illustrates that we need a clearer way forward, based on research of solid design, with low risk of bias and using appropriate outcome measures. And, as we move forward it is essential to keep in mind that doing something is not the same as doing the right thing.

Links

Primary paper

Zalsman G, Hawton K, Wasserman D, van Heeringen K, Arensman K, Sarchiapone M, Carli V, Höschl C, Barzilay R, Balazs J, Purebl G, Kahn JP, Sáiz PA, Bursztein Lipsicas C, Bobes J, Cozman D, Hegel U, Zohar J. (2016) Suicide prevention strategies revisited: 10-year systematic review, The Lancet Psychiatry, Available online 8 June 2016, ISSN 2215-0366, http://dx.doi.org/10.1016/S2215-0366(16)30030-X.

Other references

Bondy B, Buettner A, Zill P. (2006). Genetics of suicide. Molecular Psychiatry. 11, 336-351

Turecki G. (2014). The molecular bases of the suicidal brain. Nat Rev Neurosci. 15; 802-816.

Turecki G, Brent DA. (2015) Suicide and suicidal behavior. Lancet. 387; 1227-1239.

The Cochrane Collaboration. Cochrane Handbook of Systematic Reviews for Interventions. Version 5.1.0 (updated March 2011): http://handbook.cochrane.org/.

Harrod CS, Gross CW, Stallones L, et al. (2014) Interventions for primary prevention of suicide in university and other post-secondary educational settings (Review). Cochrane Library. 10. Art. No: CD009439

Wei Y, Kutcher S, LeBlanc J. (2015). Hot idea or hot air: a systematic review of the evidence for two widely marketed youth suicide prevention programs and recommendations for implementation. J Can Acad Chid Adolesc Psychiatry. (24); 5-16.

Kutcher S, Wei Y, Behzadi P, (2016) School-and community-based youth suicide prevention interventions: hot idea, hot air, or sham? Canadian J Psychiatry. Pre-publication available on-line.

Shea BJ, Grimshaw JM, Wells GA, et al. (2007). Development of AMSTAR: a measurement too to assess the methodological quality of systematic reviews. BMC Med Res Methodol 7: 10 – 29.

Bennett K, Rhodes A, Duda S, et al. (2015). A youth suicide prevention plan for Canada: A systematic review of reviews. Canadian J Psychiatry. (60): 245-257.

Ciffone J. (2007). Suicide prevention: an analysis and replication of a curriculum-based high school program. Social Work. (52): 41-49.

Freedenthal S, (2010) Adolescent help-seeking and the yellow ribbon suicide prevention program: an evaluation. Suicide Lief Threat Behav. (40): 628-639.

Hooven C, Herting JR, Snedker KA. (2010). Long-term outcome for the promoting CARE suicide prevention program. Am J Health Behav. (34): 721-736.

Klimes-Dougan B, Klingbeil DA, Meller SJ. (2013). The impact of universal suicide-prevention prorams on the help-seeking attitudes and behaviors of youths. Crisis (34): 82-97.

Cusimano M, Sameem M (2011): Injury Prevention. (17): 43-49.

Katx C, Bolton S-L, Katz L. et. Al. (2013). A systematic review of school-base suicide prevention programs. Depression and Anxiety. (30): 1030-1045.

Robinson J, Cox G, Malone A, et al. (2013). A systematic review of school-based interventions aimed a preventing, treating, and responding to suicide-related behavior in young people. Crisis. (34): 164-182.

Wilcox HC, Sheppard GK, Brown H, et. Al. (2008). Drug Alcohol Depend. (95: S-1): S60-S70.

Aseltine RH Jr., James A, Schilling EA, et al. (2007) Evaluating the SOS suicide prevention program: a replication and extension. BMC Public Health. 2007;7:161.

Schilling EA, Lawless M, Buchanan L, et al. (2014) Signs of Suicide shows promise as a middle school suicide prevention program. Suicide Life Threat Behav. 2014; 44(6): 653-67.

Reiss C, Cornell D (2008). An evaluation of suicide gatekeeper training for school counselors and teachers. Professional School Counseling. (11): 386-394.

Wyman PA, Brown CH, Inman J, et al. (2008). Randomized trial of a gatekeeper program for suicide prevention: 1-year impact on secondary school staff. Journal of Consulting and Clinical Psychology. (76): 104-115.

Tompkins TL, Eggert LL, Randell BP. (2009). Does a gate-keeper suicide-prevention program work in a school setting? Suicide and Life-Threatening Behavior. (39): 671—681.

Clark TR, Matthieu MM, Ross A, et al. (2010). Training outcomes from the Good Samaritans of New York Suicide Awareness and Prevention Program among community and school-based staff. Br J Social Work. (40): 2223-2238.

Wyman PA, Brown CH, LoMurray M, et al. (2010): An outcome evaluation of the Sources of Strength suicide-prevention program delivered by adolescent peer leaders in high schools. Am J Public Health. (100): 1653-1661.

Gould MS, Marrocco FA, Kleinman M, et al. (2005) Evaluating iatrogenic risk of youth suicide screening programs: a randomized controlled trial. JAMA. 293: 1653-1661.

Robinson J, Hetrick SE, Martin C. (2011) Preventing suicide in young people: systematic review. Aust N S J Psychiatry. 45: 3-26.

Miller DN, Eckert TL, Mazza JJ. (2009) Suicide prevention programs in the schools: a review and public health perspective. School Psychol Rev. (38): 168-188.

Portzky G, van Heeringen K. (2006) Suicide prevention in adolesents: a controlled study of the effectiveness of a school-based psycho-educational program. J Child Psychol Psychiatry. 47: 910-918.

Wasserman D, Moven CW, Wasserman C, et al. (2015). School-based suicide prevention programs: the SEYLE cluster-randomized controlled trial. Lancet. (385): 1536-1544. (Read the Mental Elf blog of this study here).

Mann JJ, Apter A, Bertolote J. et al. (2005). Suicide prevention strategies: a systematic review. JAMA. (294): 2064-2067.

Mellanby RJ, Hudson NPH, Allister R, et al. (2010). Vet Rec. (167): 730-734.

Kutcher S. (2016). Suicide Risk Management. BMJ: https://online.epocrates.com/u29111016/Suicide+risk+management

Barber CW, Miller MJ. (2014). Reducing a suicidal person’s access to lethal means of suicide: a research agenda. Am J Prev Med. 47: 5264-5272.

Goldston D, Daniel S, Erkanli A, Neilbron N, Doyle O, Weller B, Sapyta J. (2015) Suicide attempts in a longitudinal sample of adolescents followed through adulthood: Evidence of escalation. J Consult Clin Psychol. 83: 253-264.

Health Canada (2008). National Aboriginal Youth Suicide Prevention Strategy (NAYSPS) Program Framework. http://www.hc-sc.gc.ca/fniah-spnia/alt_formats/pdf/pubs/promotion/_suicide/strat-prev-youth-jeunes-eng.pdf

Photo credits

• Daniel Wehner CC BY 2.0
• Jason Taellious CC BY 2.0
• Desi CC BY 2.0