Episode 2: Pollution and Suicide - A huge meta-analysis makes huge errors

In 1969, psychiatrist Jerome Motto began sending short, kind letters to people who had refused care after a suicidal crisis, and his results made “caring letters” one of the most cited ideas in suicide prevention. This month, the largest test ever, more than 227,000 veterans who called the Veterans Crisis Line, found the letters made no difference to suicide deaths. Dr. Tyler Black goes back to Motto’s original numbers, walks through the new JAMA Psychiatry study, and asks why the ideas we most want to work deserve our hardest look, with a few words for Zero Suicide along the way.Notes and corrections noticed since recording:* 1) I swapped the ages of the two study populations. Motto’s patients averaged about 34 years old; the veterans in the VCL trial averaged about 54.* 2) The lead author of the VCL trials is Mark Reger, and the program serves callers to the Veterans Crisis Line.* 3) On Motto’s statistics: on reflection, I think a survival test was a reasonable choice. The issue is that his significant result was one-tailed and limited to the first two years. The two-sided version of the same test gives roughly p = .09, and no version is significant over the full five years.Reference list:Reger et al., Caring Letters and Suicide Mortality Among US Veterans Who Contacted the Veterans Crisis Line: A Prespecified Secondary Analysis, JAMA Psychiatry, 2026Motto et al., A Randomized Controlled Trial of Postcrisis Suicide Prevention, Psychiatric Services, 2001Reger et al., Caring Letters Sent by a Clinician or Peer to At-Risk Veterans: A Randomized Clinical Trial, JAMA Network Open, 2024Carter et al., Postcards From the EDge Project: Randomised Controlled Trial of an Intervention Using Postcards to Reduce Repetition of Hospital Treated Deliberate Self Poisoning, BMJ, 2005Luxton et al., Caring E-mails for Military and Veteran Suicide Prevention: A Randomized Controlled Trial, Suicide and Life-Threatening Behavior, 2020Comtois et al., Effect of Augmenting Standard Care for Military Personnel With Brief Caring Text Messages for Suicide Prevention: A Randomized Clinical Trial, JAMA Psychiatry, 2019Skopp et al., Caring Contacts for Suicide Prevention: A Systematic Review and Meta-Analysis, Psychological Services, 2023Zero Suicide Institute (Education Development Center), Evidence Base, zerosuicide.edc.org, accessed 2026 Get full access to Suicidology Data and Evidence at suicidology.substack.com/subscribe

Can we predict who will die by suicide? In 1983, Alex Pokorny tried with 4,800 psychiatric inpatients and nearly every rating scale available. More than 96% of the men he flagged as high risk did not die by suicide, and more than half of those who did had been rated low risk. Four decades later, we have machine learning, health records on millions of people, and a Harvard smartphone study reported this year as a “weather app” for suicide risk, with accuracy figures as high as 94%. Dr. Tyler Black works through what those numbers mean, starting with a plain-language primer on sensitivity, specificity and positive predictive value, then looks at what happens to the Harvard model when it meets patients it hasn’t seen before. Along the way he covers the 2025 meta-analysis of 53 machine learning studies, the US Army’s 2026 algorithm, and the VA’s REACH VET program, which has been flagging veterans since 2017. He also argues for what clinicians should do instead: give everyone the basics that work, and build every assessment around the individual person’s own risk and protective factors. If you are struggling, in Canada and the US call or text 988.Reviewing and checking the audio, and responding to feedback, please see the following corrections* US requirements: There is no US law requiring suicide risk stratification. The requirement comes from the Joint Commission’s accreditation standard (NPSG.15.01.01). It requires accredited hospitals and behavioural health organizations to screen for suicide risk, assess anyone who screens positive, and document each patient’s “overall level of risk.”* Large et al. (2016): The 4.84 figure is a pooled odds ratio, not a relative risk. The 5.5% figure (the share of high-risk patients who died by suicide) was reported by the authors themselves.* Spittal et al. (2025): The modelled PPV for suicide death goes above 5–6%. It reached about 17% in people discharged after self-harm. Even there, more than 8 in 10 flags are wrong.* Nock et al. (2026): The held-out range of 0.32–0.80 is for suicide-related events. The strongest predictors were a recent suicide-related event and agitation; aggression was not a predictor. Participants were paid $1 per survey.* SPOT (Simon et al., 2022): The hazard ratio of 1.29 for the online DBT skills arm applies to everyone offered the program, not only those who engaged with it. Engagement was 31% in the care management arm and 39% in the skills training arm.References for this podcast:Pokorny et al., (1983) Prediction of suicide in psychiatric patients. Report of a prospective study, Archives of General Psychiatry.Large et al., (2016) Meta-analysis of longitudinal cohort studies of suicide risk assessment among psychiatric patients: heterogeneity in results and lack of improvement over time, PLOS ONE.Belsher et al., (2019) Prediction models for suicide attempts and deaths: a systematic review and simulation, JAMA Psychiatry.Spittal et al., (2025) Machine learning algorithms and their predictive accuracy for suicide and self-harm: systematic review and meta-analysis, PLOS Medicine.Nock et al., (2026) Using smartphone surveys to predict next-week suicide attempts, Journal of Psychopathology and Clinical Science. Get full access to Suicidology Data and Evidence at suicidology.substack.com/subscribe

In the five months after Robin Williams' death in 2014, the US recorded about 1,800 more suicides than expected, most of them by the same method that was widely reported. Dr. Tyler Black goes through the history and evidence on suicide contagion, from Goethe's Werther to Vienna's subway, shares his own analyses of Robin Williams and 13 Reasons Why, and covers what clinicians, journalists and all of us can do when a famous person dies by suicide.Corrections: In the opening, the roughly 1,800 excess deaths include all methods; about 1,150–1,400 were by the method that was reported.Physicians can self-claim time spent listening to this episode or reading this post under Royal College MOC Section 2 (Personal Learning) or as non-certified Mainpro+ credits. Other health professionals should follow their own college's CPD rules. Get full access to Suicidology Data and Evidence at suicidology.substack.com/subscribe

Thanks to everyone listening so far. Please share and consider subscribing to the substack (suicidology.substack.com) to support this.Resource list:Our paper today: Jones SM, Garcia L, Cleland CL, et al. Environmental exposure and risk of death by suicide, attempted suicide and suicide ideation: a systematic review and meta-analysis .J Epidemiol Community Health Published Online First: 15 September 2026. doi: 10.1136/jech-2025-225491* EurekAlert press release* Kim et al. (2010), Am J Psychiatry* Szyszkowicz et al. (2010), Environmental Health Insights* Braithwaite et al. (2019), Environmental Health Perspectives* Heo et al. (2021), IJERPH* Persico & Marcotte (2026), Journal of Health Economics* Frangione et al. (2022), Environmental Research* Burke et al. (2018), Nature Climate Change* Ro et al. (2026), Nature Mental Health Get full access to Suicidology Data and Evidence at suicidology.substack.com/subscribe

As always, please consider subscribing to this podcast, and supporting ime suicidology.substack.com, so I can continue to create content for suicidology education.References for this show:Bagley, C. (1968). The evaluation of a suicide prevention scheme by an ecological method. Social Science & Medicine, 2(1), 1–14.Garnett, M. F., et al. (2025). Changes in suicide rates in the United States from 2022 to 2023 (NCHS Data Brief No. 541). National Center for Health Statistics.Gould, M. S., et al. (2007). An evaluation of crisis hotline outcomes. Part 2: Suicidal callers. Suicide and Life-Threatening Behavior, 37(3), 338–352.Hoffberg, A. S., et al. (2020). The effectiveness of crisis line services: A systematic review. Frontiers in Public Health, 7, Article 399.Jennings, C., et al. (1978). Have the Samaritans lowered the suicide rate? A controlled study. Psychological Medicine, 8(3), 413–422.Miller, H. L., et al. (1984). An analysis of the effects of suicide prevention facilities on suicide rates in the United States. American Journal of Public Health, 74(4), 340–343.Pirkis, J., et al. (2015). Interventions to reduce suicides at suicide hotspots: A systematic review and meta-analysis. The Lancet Psychiatry, 2(11), 994–1001.Public Health Agency of Canada. (n.d.). Suicide and self-harm: Suicide mortality. Health Infobase. Retrieved September 21, 2026.Shin, S., et al. (2026). Effectiveness of fixed landline crisis phone boxes in preventing suicides: Evidence from Australia and the United States. Epidemiology and Psychiatric Sciences, 35, e59.Stack, S. (2015). Crisis phones – suicide prevention versus suggestion/contagion effects: Skyway Bridge, 1954–2012. Crisis: The Journal of Crisis Intervention and Suicide Prevention, 36(3), 220–224.Too, L. S., et al. (2025). Impact of interventions at frequently used suicide locations on occurrence of suicides at other sites: A systematic review and meta-analysis. Psychological Medicine, 55, e168.Zarkowski, P. (2008). The Aurora Bridge. American Journal of Psychiatry, 165(9), 1126. Get full access to Suicidology Data and Evidence at suicidology.substack.com/subscribe