Researchers led by psychologist Matthew K. Nock at Harvard have developed a prediction model that forecasts 75 percent of suicide attempts and 87 percent of suicide-related crises within a week of their occurrence—a substantial improvement over existing methods, according to findings forthcoming in the October issue of the Journal of Psychopathology and Clinical Science.
The study involved 600-plus high-risk adults and adolescents recruited from two groups: adults who received emergency room psychiatric treatment and young people aged 12 to 19 treated at an inpatient clinic for suicidal thoughts or behavior. Participants completed surveys via an app starting immediately after hospital or clinic release, with six optional surveys daily for three months, then one daily for an additional three months.
Surveys included 20 questions using 0-10 sliders, with three questions measuring suicidal thinking (urge, intent, and ability to resist) and 17 addressing affective states such as negative mood, hopelessness, feeling trapped, isolation, anger, agitation, worry, fatigue, and energy levels. Nearly 500 participants completed at least one full survey, collectively answering more than 77,000 surveys total.
Researchers analyzed both survey responses and metadata, including response timing patterns. The study incorporated a real-time alert system for intervention when participants indicated high suicidal intent.
Among affective states examined, agitation emerged as a far stronger indicator of suicide risk than depression, with an 11 percent increase in suicide attempt likelihood for each additional point of agitation on the sliding scale. This aligns with prior research from Nock’s lab finding that 90 percent of suicide attempt survivors reported feeling an urge to alleviate psychological agitation and pain that was temporary in nature.
Traditional prediction models typically forecast suicide risk over six months, years, or even decades based primarily on self-reported survey data. Nock’s approach uses real-time monitoring to capture fluctuating suicidal thoughts and emotions, reflecting his earlier research demonstrating that suicidal thoughts vary moment to moment for most people.
Nock, the Edgar Pierce Professor of Psychology, noted that 50 percent of people who died by suicide saw a clinician in their final month, highlighting what he describes as a tragic gap in a clinical system based on scheduled appointments that may fail to detect the variable and fleeting nature of suicidal thoughts.
The research team included 20 co-authors from Nock’s lab, Harvard-affiliated hospitals, Harvard T.H. Chan School of Public Health, and other institutions. Nock’s broader research initiatives include using wearable sensors to monitor sleep, heart rate variability, and voice signatures, along with “just-in-time” interventions designed to prompt coping strategies or outreach during periods of elevated risk. Research was partially supported by federal funding from the National Institute of Mental Health.
Key facts
- Harvard researchers predicted 75% of suicide attempts and 87% of suicide-related crises in the week before they occurred
- The study involved 600-plus high-risk adults and adolescents who completed over 77,000 surveys via an app
- Agitation was found to be a stronger suicide risk indicator than depression, with an 11% increase in attempt likelihood per additional point of agitation
- Participants were recruited from emergency psychiatric treatment and inpatient clinics for suicidal thoughts or behavior
- The research is forthcoming in the October issue of the Journal of Psychopathology and Clinical Science