Harvard model accurately predicted 75% suicide attempts 7 days before they happened

A Harvard model successfully predicted 75 per cent of suicide attempts and 87 per cent of suicide-related events a week before they happened. The study answers a crucial question: Who is most likely to attempt suicide and when? Boston Globe reported that Harvard researchers followed 619 adults and adolescents who had visited two Boston hospitals after having suicidal thoughts. They used frequent phone check-ins to predict suicide attempts in the next seven days.

Participants answered 20-question surveys on their phones six times a day for three months. They were asked to rate how hopeless and agitated they felt, how strong their urge to kill themselves was and how capable they felt of resisting that urge on a scale of 1 to 10. By the end, they had 80,000 surveys. The data was fed into machine-learning models and revealed that agitation, not depression, was the main trigger for suicide.

Agitation, not depression, was the trigger for suicide attempt

For every one-point increase in agitation among adults, the chances of a suicide attempt the next week went up by 11 per cent. Agitation is a feeling of extreme irritability and discomfort which caused people to take the extreme step. "While depression is an important risk factor, agitation also seems to be a really strong predictor of near-term suicide attempt risk," study author Matthew Nock said.

Harvard psychology professor Matthew Nock is a leading expert in studying suicide risk. His latest study will be published in the Journal of Psychopathology and Clinical Science in October. Nock’s studies show that half of those who died by suicide visited a clinician in their final month.

While even a one-point increase in agitation increased the chances of a suicide attempt by 11 per cent, a person’s admission that they were able to resist suicidal urges was associated with lower risk.

Those who did not respond were also at higher risk

But not every participant opened every questionnaire, even though they were paid $1 for each of them. Nock said fewer than half of the surveys sent were opened, and participation fell as the weeks progressed. He stated that those who did not answer were also at a higher risk of suicide, and they are trying to understand whether knowing about them can prove useful.

The study predicted nonfatal suicide attempts and suicide-related events. However, two people died during the study, Nock said. He will now develop “just-in-time interventions” to provide help during the time the risk is extremely high. “Now that we’re getting better at identifying when people are at risk, we’re taking steps to try and test out what the most effective ways are to intervene with people in moments of suicidal crises,” he said.

He reiterated that studies have shown that simply bringing up or asking someone about suicide does not trigger them, or make them more likely to end their life. So talk when you think it is important, and seek help.