New AI tool can assess suicide risk

A new AI-powered assessment tool is able to predict the likelihood that someone is exhibiting suicidal thoughts and behaviours, a study has found.

The tool, developed by researchers at Northwestern University, the University of Cincinnati, Aristotle University of Thessaloniki and Massachusetts General Hospital/Harvard School of Medicine, asks people to perform a task that involves ranking pictures, and combines these with a small set of contextual/demographic variables to predict risk. The research found that the tool was on average 92% effective at predicting four variables related to suicidal thoughts and behaviours.

The study, published in the journal Nature Mental Health, concludes that a small set of behavioural and social measures plays a key role in predicting suicidal thoughts and behaviours. The AI tool being developed as a result could become an app for medical professionals, hospitals or the military to provide assessment of who is most at risk of self-harm.

Data was collected from surveys completed in 2021 by 4,019 participants aged 18 to 70 across the United States. Identities of participants were protected and not shared with researchers, and participants gave informed consent.

Participants were asked to rank a random sequence of 48 pictures on a seven-point like-to-dislike scale of 3 to -3 in six categories: sports, disasters, cute animals, aggressive animals, nature and adults in bathing suits. Researchers also collected a limited set of demographics about age, sex assigned at birth, race or ethnicity, highest education level achieved and handedness.

As well as rating the pictures, participants completed a set of mental health questions and were asked to rank perceived loneliness on a five-point scale.

A ‘quantitative’ understanding of mental health

When the data was fed into the AI tool, the software was able to predict four measures of suicidal thoughts and behaviours: passive suicidal ideation (desire without a plan); active ideation (current and specific thoughts); planning for suicide; and planning coping strategies to prevent self-harm.

“The usage of a picture-rating task may seem simple but understanding individual preferences and how one evaluates reward and aversion plays a large role in shaping personality and behaviour,” said co-author Aggelos Katsaggelos of Northwestern University. “We find that our results in predicting suicidality exceed typical methods of measurement without using extensive electronic health records or other forms of big data.”

Shamal Shashi Lalvani, a doctoral student at Northwestern University, and first author on the study, said: “A system that quantifies the judgment of reward and aversion provides a lens through which we may understand preference behaviour. By using interpretable variables describing human behaviour to predict suicidality, we open an avenue toward a more quantitative understanding of mental health and make connections to other disciplines such as behavioural economics.”

Hans Breiter, professor in computer science and biomedical engineering at the University of Cincinnati, said: “It’s reported we have about 20 suicides daily among veterans in the US, and a salient number of students. We all can cite statistics to how the American medical system is at a breaking point. I wish we’d had this technology sooner. The data strongly argues it would change outcomes.

“People have developed good techniques with big data, but we have problems interpreting the meaning of many predictions based on big data. Having a small number of variables grounded in mathematical psychology appears to get around this issue and is needed if current machine learning is ever going to approach the issue of artificial general intelligence.”

One potential limitation of the study, the researchers said, was that the surveys were self-reported – though as they pointed out, it’s difficult to see how a prospective study of suicide might be performed. They also noted that the cohort was sampled during the pandemic, at a time that has seen higher-than-normal rates of loneliness and self-harm.