New AI tool assesses risk of self-harm

Instead of sorting through reams of psychological data, the system focuses on basic demographics, short mental health surveys and a simple ranking task where the patient assigns positive or negative ratings to pictures.

The tool proved as much as 92% effective at predicting four variables relating to suicidal thoughts and behaviors.

UC's research team included Breiter, senior research associates Sumra Bari and Nicole Vike and research associate Byoung-Woo Kim. They collaborated on the study with Northwestern University, Aristotle University of Thessaloniki and the Harvard School of Medicine.

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

Researchers collected demographics about age, gender at birth, race or ethnicity, highest education level achieved and handedness. Patients 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 food.

The patients completed a patient health questionnaire and portions of the State-Trait Anxiety Inventory. Researchers also asked the patients to rank their perceived loneliness on a five-point scale.

Using artificial intelligence developed by UC and Northwestern, the diagnostic tool was able to predict four measures of suicidal thoughts and behaviors: passive suicidal ideation, active suicidal ideation, planning for suicide and planning ways to prevent self-harm.

“The usage of a picture rating task may seem simple, but understanding individual preferences and how one evaluates reward plays a large role in shaping personality and behavior,” said co-senior author Aggelos Katsaggelos, a professor of electrical and computer engineering at Northwestern University.

“We find that our results in predicting suicidality exceed typical methods of measurement without using extensive electronic health records,” he said.

The data was collected from surveys completed in 2021 by 4,019 participants ages 18 to 70 across the United States through Gold Research. The identities of the participants were protected and not shared with researchers. Participants provided informed consent.

Researchers noted that respondents in other countries could have unique cultural influences that affect prediction success. Another potential limitation is the surveys were self-reported rather than through clinical assessments. And the cohort was sampled during the COVID-19 pandemic at a time that has seen higher-than-normal rates of loneliness and self-harm.

Researchers suggested future examinations might be able to leverage contextual risk factors relating to a patient’s age or other demographics to improve predictions. Additional authors include Leandros Stefanopoulos, Martin Block and Nicos Maglaveras.

The study was funded by philanthropist Jim Goetz, UC and the U.S. Office of Naval Research.

Featured image at top: University of Cincinnati researchers developed a diagnostic tool that leverages artificial intelligence to predict whether patients exhibited suicidal thoughts and behaviors. Photo/Unsplash

Editor's note: If you are experiencing a mental health crisis, call or text 988 to connect directly to the Suicide and Crisis Lifeline.