Duke’s AI predictive model identifies teens with mental health risks

In an American Hospital Association podcast, representatives from Duke University in Durham, N.C,. discuss its artificial intelligence-powered predictive model designed to identify teenagers at high risk for developing a psychiatric illness in the next 12 months. The tool uses standard questionnaires collected in primary care settings. It determines which patients are at risk and what factors are contributing to that risk, such as lack of sleep or excessive device use. Other factors could include conflict within the family or school setting, a history of childhood adversity or family history of psychiatric illness. The tool would enable providers to identify teen mental health risks more efficiently to enable providers to intervene earlier. They believe the tool could be particularly beneficial in assessing patients in underserved areas.