AI could help hospitals predict PTSD risk in the emergency department

Artificial intelligence could help hospitals identify which emergency department patients are likely to develop post-traumatic stress disorder a month later, new research suggests. The study, published in the journal Nature Communications, describes a tool called PREVIA that analyses a patient's own account of the traumatic event that brought them to the emergency department and predicts their PTSD risk with expert-level accuracy. Millions of people develop PTSD each year, and researchers say the tool could help often-overloaded emergency teams support patients before mental health disorders take hold. The tool remains at the research stage, and further evaluation is needed before routine hospital use. The evaluation analysed narratives from 145 emergency-department patients. In one comparison it flagged all 30 people in that group who developed PTSD. This small-sample result is not proof of perfect accuracy in other hospitals or a diagnosis for an individual patient; prospective testing and checks for missed cases and false alarms remain important.