AI tool PREVIA predicts PTSD risk from emergency room narratives

A single emergency room question provides the signal

PREVIA analyzes patients' recorded and transcribed answers to the routine question, "What brought you to the ED?" The researchers say these spontaneous narratives contain language associated with later PTSD risk, including symptoms during the initial trauma, such as loss of control or amnesia, and early symptoms in the emergency department, such as feeling disconnected from the body.

In a head-to-head comparison involving 30 patients from the same group who later developed PTSD, PREVIA flagged 100 percent of the cases. An emergency department clinician specializing in PTSD correctly identified 89 percent of them.

"This is the first study to show that a patient's spontaneous answers to a single, routine care question ... carries enough of a psychological signal to forecast their PTSD risk a month later," said senior study author Katharina Schultebraucks, an associate professor in the departments of psychiatry and population health at NYU Langone.

How PREVIA builds its risk assessment

The system uses a pipeline of AI models, with different models assigned to different parts of the assessment. Its first layer examines the patient's narrative for trauma-related and early post-trauma symptoms. A second layer considers risk factors such as pre-existing mental health problems and lower socioeconomic status, as well as protective factors including coping skills and strong social support.

A third model combines those assessments into a structured summary. A final layer uses the summary alongside demographic information, trauma type and known PTSD prevalence rates to produce a risk score and a plain-language explanation.

The researchers said this three-layer design consistently outperformed simpler configurations across most performance metrics. They also said the tool is designed to explain its reasoning and connect its predictions to established PTSD risk factors. The results were validated against diagnoses made with the Clinician-Administered PTSD Scale for DSM-5, known as CAPS-5, which the study describes as the gold standard clinical interview for diagnosing PTSD.

Why early screening matters in emergency departments

About 43 million patients are admitted to US emergency departments each year with physical injuries, according to past studies cited in the announcement, and more than a quarter develop PTSD. The researchers describe the first month after trauma as an important period for identifying people at risk, because many patients are no longer in contact with the hospital after that time.

Existing questionnaires, including the Peritraumatic Dissociative Experiences Questionnaire and the Posttraumatic Adjustment Scale, can forecast PTSD development. However, only about 28 percent of emergency departments routinely screen for PTSD, and only two-thirds of those use validated tools. The study authors attributed this in part to overcrowding and the roughly 15, 20 minutes clinicians typically have per patient, which can leave immediate medical needs ahead of lengthy mental health assessments.

"PREVIA can serve as an AI screener that identifies at-risk patients and explains why, without adding any time to intake in busy emergency rooms," Schultebraucks said.

Open-source design could allow independent testing

According to Schultebraucks, PREVIA was built on open-source models and required no training or fine-tuning. That means other hospitals can test the system independently without access to proprietary software or the researchers' patient data, she said.

The study was led by researchers from NYU Langone Health and funded by a National Institute for Mental Health grant, R01 MH129856, with Schultebraucks as principal investigator. The findings support the use of patients' existing emergency room narratives as a possible screening input, while the announcement presents PREVIA as a tool for identifying people who may need timely follow-up care rather than as a replacement for clinical diagnosis.