Researchers develop AI tool to predict patients at risk of intimate partner violence | National Institutes of Health (NIH)
Publish Date: 2026-03-13 09:17:00
Source Domain: www.nih.gov
- NIH-funded researchers developed an AI tool that aids clinicians in detecting intimate partner violence (IPV) risk based on routinely collected medical data.
- The AI model used machine learning techniques and showed high accuracy in detecting IPV among patients.
- IPV, affecting millions in the U.S., often goes undetected due to patients’ reluctance to disclose abusive relationships.
- The study introduced three AI models for IPV detection, with the multimodal fusion model achieving the highest accuracy (88%).
- The fusion model effectively detected IPV risk over three years before patients seek help at domestic abuse intervention centers.
- Researchers emphasized that the AI tool supports timely clinician-patient conversations about IPV, not making definitive diagnoses.
- The tool, part of a proactive approach to IPV intervention, aims to initiate earlier, safer, and more informed discussions.
- The study hopes the AI tool will improve long-term health outcomes for at-risk patients when used in a supportive patient-centered manner.
