AI Tool Spots Domestic Abuse Risk 3 Years Early

Doctors now have a powerful new ally in the fight against domestic violence: an AI tool that can spot warning signs years before victims are ready to ask for help.

Researchers in the United States developed a machine learning system that analyzes hospital records to identify patients at risk of intimate partner violence. Published in Nature, the technology flagged potential abuse cases more than three years before many patients later entered hospital-based intervention programs.

The AI system studied records from nearly 850 women who had experienced partner violence and compared them to more than 5,200 similar patients. It looked for patterns in medical history, doctor's notes, and injury reports that often go unnoticed during routine care.

The most accurate version combined all available data and correctly identified risk in 88% of cases. That matters because current screening methods rely on doctors asking direct questions, which many victims cannot answer honestly due to fear or safety concerns.

The technology doesn't diagnose abuse or force anyone to disclose information. Instead, it gives healthcare providers a quiet signal that a patient might need extra support, allowing them to approach the topic with care and offer resources when appropriate.

"This clinical decision support tool could make a significant impact on prediction and prevention of intimate partner violence," said Qi Duan, program director at the National Institute of Biomedical Imaging and Bioengineering. About 18% of European women with partners experienced physical or sexual violence in 2021, according to the European Commission.

Why This Inspires

This breakthrough represents hope for millions who suffer in silence. By working behind the scenes, the AI gives doctors information they need without putting patients in danger or pressuring them to speak before they're ready.

The tool recognizes something crucial: victims often interact with the healthcare system for years before finding the courage to disclose abuse. Every emergency room visit, every unexplained injury, every pattern in the records tells a story that human eyes might miss in the rush of daily care.

Researchers are now working to integrate the technology into electronic medical record systems for real-time assessments during routine appointments. The goal is supporting clinicians, not replacing their judgment or the human connection that makes disclosure possible.

"Our work represents a shift towards recognizing risk earlier using information already present in healthcare data," said Bharti Khurana, an emergency radiologist at Mass General Brigham and Harvard Medical School.

For people trapped in dangerous situations, this could mean the difference between years of suffering and getting help when it matters most.