April 28, 2027
A groundbreaking AI tool shows promise in predicting which patients might face intimate partner violence (IPV) long before they ask for help. Developed by a team at Mass General Brigham, this tool could give healthcare providers the ability to identify at-risk patients and offer timely support. Currently, IPV affects millions in the U.S. and can lead to serious physical and mental health issues.
Many victims of IPV are hesitant to speak up due to fear, stigma, and safety concerns. Traditional screening methods often rely on self-reporting, which means many cases go unnoticed. Early detection, however, can lead to quick intervention, improving health outcomes significantly.
Dr. Bharti Khurana’s research team created this AI tool using machine learning. They analyzed electronic medical records from 841 patients involved in domestic abuse interventions, compared against 5,212 non-IPV patients. They developed three predictive models: one using structured data, another utilizing unstructured notes, and a third combining both.
During testing, all models displayed over 80% accuracy, with the best model identifying IPV risk as much as three years before patients sought help. Findings showed particular risk indicators: mental health issues, painkiller usage, and a lack of social support were linked to higher IPV risks. Interestingly, patients who consistently pursued preventive screenings had a lower risk, possibly due to better access to healthcare.
According to Dr. Khurana, the AI tool isn’t meant to diagnose IPV directly. Instead, it aids healthcare providers in identifying patients who may benefit from conversations about IPV and related resources. “This approach encourages supportive dialogue without forcing anyone to disclose,” she explains.
This advancement is crucial. Recent studies show that almost 1 in 4 women and 1 in 9 men in the U.S. experience severe intimate partner physical violence. Yet, only a small percentage of these cases are reported to healthcare providers. By using technology like this AI tool, the goal is to bridge that gap and provide much-needed support.
Experts believe that AI in healthcare can revolutionize how providers approach sensitive issues, allowing for more informed and compassionate interactions with potentially vulnerable patients. As we develop these technologies, their application must be broadened to diverse populations to ensure their effectiveness and accessibility.
This emerging tool could not only change individual lives but might also impact the healthcare system broadly, paving the way for clearer communication and better outcomes for IPV victims.
Related Links
References
Leveraging multimodal machine learning for accurate risk identification of intimate partner violence, Gu J, Carballo KV, Ma Y, Bertsimas D, Khurana B. NPJ Womens Health. 2026;4(1):15. doi: 10.1038/s44294-025-00126-3. Epub 2026 Mar 13. PMID: 41836047.
Funding
NIH’s National Institute of Biomedical Imaging and Bioengineering (NIBIB), Office of the Director (OD).
Source link
Artificial intelligence, disclosure, mental health, intimate partner violence, machine learning, early intervention, diagnosis, radiology, risk assessment, Bharti Khurana, npj Women’s Health
