Researchers at Mass General Brigham have developed innovative artificial intelligence tools designed to pinpoint individuals susceptible to intimate partner violence (IPV). These tools analyze electronic medical records (EMRs) and could potentially identify IPV risks up to four years before a patient seeks assistance from a domestic violence treatment facility, according to a study in npj Women’s Health.
The initiative aims to enhance proactive screening, allowing healthcare providers to engage in critical discussions regarding IPV with their patients. Bharti Khurana, MD, MBA, who led the research, emphasized the importance of early detection, which might enable clinicians to act sooner to mitigate serious mental and physical health issues associated with IPV.
The research team worked alongside experts from the Massachusetts Institute of Technology (MIT), including Dimitris Bertsimas, PhD, to create three machine learning models. These models were trained using data from 673 women who sought help from an abuse intervention center between 2017 and 2022, along with 4,169 matched controls. In testing, the models incorporated both structured and unstructured data, demonstrating a comprehensive approach to identifying IPV.