A new AI model developed by researchers in the U.S. aims to enhance early detection of intimate partner violence (IPV) risks among patients. By analyzing comprehensive hospital data, the tool identifies patterns consistent with abuse, allowing healthcare professionals to intervene sooner. IPV affects 18% of women, leading to serious physical and mental health issues. The study, published in Nature, analyzed records of nearly 850 IPV survivors and over 5,200 control patients. Three machine learning models were trained: one focused on structured data, a second on clinical notes, and a combined model that achieved an impressive 88% accuracy in identifying risk. Remarkably, the system can flag potential abuse over three years in advance, enabling timely support for at-risk individuals. This breakthrough underscores the potential of AI in healthcare, particularly in addressing domestic violence. For more information about AI advancements and their impacts, explore our Diplo chatbot today!
