Research from the World Health Organization reveals that almost one in three women has encountered physical or sexual violence from a partner at some point in their lives. Despite this alarming statistic, many victims refrain from seeking assistance from law enforcement or social services. When they do find themselves in medical facilities, numerous women frequently conceal the true source of their injuries. In response to this pressing issue, a team of American researchers has created an artificial intelligence tool designed to recognize indicators of domestic violence by analyzing a patient's complaints alongside their medical history. This system successfully identified risks in over 75% of women who were genuinely in peril. To develop the model, researchers utilized medical records from more than 800 women who sought treatment in hospitals for intimate partner violence during the period from 2017 to 2023. These records included detailed radiology findings, medical backgrounds, and prescriptions, which were then cross-referenced with data from a control group comprising over 5,200 women. The findings indicated that victims of domestic abuse frequently report symptoms like depression and various psychological disorders, in addition to physical complaints such as chest pain. Furthermore, these individuals tend to undergo X-rays and CT scans for injuries affecting their faces, necks, and upper bodies more often, and are regularly prescribed pain medication. The AI model underwent real-world testing on women hospitalized due to violence and demonstrated the capability to pinpoint signs of physical and sexual abuse 3 to 5 years earlier than the typical timeline when such individuals are likely to seek help independently. The question remains: Should hospitals implement such systems? Advocates argue for their use, while others raise concerns about potential privacy infringements.
Informational material. 18+.
