AI Gets Better At Writing Patient Histories When Physicians Engineer The Prompts

When doctors improve the start-up prompts, the ChatGPT artificial intelligence (AI) programme grows better at summarizing patient records. Researchers from Duke and Stanford Universities compared histories of current diseases (HPIs) created by senior internal medicine students to those created by the well-known AI programme. Over three rounds, the outcomes improved, with grades for ChatGPT and human doctors varying on a 15-point aggregate scale by less than a point. The study's research letter, which was published in JAMA Internal Medicine, stated that "these findings underscore the potential of chatbots to aid clinicians with medical documentation." The research team used ChatGPT in January and February to produce HPIs based on three patient interview scripts depicting various types of chest discomfort.