For decades, the standard practice in medical research and data sharing involved “anonymizing” patient records. The process seemed simple and secure: remove explicit identifiers like names and addresses, and the data could be safely used to train models, study populations, … Read More