OpenMed PII

LLMs outperform traditional systems in de-identifying sensitive health data

Researchers have developed a new method using large language models (LLMs) to identify and recover protected health information (PHI) that traditional de-identification systems often miss. By employing institution-speci…