DeIDClinic: A Risk-Aware Pseudonymization Framework for Clinical Text De-identification and Re-identification Risk Assessment
Researchers have developed DeID-Clinic, a framework designed to pseudonymize clinical text and assess re-identification risks. The system integrates transformer models like BioBERT and ClinicalBERT to identify and mask protected health information (PHI). It also includes a novel module for quantifying residual risk using various privacy metrics, aiming to support compliant data sharing. AI
IMPACT Enhances privacy preservation for clinical data, potentially enabling broader research and data sharing.