Issue Data linkage law, fee-for-service payment and AI hallucination as barriers to population health management

Summary

Under Korea's Medical Service Act and Personal Information Protection Act, procedures for combining medical and non-medical data are demanding, and fears of re-identification limit data sharing between institutions. Under fee-for-service payment there is little economic incentive to introduce prevention-oriented population health management, because no suitable reimbursement system rewards prevention and management. KISTEP expects a move toward payment based on patient health outcomes, such as prevention and fewer complications, once population health data have been validated. Technically, transmission standards such as FHIR exist, but contextual differences among clinical, genomic, lifelog and social determinant data degrade quality when they are simply merged. Hallucination and black-box behaviour in generative foundation models can produce false health advice, limiting safe use in medicine.