Researchers found a way to detect when hospital software simultaneously accesses patient records — something the current standard ignores entirely

Hospital systems using the industry standard for sharing patient data (HL7 FHIR) have no built-in protection against race conditions — situations where multiple systems try to read or write the same patient record at the same time. Researchers built a formal model to detect these collisions and tested it on synthetic hospital data, catching 90% of concurrent access problems that existing methods miss entirely.

Right now, hospitals assume patient data access happens one at a time, sequentially. In reality, an EHR system, a pharmacy system, and a lab system might all grab the same patient record simultaneously — and nothing stops them from corrupting it or making conflicting changes. This matters because race conditions in patient data can cascade: a drug interaction check runs on stale information, a lab result overwrites a treatment decision, or two systems write conflicting dosing instructions. The model is still theoretical, but it names a structural gap in the standard that affects every hospital system running concurrent access without knowing it.