Matching Duplicate FHIR Resources in AWS HealthLake

Details

This article describes the technical overview of 'Resource matching' within AWS HealthLake. The feature is designed for scenarios where data from various systems are aggregated, and a single real-world entity might exist as several separate FHIR resources. Resource matching establishes links by identifying the same entity and creating a dedicated 'FHIR Linkage resource' that points to all linked records, leaving the originals unchanged. This allows users to build an accurate, unified view without needing to operate a full master data management system (MDM). A key technical detail is that this linkage relies specifically on healthcare identifiers present in the data; it does not perform probabilistic or demographic fuzzy matching on attributes like names, dates of birth, or addresses. The documentation notes that this capability is currently available as a preview and may be subject to change before general availability.