Redefining FHIR's Role: Adapting to Analytical Structures in a Data Mesh
Finding a proper role for FHIR in an emerging analytical structure
Summary
While excellent for transactional exchange, FHIR is inherently ill-suited for large-scale analytical processing. Organizations must adopt a 'boundary thesis,' using FHIR as an interface while employing optimized formats within the domains.
Details
This article addresses the architectural challenge of determining if FHIR should serve as the native data model in emerging data mesh structures. Although mandated by regulations like the 21st Century Cures Act, FHIR has inherent limitations for analytical use due to its structural nature. The core issue is that FHIR was built for Online Transactional Processing (OLTP), excelling at small-scale operations like single record lookups or updates. However, a data mesh requires Online Analytical Processing (OLAP) capabilities—the ability to aggregate patterns across millions of patient records. This mismatch creates three major bottlenecks: 1) The performance penalty of deep nesting, where complex qualifiers and extensions force analytical engines to expend immense power parsing deeply nested JSON paths. 2) The
Original content copyright by respective publishers