Understanding the FHIR schema deeply pays off for integration engineers. Six concepts to master.
Concept 1: Resource model. 140+ resources representing clinical, administrative, financial healthcare concepts. Core resources (Patient, Encounter, Observation, Condition, MedicationRequest, Practitioner) cover 80% of use cases.
Concept 2: Data types. Primitives (string, integer, dateTime), complex (HumanName, Address, Quantity), reference types.
Concept 3: Extensions. Custom fields via Extension. Each has a canonical URL and StructureDefinition.
Concept 4: Profiles. Profiles constrain base resources. US Core is a profile suite.
Concept 5: Bundles. Bundle types — transaction (atomic), batch (independent), searchset (query response), collection (grouping).
Concept 6: Terminology bindings. Coded fields bound to ValueSets. Terminology module provides operations.
Practical skills
1. Read a StructureDefinition (base resource or profile). 2. Author an Extension StructureDefinition. 3. Use $validate for conformance checking. 4. Distinguish Bundle types by semantic. 5. Follow references (_include, _revinclude). 6. Understand versioning (ETag, If-Match).
Common misunderstandings
1. Treating FHIR resources as generic JSON. 2. Custom fields without StructureDefinition. 3. Ignoring reference type constraints. 4. Bundle type confusion. 5. Missing profile conformance.
Learning resources
1. HL7 FHIR spec — reference. 2. SMART tutorials — auth patterns. 3. FHIR chat — community Q&A. 4. Vendor tutorials (Aidbox, Medplum, HAPI). 5. Inferno — conformance testing framework.
Advanced topics
1. StructureMap for cross-profile transformation. 2. ConceptMap for terminology translation. 3. CQL for clinical logic expression. 4. GraphQL over FHIR. 5. FHIRPath expressions.
FHIR schema mastery compounds. The six concepts above unlock understanding of everything else.
