Genomic data in FHIR uses specific resources and profiles (Genomics Reporting IG) that differ from routine clinical data. Understanding the model prevents force-fitting genomic data into the wrong resources.
Core FHIR genomics resources
1. Observation with genomics profile. Genetic variants reported as Observation with specific SNOMED CT codes and value ranges. 2. **MolecularSequence — Raw sequence data (deprecated in R5 in favor of updated approach). 3. DiagnosticReport for genomic reports. Aggregate results of a genomic panel. 4. Task for genetic counseling referrals. 5. Condition with genetic disease coding** — Diagnosed genetic conditions.
Genomics Reporting IG
The HL7 Genomics Reporting IG defines the profiles for reporting genetic test results. Includes profiles for variant assertions, genomic study, and clinical impact statements.
Integration patterns
1. Laboratory Information System (LIS) to FHIR. LIS reports genetic variants via HL7v2 ORU; HL7v2-to-FHIR converter translates. See HL7v2-to-FHIR IG. 2. Bioinformatics pipeline to FHIR. Variant call format (VCF) files transformed into Observation resources for structured reporting. 3. EHR display of genomic data. SMART on FHIR apps render genomic data in clinician context. 4. Genomic decision support. CDS Hooks fires drug-gene interaction alerts.
Common data challenges
1. Variant coding heterogeneity. HGVS vs. ClinVar identifiers vs. custom names. Standardize on HGVS. 2. VCF-to-FHIR translation. No universal standard; custom pipelines dominate. 3. Terminology binding. SNOMED, LOINC for genetic testing; HGVS for variants. 4. Consent complexity. Genomic data has separate consent regimes.
Vendor tooling
Open Data Platform for Genomics and MITRE genomics tooling are open options. Commercial genomic FHIR platforms are emerging but not consolidated.
FHIR genomics is a real but specialized use case. Sites that need it should invest in the specific IG conformance and dedicated pipelines.