EMR modernization to FHIR-first happens via six patterns. Understanding them shapes 3-5 year roadmap.
Pattern 1: FHIR facade on legacy EMR. Custom adapter translates legacy API to FHIR REST. Best for legacy EMRs with no replacement plan.
Pattern 2: HL7v2 pipeline to FHIR store. HL7v2-to-FHIR IG drives conversion. Best for HL7v2-heavy environments.
Pattern 3: Dual-write during migration. New writes land in legacy + FHIR. Progressive reader migration.
Pattern 4: FHIR-native replacement. Build on Aidbox, Medplum, or HAPI.
Pattern 5: Commercial FHIR-first EMR. Emerging category.
Pattern 6: Hybrid (multiple patterns). Different pattern per workflow area.
Decision matrix
| Situation | Path |
|---|---|
| Legacy end-of-life | Rebuild (4 or 5) |
| Legacy meets needs | Refactor (1) |
| Multi-year horizon | Hybrid (3 or 6) |
| HL7v2 heavy | HL7v2 pipeline (2) |
| Greenfield | FHIR-native (4) |
Timeline realities
| Path | Time to production |
|---|---|
| FHIR facade | 6-12 months |
| HL7v2 pipeline | 6-12 months |
| Dual-write | 24-48 months |
| FHIR-native rebuild | 18-36 months |
| Commercial FHIR-first | 12-24 months |
Common mistakes
1. Rebuild without migration plan. 2. Weak FHIR facade. 3. Hybrid without dual-write discipline. 4. Underestimating learning curve. 5. Custom auth in rebuild.
Team implications
1. FHIR facade: 2-5 FTE integration engineers. 2. HL7v2 pipeline: 2-3 FTE + terminology. 3. FHIR-native rebuild: 5-10 FTE FHIR engineers.
EMR modernization is a strategic decision. Match path to your specific situation and organizational capacity.