EMR modernization to FHIR-first can happen via rebuild or refactor. Understanding when each applies shapes 3-5 year roadmap decisions.
Rebuild (greenfield FHIR-native)
When: legacy EMR is end-of-life, technical debt makes refactor untenable, team has FHIR expertise available.
Approach: Build on Aidbox, Medplum, or HAPI FHIR. Data migration is the biggest project component.
Timeline: 18-36 months for functional parity with legacy.
Refactor (legacy + FHIR facade)
When: legacy EMR meets clinical needs, migration cost high, incremental modernization preferred.
Approach: FHIR facade on top of legacy EMR. Custom adapters translate legacy API to FHIR REST. Progressive migration of features.
Timeline: 12-24 months for feature-by-feature FHIR exposure.
Hybrid (dual-stack)
When: rebuild horizon exists but incremental value needed now.
Approach: Both systems run; new writes dual-written; reads gradually migrated.
Timeline: 24-48 months for full migration.
Decision matrix
| Situation | Path |
|---|---|
| Legacy end-of-life | Rebuild |
| Legacy meets needs | Refactor |
| Multi-year migration horizon | Hybrid |
| No FHIR expertise | Refactor first |
| Startup or greenfield opportunity | Rebuild |
Common execution mistakes
1. Rebuild without migration plan. 2. Refactor with weak FHIR facade. 3. Hybrid without dual-write discipline. 4. Underestimating team learning curve. 5. Custom auth in rebuild path.
Data migration considerations
1. Terminology mapping (SNOMED/LOINC/RxNorm alignment). 2. Reference integrity across systems. 3. Historical data quality vs. new-write standards. 4. Consent state migration. 5. Audit trail continuity.
Team implications
1. Rebuild: 5-10 FTE FHIR engineers. 2. Refactor: 2-5 FTE integration engineers. 3. Hybrid: both teams during migration.
EMR modernization to FHIR-first is a strategic decision. Match the path to your specific situation and organizational capacity.
