Healthcare Interoperability: Landscape and Approach
Healthcare Interoperability: Landscape and Approach
A patient visits their primary care doctor, who orders lab work. The lab sends results to a specialist, who prescribes medication that the pharmacy needs to verify against the patient’s allergy records. Each system—the EHR, the lab system, the pharmacy system—speaks a different language and stores data differently.
Healthcare interoperability is the ability of these systems to exchange data and use that data once received. When it works, clinicians see complete patient histories regardless of where care was delivered. When it fails, critical information stays trapped in silos, care quality suffers, and administrative costs multiply.
The standards landscape
Most organizations use multiple standards simultaneously:
| Standard | What It Does | When You’ll Use It |
|---|---|---|
| HL7 v2 | Real-time event messaging | Lab results, admissions, pharmacy orders |
| HL7 CDA | Clinical documents | Care transitions, HIE document exchange |
| FHIR | RESTful APIs | Patient apps, new integrations, modern infrastructure |
| DICOM | Medical imaging | Radiology, cardiology, imaging workflows |
| X12 | Administrative transactions | Billing, payer integration, claims |
See: Healthcare Standards Primer, What is FHIR
Integration patterns
| Pattern | Best For | Trade-off |
|---|---|---|
| Point-to-Point | Small environments (2-5 systems) | Scales poorly (n² connections) |
| Hub-and-Spoke | Medium/large (5+ systems) | Single point of failure |
| Event-Driven | High volume, multiple consumers | Complex debugging |
| Batch ETL | Analytics, reporting | Hours/days latency |
Most organizations use hybrids based on scale and latency requirements.
Choosing an approach
Match your approach to use case, constraints, and scale:
- Patient apps: FHIR + OAuth (regulatory mandate)
- Care coordination: C-CDA documents, FHIR queries
- Operational: HL7 v2 + integration engine
- Analytics: Batch ETL or FHIR Bulk Data
- Small orgs (1-3 systems): Point-to-point or vendor bundles
- Mid-size (4-20 systems): Integration engine + ETL
- Large (20+ systems): Hybrid architecture
Where to start
Start with one use case, deliver in 8-12 weeks, then expand:
-
Patient access: Deploy FHIR server; implement US Core resources; add OAuth + SMART App Launch. Mistake: skipping profiles and consent.
-
Legacy integration: Choose integration engine (eg. Mirth); implement one interface end-to-end with monitoring. Mistake: underestimating HL7 v2 vendor differences.
-
Orchestration: Map workflows; implement observability first; start with one workflow. Mistake: building logic in application code.
-
Analytics: Identify sources; build one pipeline completely; validate quality upfront. Mistake: ignoring source semantics.
See: Legacy-to-FHIR Mapping, Data Quality and Validation
Common pitfalls
- Underestimating terminology: Budget 30-40% of time for mapping and code normalization. See Legacy-to-FHIR Mapping and Healthcare Data Semantics
- Ignoring identity matching: Jane Doe ≠ J. Doe without probabilistic matching (EMPI). See FHIR References
- “Boil the ocean” scope: Connect everything → 18 months → nothing works. Start small, deliver in 8-12 weeks.
- Not testing with real data: Sample data is clean; production data has typos, missing fields, unexpected values.
- One-time project mindset: Staff ongoing operations, document everything, automate regression testing. See Data Quality and Validation
Readiness checklist
Before starting, ensure you have:
Foundation:
- Clear business objectives and success metrics
- Regulatory requirements identified
Technical:
- Current state documented (systems inventory, existing interfaces)
- Standards and architecture pattern selected
- Terminology strategy defined (code systems, value sets, mapping)
- Identity matching approach planned
Operational:
- Integration team
- Monitoring and alerting
- Documentation standards established
See also
Continue learning:
- Healthcare Standards Primer - Detailed HL7 v2, CDA, FHIR, X12, DICOM comparisons
- Healthcare Data Semantics - Why terminology and coding matter
- What is FHIR - Modern API-first standard
Implementation:
- Legacy-to-FHIR Mapping - Mapping patterns for legacy data models
- Data Quality and Validation - Data quality strategies for pipelines
- FHIR Profiling - Constraining FHIR resources
- FHIR Terminology - Code systems and value sets