Interoperability

HL7, FHIR, and DICOM in context

Interoperability is a product requirement, not an integration afterthought.

Workbook: 30 minutes

What each standard addresses

For non-experts, interoperability means deciding how your product fits into systems that already exist. Some products can succeed as standalone tools early, while others need deep EMR/EHR integration from the beginning. The right strategy depends on your target workflow, customer expectations, and deployment timeline. Use the table as a map, not a substitute for a concrete interface plan with your customers’ IT and clinical stakeholders.

Standards overview
StandardCommon roleFounder concern
HL7 v2Legacy hospital event and order messaging (ADT, ORU, etc.)High site-to-site variability; often mediated by an interface engine
FHIRModern HTTP/API-oriented exchange of clinical resourcesWhich resources, which implementation guides, which partner profile
DICOMMedical imaging objects, metadata, storage, and network transferPACS/RIS workflow, bandwidth, long-term archive sizing

What this page helps you decide

This page helps learners understand integration before promising it. HL7, FHIR, DICOM, IHE profiles, interface engines, and EHR workflows solve different problems and create different maintenance obligations.

Use it before telling a customer the product integrates with their systems. Integration is a workflow, data, security, testing, and support commitment, not only an API feature.

HL7 v2 versus FHIR in practice

HL7 v2 messages are still common in hospitals: pipe-delimited records that carry orders, results, and demographic events. “Standard” does not mean uniform. Local conventions (Z-segments, custom tables) and aging documentation mean two “HL7” feeds can behave quite differently. Your team should plan for mapping work, test environments, and clarification cycles with each site.

FHIR (Fast Healthcare Interoperability Resources) wraps clinical concepts (patients, observations, diagnostic reports) in REST-style resources. That makes modern app integration easier in principle, but you still must answer which implementation guide applies (for example, US Core, national profiles, or payer-specific rules), whether SMART on FHIR is in scope for app launch inside an EHR, and how you handle authentication and consent. FHIR reduces some friction; it does not remove governance or security review.

DICOM and imaging workflows

DICOM is the lingua franca of imaging: acquisition modalities, structured reports, and storage semantics (PACS, VNA) depend on it. If your product touches images or derived measurements, you need a clear story for transfer syntax, routing, priors retrieval, and who owns long-term retention. Imaging integrations often stress networks and storage harder than text-only clinical workflows, so performance and latency targets belong in the same conversation as clinical requirements.

Profiles, IHE, and the interface engine

Standards gain teeth through profiles and testing: IHE (Integrating the Healthcare Enterprise) publishes integration profiles that mix HL7, DICOM, and FHIR in real-world scenarios. An interface engine (also called an integration engine) sits between systems to translate, route, filter, and retry messages. Many “integration problems” are really configuration, monitoring, and ownership problems around that middle layer. When scoping work, ask who operates the engine, how errors surface, and what uptime is guaranteed for your path.

Integration brief before build

Before promising integration, write down the exact system, standard, message or resource type, owner, test environment, authentication method, expected volume, error handling, and support contact. “We integrate with the EHR” is not a requirement. It is the start of a discovery conversation.

Worksheet

Use the integration brief worksheet before estimating build effort or making customer commitments.

Standalone vs integrated

A standalone product can be faster to deploy and easier to control. An integrated product can improve user adoption by fitting existing workflows but usually requires more implementation effort. Many teams use a phased strategy: start standalone with clear value, then integrate where it drives measurable adoption gains. Pair this page with Health Canada and ecosystem strategy when regulatory classification and data location also depend on integration depth.

Map of HL7 FHIR and DICOM interactions in clinical ecosystem.

Pitfalls that catch teams

Standards and learning links

Use these as neutral references for vocabulary and scope. Your customer’s IT policies, national profiles, and contractual testing gates still drive what “done” means.

Practical next step

Write an integration brief that names source systems, target systems, data fields, timing, errors, owners, test environment, and support process.

Previous: Monitoring and model changeNext: Cybersecurity threat modeling