Every time I talk about interoperability, I try to start with the essentials: interoperating isn't about “connecting systems.” It's about aligning an ecosystem across four dimensions: legal, organizational, semantic and technical. (And yes: the technical piece is only one part.)

In my experience, when a country or a health network takes interoperability seriously, the real investment looks more like this:

~75%: enabling frameworks (legal/regulatory), organizational agreements, governance, and semantic decisions (data + terminologies).

~25%: technical implementation (platforms, APIs, integration, infrastructure).

So… why do we talk so much about HL7 FHIR when we talk about interoperability?

Because it's the most visible, most “tangible” part. And sometimes we repeat the same old bias: believing that interoperating = software (the same way people believe digital transformation = buying a system).

Adopting HL7 FHIR isn't the starting point. It's a consequence.
Before talking about FHIR, you should normally already have this resolved (or under way):

  • ✅ Use cases: what health problem needs data exchange.
  • ✅ Data sets (primary and secondary use): what data is required and at what quality.
  • ✅ Data structuring strategy (modeling, quality, traceability).
  • ✅ Clinical and administrative terminologies: which ones are adopted and how they're operated (ValueSets, mappings, governance).
  • ✅ Governance: who generates/consumes the data, responsibilities, auditing, consent/security, and operation of shared services.
  • ✅ Enabling legal aspects: legal basis for exchange, privacy, secondary use, retention, etc.
  • ✅ Cross-functional teams (not just IT): clinical, management, legal, statistics/epidemiology, and whoever runs the process.

So what does the decision to use FHIR actually drive?

It puts part of the technical path in order and speeds up the standardization of exchange: it defines a data “language” and a modern way to expose it, with a common core and extension mechanisms for what's specific to each context.

FHIR is key, but it arrives late in the story. If we only talk about FHIR, we're looking at the tip… and leaving underneath what actually determines success.

Because interoperating in health is like cooking well: you can't do it with a single ingredient. Adopting a standard without resolving everything else can be a recipe… for frustration.

When we talk about HL7 FHIR we're not talking about “a piece of software.” We're talking about data: how it's generated, governed, coded, shared and used.