Imagine agreeing on a national blueprint for a kitchen.
The blueprint describes what a functional kitchen should contain, which elements belong together and what each part is intended to represent. It creates a shared understanding before anyone chooses the software, materials or technical construction.
But every existing house already has a kitchen.
Some were built decades ago. Some combine several functions in one cabinet. Others use local solutions that make perfect sense to the people living there, but are difficult for anyone else to interpret.
Zibs work in much the same way. A zib describes which healthcare information should be recorded for a particular concept and how that information is structured. It does so independently of the technology used to store or exchange it. In Dutch healthcare, zibs provide reusable national building blocks for structured recording and information exchange.
A zib can then be represented and implemented technically in different ways. For FHIR-based exchange, the content of zibs is expressed through FHIR profiles and further refined for particular information standards and use cases. The FHIR representation is therefore not the zib itself, but one technical implementation of the underlying information model.
The national blueprint defines what we have agreed to mean.
The local implementation still has to make that meaning real.
That is where the difficult work begins. Existing systems may store the information differently, combine elements, use local codes or leave important context hidden in free text. Before that information can be represented according to a zib and exchanged through a FHIR profile, someone must interpret the local registration and map it to the shared model.
A system may therefore conform to the technical profile, while the underlying local information was never captured with the same distinctions, context or level of detail that the zib expects.
The structure may have been implemented.
The original meaning may still not fit the blueprint.
This is why adopting a national model is not the same as implementing it consistently. The shared agreement may be clear, while local systems and workflows still shape how it is interpreted in practice.
Perhaps a national model cannot guarantee identical implementation. But it can make the intended meaning explicit enough for us to see where implementations differ — and where local decisions still need to be examined.
The agreement may be national.
The implementation is still local.
.
Next week: Governance
Who maintains meaning once other people and systems depend on it?
SemanticWeekly ConceptMap SemanticInteroperability HealthInformatics
