Denmark's digital health system is admired everywhere and copied almost nowhere. The reason is not technology.

Every year, health system delegations travel to Copenhagen to study a national digital health infrastructure that genuinely works. Every year, most of them return and build something that does not.

The failure is not one of effort or of engineering. It is that the visible layer, the portal and the shared record, rests on two foundations that are almost never transferable.

Foundation one: a real national identifier

The Danish personal identification number is used across health, tax, banking and municipal services, it is universal, and it is socially accepted. A health system attempting the same architecture without a national identifier has to solve patient matching probabilistically, and probabilistic matching at national scale produces exactly the class of errors that destroys clinical trust.

Foundation two: patient expectation

Danish patients have been able to read their own records for long enough that reading them is unremarkable. That expectation does quiet work. It creates a continuous, distributed audit of data quality, and it removes the political energy from arguments about disclosure that consume years elsewhere.

What does transfer

Three things travel well. A single national prescription registry is technically modest and clinically transformative, and it can be built without solving the identifier problem completely.

Standard national terminology for a small number of high-traffic clinical concepts, agreed centrally rather than negotiated per vendor, is achievable in a few years.

And a published, enforced interoperability requirement in public procurement changes vendor behaviour faster than any standards body.

What does not

The portal. Every delegation wants the portal, because the portal is what they were shown. The portal is the last thing a system should build, and building it first produces a well-designed window onto data that is not yet worth reading.