Practitioner.communication has required binding to human-language-1-0.html, whereas FHIR STU Practitioner has required binding to IETF language tags. This means that any valid gpconnect-practitioner using *.communication can not simultaneously be a valid Practitioner.
I think this needs amending. Consider:
a) provide an extension with the UK preferred ValueSet
b) amend cardinality to 0…2 and carry UK preferred ValueSet as a second code within the CodeableConcept
It’s worth noting that Patient.communication isn’t defined, but for consistency it would seem to make sense to add it.
I’ve got no particular preference for IETF or ISO-639 language codes, but FHIR spec has a required binding to IETF, so Practitioner.communication would be invalid if we provide IS0-639 coding without the required IETF coding.
I’m concerned about how we would validate the resource instance. If we use standard libraries we’ll have to reject the instance as it’s invalid. I’m not keen on having to custom validation routines when FHIR profiles can provide computable metadata that can be interpreted by standard libraries …