I am not convinced that DV_IDENTIFIER.type should be optional. I think we need to have at least some means to identify what the identifier is. Having said that, I recognise the difficulty in needing to define ID types all the time and there are use cases where we just want to provide an identifier doe a document or referral etc and we currently use DV_TEXT for this.
If there is some thought that we may align with FHIR Identifier which as a system attribute then I could probably accept this, but I tend to think we need to provide clear guidance around the use of id type and assigner as being recommended until we include a system attribute. Issuer is the one I don't see much value in and is redundany.
What do Ocean, Marand, Code24, others currently do in their back-ends for DV_IDENTIFIER.type?
Ocean uses HL7 V2 table 0203 where applicable, this enables adopting national HL7 impelmentation guides and direct transformation from HL7 messages.
Ian - leave optional but with advice on best way to populate.
Bjorn - whether its used or not depends on use case case.
Heath - can be used in archetypes - sometimes constrain type, issuer. But needed in demographic data. E.g. EMR system ids. Recommend - type + assigner. National ID shouldn't need a scoping 'assigner', but others will.