Stream: v2 to FHIR
Topic: Code System Mapping to fixed bindings
Hans Buitendijk (Aug 31 2020 at 18:37):
@Rob Hausam , @Lloyd McKenzie : In v2-to-FHIR we have various fields where an HL7 v2 table value set needs to be mapped to a value set that is fixed and involves a mandatory field. The project team is seeking clarity on the following options and perhaps alternatives to maintain the fidelity of the data sent and passed on.
- Are null flavors acceptable values to populate such FHIR attributes when "unknown" or similar catch-all are not included?
- If null flavors are not acceptable values and no reasonable catch-all is available, and the owner of the field cannot add a catch-all or actual value, how are we to populate such mandatory FHIR attribute.
- When using a null flavor (if acceptable) or other catch-all, is it acceptable to also have an extension on the FHIR attribute at hand to maintain the original value from the v2 message?
Lloyd McKenzie (Aug 31 2020 at 19:10):
Not sure what you mean by 'null flavors'. In FHIR, if you have a required binding, you must send one of the specified codes if you include the element. If an element is mandatory and you don't have a value that's part of the allowed list, you can't create a conformant instance.
Lloyd McKenzie (Aug 31 2020 at 19:10):
You're free to send extensions too, but if the binding strength is required, you can't send an extension instead of a value from the bound value set.
Lloyd McKenzie (Aug 31 2020 at 19:11):
Obviously if you find a place where there's a required binding on a mandatory element in core that you're concerned you won't have a value for, you should raise that ASAP...
Hans Buitendijk (Aug 31 2020 at 20:30):
So we'll have to submit vocabulary update requests then where we run into this. As we work through, we'll do so.
Lloyd McKenzie (Aug 31 2020 at 22:39):
Either that or requests to loosen the binding strength and/or cardinality expectations.
Last updated: Apr 12 2022 at 19:14 UTC
