FHIR Chat · Value set binding in a slice · australia

Stream: australia

Topic: Value set binding in a slice


view this post on Zulip John Carter (Mar 28 2018 at 00:35):

How should I interpret the value set bindings associated with the Specialty element in the AUBase Healthcare Service profile here: http://hl7.org.au/fhir/base2018Mar/StructureDefinition-au-healthcareservice.html? I think the intention is to not use the HL7 value set and instead use the SNOMED value set defined in AU, which makes sense to me. But as written, it looks like there's a value set bound (preferred) to the 'container' of the slice, and then a separate value set bound (required) to the actual slice, which doesn't make sense to me. And indeed, I wonder if it ever makes sense to have a value set binding attached to the top level of a slice?

view this post on Zulip John Carter (Apr 03 2018 at 04:38):

How should I interpret the value set bindings associated with the Specialty element in the AUBase Healthcare Service profile here: http://hl7.org.au/fhir/base2018Mar/StructureDefinition-au-healthcareservice.html? I think the intention is to not use the HL7 value set and instead use the SNOMED value set defined in AU, which makes sense to me. But as written, it looks like there's a value set bound (preferred) to the 'container' of the slice, and then a separate value set bound (required) to the actual slice, which doesn't make sense to me. And indeed, I wonder if it ever makes sense to have a value set binding attached to the top level of a slice?

Was this the right forum to ask this question?

view this post on Zulip Brian Postlethwaite (Apr 03 2018 at 04:50):

Yes this is the correct place to ask.
It means you are required to have the SNOMED coding included, but gives the flexibility to be able to have other codings also included.
(permits a local coding, and SNOMED coding in the same value)

view this post on Zulip Richard Townley-O'Neill (Apr 03 2018 at 05:10):

@Brian Postlethwaite I think that slicing means that when the slice applies, the constraints for that slice apply.
That would mean that when the snomedRole slice applies (when the value of coding.system='http://snomed.info/sct'), the value of coding.code must come from the value set Provider Specialties (http://hl7.org.au/fhir/ValueSet/snomed-practitioner-role).
That would mean that this design limits the use of SNOMED CT codes to those from the value set, but does not require that a SNOMED CT code be used.
I do not know what was intended.

view this post on Zulip Michael Lawley (Apr 03 2018 at 05:29):

It would be useful if intent was also documented :)

view this post on Zulip John Carter (Apr 03 2018 at 21:24):

Thank you all for the answers. I guess it just looks funny to me... when I look at the spec page, it reads (to me) that specialty is "preferred" to be coded with the HL7 value set (which itself is a US artifact I helped create a long time ago and therefore of questionable value in AU), but it's "required" to be coded with the SNOMED value set. To get the intent that @Brian Postlethwaite describes above, would it be better to put the required SNOMED at/near the top of the slice and then include an Extension to allow an optional local coding to be specified by the user? That model is used for some of the identifier slices in this group of profiles (although without the added constraint of a value set binding).

view this post on Zulip Richard Townley-O'Neill (Apr 03 2018 at 23:16):

The preferred binding applies to all uses of specialty. Only in the defined slice is the binding to SNOMED required. This was more obvious to me after I looked at the structure definition.


Last updated: Apr 12 2022 at 19:14 UTC