Stream: IPS
Topic: The emptyReason element in IPS Composition
Thomas Zhou (Dec 09 2021 at 18:44):
The description of the Composition.section.emptyReason in the International FHIR profile shows that it serves for the first layer section sitting in the composition.
However, The Invariant affecting both Composition.section.entry and Composition.section.emptyReason
(https://hl7.org/fhir/composition-definitions.html#Composition.section.emptyReason
RuleA section can only have an emptyReason if it is emptyemptyReason.empty() or entry.empty()
)
has not been conformed in the IPS profile properly.
(https://build.fhir.org/ig/HL7/fhir-ips/StructureDefinition-Composition-uv-ips-definitions.html#Composition.section.entry
ele-1: All FHIR elements must have a @value or children (: hasValue() or (children().count() > id.count()))
)
This issue needs a discussion in the IPS working group.
Here are a few points to bring to the conversation.
First, the emptyReason element should be added in the section:All Slices in the IPS Composition with its full description.
Second, how to handle and address the relationship between the mandatory sections and the emptyReason element in the IPS Composition? ( emptyemptyReason.empty() or entry.empty() ).
This request comes from the PS_CA case. In that, the emptyReason element is used to handle the challenge of the IPS adoption with the relaxation of the mandatory sections.
Rob Hausam (Dec 09 2021 at 19:31):
@Thomas Zhou Currently in the IPS Composition profile we have constrained out the use of Composition.section.emptyReason. We had some recent discussion about whether to revisit that - but I think that we decided to keep it constrained out as it is now. The cmp-2 invariant comes from the base Composition resource, not the IPS Composition profile - and in this case emptyReason.empty() always will be true, so the invariant is satisfied. I assume that you are seeing the same thing?
Thomas Zhou (Dec 10 2021 at 15:56):
Hi Rob, Thank you for your help.
If it has to relax the mandatory sections of the composition in the PS_CA case, can we only keep and use the display element as an alternative to the EmptyReason element?
I look forward to your comment.
Giorgio Cangioli (Dec 10 2021 at 16:19):
Hi @Thomas Zhou .. not sure to have understood what you are proposing.
But the expectation is that implementers provide in any case an entry with a "no infos" or a "not known" code . E.g an AllergyIntolerance resource with a 'no-allergy-info' code (No information about allergies)
Giorgio Cangioli (Dec 10 2021 at 16:23):
Rob Hausam said:
but I think that we decided to keep it constrained out as it is now. ..
Hi @Rob Hausam this is what I recall as well.
The current build has however emptyReason 0..1 and no more 0..0 as it was in the published version.
.. to be discussed in one of the next calls...
Rob Hausam (Dec 10 2021 at 17:35):
Oh, ok. That's the issue. I'll need to see if I can find what we last discussed in the minutes/recordings (I don't recall exactly what the date or dates might have been at the moment).
@Giorgio Cangioli @Thomas Zhou
Last updated: Apr 12 2022 at 19:14 UTC
