Stream: CDA IG Publisher
Topic: CDA template design
Giorgio Cangioli (Aug 27 2020 at 13:01):
I really appreciated the very interesting presentation made yesterday by @Sean McIlvenna about the Consolidated CDA Release 2.2 guide (http://build.fhir.org/ig/HL7/cda-ccda-2.2).
I hope that someone (@Grahame Grieve ?) can help me to better understand how CDA templates are designed.
Giorgio Cangioli (Aug 27 2020 at 13:02):
If I look, for example, to any of the section templates (profile), what I understand is that there is an implementable profile constraining a resource of type Section
<kind value="resource"/>
<type value="Section"/>
<baseDefinition value="http://hl7.org/fhir/cda/StructureDefinition/Section"/>
<derivation value="constraint"/>
with the Section that is a logical model
<kind value="logical"/>
<type value="Section"/>
<baseDefinition value="http://hl7.org/fhir/StructureDefinition/Base"/>
Giorgio Cangioli (Aug 27 2020 at 13:05):
It is not clear to me however what is the actual meaning of a "resource" specializing a "logical model" ?
Are not they belonging to two different layers (implementation and logical) ?
Oliver Egger (Aug 27 2020 at 13:09):
I assume this is an error and should also be of kind 'logical' in cda-ccda-2.2, with a constraint you should not be able to change the kind.
Giorgio Cangioli (Aug 27 2020 at 13:12):
happy @Oliver Egger we are on the same page :-)
Grahame Grieve (Aug 27 2020 at 21:46):
@Giorgio Cangioli I expect that this is historical; we started working on the CDA IG early. It should definitely be logical not resource (@Sean McIlvenna )
Sean McIlvenna (Aug 27 2020 at 21:56):
the c-cda profile should be type "logical"?
Grahame Grieve (Aug 27 2020 at 22:29):
yes I think so
Giorgio Cangioli (Aug 28 2020 at 12:52):
If CDA templates are logical models are they are supposed to be used for validating the CDA instances ?
Giorgio Cangioli (Aug 28 2020 at 12:57):
I believe that one of the main scopes of these CDA IGs will be that of allowing the conformance test, as you can do now with the FHIR IGs...
Oliver Egger (Aug 28 2020 at 15:21):
Giorgio Cangioli said:
If CDA templates are logical models are they are supposed to be used for validating the CDA instances ?
you certainely can do that then
Sean McIlvenna (Aug 28 2020 at 17:22):
so, if it's working like it is now, is there a specific reason to do this? what would be the value of it, beyond being "accurate"?
Grahame Grieve (Aug 28 2020 at 21:43):
well, there's future risk, in that we might start making rules about 'resources' but it's not a resource that should be subject to any rules we make about resources
Grahame Grieve (Aug 28 2020 at 21:44):
otoh, now that I think about, there's something in my past might about the idea that if a ClinicalDocument is a resource, then you could potentially do GET [base]/ClinicalDocument/[id]. The problem is that there isn't an id which brings us back to rules about Resources
Giorgio Cangioli (Aug 28 2020 at 22:17):
I'm still a little bit struggling in understanding how a 'logical model' represents a conformance resource for an implementable specification.; but I agree that a logical model can be instantiated (i.e provide a concrete example with values for that model) and you should be able to validate the compliance of that example against a model.
These instances should in principle belong to the logical level (i.e do not represent concrete implementations), but from a practical perspective, if the tool support this, it is at the end only a formal difference.
Giorgio Cangioli (Aug 28 2020 at 22:18):
Respect to the discussion made in April (https://chat.fhir.org/#narrow/stream/179252-IG-creation/topic/Logical.20Model.20instances) is anything changed.
This is what I get from that conversation :
- you can validate instances against a logical model using the fhir validator
- model instances cannot be included in the IG
Is my understanding correct ?
Grahame Grieve (Aug 28 2020 at 22:19):
they can be included in an IG, but it's a bit round about since they are not 'resources' in the same sense. It's kind of on my to do list to address that, but that also brings us back to id: I couldn't do it if they don't have an id in them
Last updated: Apr 12 2022 at 19:14 UTC
