Stream: terminology
Topic: Expansion Profile
Grahame Grieve (Oct 15 2016 at 19:43):
@Michael Lawley - I am implementing Expansion Profile, or at least, an aspect of it. I am using it to allow the expander to request a particular edition of SCT be used. But it seems expansion profile is doing some thing different in the .codeSystem element, specifying to inlude or exclude codes, and I don't really follow quite what the logic of 'include' is.
It makes sense to say
- when using this code system, make sure you use version X
- I don't want you to use any codes out of system Y
is that what the existing profile is saying? If not, what is it saying> the definitions here definitely need improving
Michael Lawley (Oct 16 2016 at 03:48):
I really don't know - I was not directly involved in teh development of the ExpansionProfile proposal - @Reuben Daniels would be the person to ask I think
Grahame Grieve (Oct 16 2016 at 06:27):
well, do you have any opinions from your usage then?
Reuben Daniels (Oct 16 2016 at 13:43):
@Grahame Grieve: Hi Grahame. I think the two bullet points you've stated was the intent of ExpansionProfile.codeSystem.include . So, to your use case, I'm not sure you can restrict to a specific (version non-specific) edition only, you would need to specify a version of that edition – but that is assuming that SNOMED CT Edition URI is not a valid version string for SNOMED CT (ML can comment if it is).
Grahame Grieve (Oct 16 2016 at 23:01):
it is a valid URI. One that I use widely
Grahame Grieve (Oct 16 2016 at 23:01):
if that's the intent, we can mnake it much simpler
Grahame Grieve (Oct 16 2016 at 23:03):
ExpansionProfile.fixedVersion [0..*]
ExpansionProfile.fixedVersion.system : uri [1..1]
ExpansionProfile.fixedVersion.version : string [1..1]
ExpansionProfile.excludedSystem : string [0..*
Michael Lawley (Oct 16 2016 at 23:04):
My understanding is that FHIR has adopted an interpretation of an Edition URI in the version field as meaning the latest Version of that Edition. I think this entirely consistent with the URI spec and it is the perogative of FHIR t o make this interpretation since it is setting the context.
Grahame Grieve (Oct 16 2016 at 23:05):
well, perhapa, but that would need to be consistent with IHTSDO intent.
Grahame Grieve (Oct 16 2016 at 23:05):
and actually, the edition URI doesn't specifically mean 'the latest version' - it just means, that edition, and by default, servers would use the most recent version available
Michael Lawley (Oct 16 2016 at 23:06):
It was certainly an interpretation of the Edition URI that was considered when the spec was developed
Michael Lawley (Oct 16 2016 at 23:07):
...right, so it is convention to use 'latest available', but it's really only saying 'any version of this Edition'
Grahame Grieve (Oct 16 2016 at 23:12):
y
Grahame Grieve (Oct 17 2016 at 04:14):
Back to this question - is the expansion profile meant to override version specific references in valuesets, or not? I have use cases for both...
Grahame Grieve (Oct 17 2016 at 04:37):
GF#12280 - for procedural reasons, I'm in a hurry to apply this and stabilise the build infrastructure
Grahame Grieve (Oct 17 2016 at 04:46):
@Rob Hausam @Reuben Daniels @Michael Lawley @Ted Klein @Robert McClure, @John Gresh opinions on GForge task GF#12280 please
Last updated: Apr 12 2022 at 19:14 UTC
