Stream: Orders and Observation WG
Topic: Discovered issue with Observation.interpretationCode
Grahame Grieve (Aug 16 2018 at 23:34):
I am just processing changes to vocab harmonization for v2, and there's a large change in there: table 0078 has been changed to an external table with the content from this v3 code system:
http://build.fhir.org/v3/ObservationInterpretation/cs.html
That's ok, but the binding for Observation.interpretation is to http://build.fhir.org/valueset-observation-interpretation.html, which is now empty. In order for Observation to pass ballot, I have to change the value set to use all codes from http://terminology.hl7.org/ValueSet/v3-ObservationInterpretation instead of http://terminology.hl7.org/CodeSystem/v2-0078.
Please confirm that OO is ok with this immediately. @Hans Buitendijk @Eric Haas @Rob Hausam
Ted Klein (Aug 16 2018 at 23:38):
yes, the value set definition changed, so the FHIR value set definition also needs to change. The original request came from OO btw...from Riki Merrick...this is a consequence of the finalizing of the unification of this terminology across the product families. Only thing left to do is to move it from v3 to UTG when that is ready...
Rob Hausam (Aug 16 2018 at 23:39):
yes, that's it
Eric Haas (Aug 17 2018 at 00:47):
Is this a breaking change?
Since Oo approved the harmonization I assumethey understand the consequences but I guess should alerton the list.
Do we need a tracker. ?
Rob or @Riki Merrick can you be the point on this since you are the expert on 0078
Grahame Grieve (Aug 17 2018 at 00:49):
I committed it - I didn't really have any choice
Eric Haas (Aug 17 2018 at 00:56):
I committed it - I didn't really have any choice
I agree but I want know if I add to the substantive change list. And what tracker to use Just a formality I think.
Eric Haas (Aug 17 2018 at 00:56):
Re being oiint is to explain to the wg the reason for the change if any questions or concerns
Grahame Grieve (Aug 17 2018 at 00:57):
Rob might know which tracker to reference; it would be a general one about 0078. Or we could just attribute it to the UTG process, which are the only changes that are authorised to happen without a task. (I think it basically is a UTG thing)
Rob Hausam (Aug 17 2018 at 01:02):
yes, thats's right - and I can find the tracker
UTG is potentially going to generate a lot more of situations like this for HL7 code systems / value sets going forward
we talked about value set urls and "authoritative source", but I don't think we really discussed changing code system urls, which is pretty much all that matters in an instance
Ted Klein (Aug 17 2018 at 01:12):
I thought that is why you struggled to get it in for this harmonization, so the new urls would get into the FHIR normative ballot...
Grahame Grieve (Aug 17 2018 at 01:26):
so the sweeping URL change yes. But moving individual code systems around, no
Rob Hausam (Aug 17 2018 at 01:28):
the tracker that was referenced in the harmonization proposal is GF#16186 - but it's only about aligning the definition / description and some note text
as I read them, neither the tracker nor the approved harmonization proposal say anything about changing from a FHIR to a V3 value set binding or changing the underlying code system that's the source of the codes
Rob Hausam (Aug 17 2018 at 01:29):
the codes themselves didn't change - they were already aligned
Ted Klein (Aug 17 2018 at 02:34):
Changing that was exactly what was in the final proposal that was approved at harmonization. The final updated package has the final proposal in it, with the final changes from OO and the approvals during the harmonization meeting. You were there when it was discussed and voted, no?
Grahame Grieve (Aug 17 2018 at 02:35):
sure this issue is narrow, about the paper trial for the change
Ted Klein (Aug 17 2018 at 02:36):
actually, the approval vote on the table 78 proposal was : RESOLUTION:
Motion to accept as amended by Riki Merrick, Rob Hausam seconds. Vote: 10/0/0, motion carries.
Ted Klein (Aug 17 2018 at 02:39):
The FHIR part of it does not mention the URL in the FHIR space, but as that is created on import and processing from the V2 content, with the removal of the V2 content and replacement by the V3 object, that surfaces this. In the V3 space, we do not have defined canonical URIs for the code systems, but since this is an HL7 authored and maintained system, then it should be one of the ones that gets the terminology.hl7.org path, no?
Eric Haas (Aug 17 2018 at 04:00):
I will attribute to UTG process for the documentation in the build.
Riki Merrick (Aug 17 2018 at 05:27):
Sorry about joining this conversation late - I am not sure what I should have put in the proposal differently - the goal was to have a single code system with all aligned definitions and coded concepts; whether we point to v2 HL70078 or v3 observationInterpretation does not matter so much to me - what I understood is that we would need to use v3 for the tooling to work right, hence we went with v3 object.
Eric Haas (Aug 17 2018 at 08:40):
Is not a problem just need to make folks aware why it changed.
Rob Hausam (Aug 17 2018 at 12:02):
Agree with @Ted Klein and @Riki Merrick . I was just saying that the proposal itself doesn't provide those details and we didn't vote with precisely how it would be executed in mind, but I agree that how it was done naturally follows. And @Eric Haas will add the explanation, so it should be good.
Last updated: Apr 12 2022 at 19:14 UTC
