Stream: terminology
Topic: Argonaut's fictional code system
Grahame Grieve (May 21 2018 at 22:26):
@Eric Haas @Rob Hausam @Brett Marquard - my attention has just been drawn to this: http://www.fhir.org/guides/argonaut/r2/ValueSet-condition-category.xml.html. Is there any reality behind the code system "http://argonaut.hl7.org"?
Brett Marquard (May 21 2018 at 22:56):
Just a local code system we used during development -- for US Core (STU3) see http://hl7.org/fhir/us/core/ValueSet-us-core-condition-category.html
Brett Marquard (May 21 2018 at 22:57):
Did we miss a step to make the codeSystem 'official'?
Grahame Grieve (May 21 2018 at 22:57):
that's much better. But it's not what the vendors are actually using...
Grahame Grieve (May 21 2018 at 22:59):
did we look at snomed?
Grahame Grieve (May 21 2018 at 23:00):
I do think the definitions could do with a little more work. Is this a classification (e.g. are the terms in here mutually exclusive?)
Robert McClure (May 21 2018 at 23:38):
@Brett Marquard you should inform the argonaut members of this. Cerner is referening the local code system. See https://fhir.cerner.com/millennium/dstu2/general-clinical/condition/ as an example
And yes, this is another one that really needs work. It seems that these codes are a mash-up of categories and situations of use.
Grahame Grieve (May 21 2018 at 23:42):
that's a version issue. it will go away when they update to a later version of argonaut (cause it will be based on US core then)
Robert McClure (May 21 2018 at 23:46):
Problem with assuming an update will fix this is argonauters are fixated on the floor being DTSU2 and that will have an impact on initial expectations for use. I'd bet that Cerner site is what they do now. Can't we get them to fix it because as is, they are describing a non-implementable spec, right?
Brett Marquard (May 21 2018 at 23:46):
"Problem" and "Health Concern" definitely have some overlap -- ONC CCDS required systems to expose both. @Robert McClure , is there a definition in SNOMED or other terminology concept we can reference?
Brett Marquard (May 21 2018 at 23:46):
I think the issue is a bit deeper then us using a local code system...
Grahame Grieve (May 21 2018 at 23:47):
well, that's an argonaut question, I'll ask on the argonaut track. But I suspect the answer will be no, not going to change. Which means that the best option is for us to back-migrate the code system definition from US-Core
Grahame Grieve (May 21 2018 at 23:48):
Brett, I didn't find anything for health concern in SCT. But I think my question and your answer suggest that there's a possibility of better definitional work
Brett Marquard (May 21 2018 at 23:48):
Definitely.
Robert McClure (May 21 2018 at 23:48):
Unfortunately the SCT definitions are not exposed. Not sure if I have solution for that. I have a bigger concern about the inclusion of situation of use ("on the problem list" and "recorded at a specific encounter")
Eric Haas (May 22 2018 at 02:10):
Here is the original codesystem in the spec. In DSTU2 there was this really sweet way to define the URI so I bunched all the codes into a single codesystem. It was part of Valueset back then. Its there after all and so no need to use US Core. I thought Rob M was overstating things a bit by describing it as a "a non-implementable spec" since I had not heard a lot a grumbling about the missing code system ... until now.
At the time we added these categories because they were not there in the base resource and they were defined in the MU regs. We spent a lot of time on this topic and came up with a couple of codes. I don't wish to revisit the definitions unless the implementers of argonaut want to. I remember that the code system resource underwent so major changes so that we had to split these codes up into separate codesystems and subsequent to USCore one of them got added to core spec and so there is overlap there. ( One reason why we are not rushing out a R4 version before the underlying specification is firmed up)
Rob Hausam (May 22 2018 at 03:36):
Eric's explanation makes sense to me. I'm not sure that there is anything in particular that needs to be done at the moment, in anticipation of eventually moving to the R4 version based on US Core.
Grahame Grieve (May 22 2018 at 04:18):
well, maybe i could fix the missing link problem (not linked in the spec). As for the definitions, I wouldn't propose to change the meanings of the code, only to be clearer about their definitions
Eric Haas (May 22 2018 at 06:05):
They are purposefully a bit ambiguous. These were after all categories designed to meet regulatory requirements.
Eric Haas (May 22 2018 at 06:12):
here is the definition in the FHIR spec :
http://hl7.org/fhir/valueset-condition-category.html:
Problem List Item 'An item on a problem list which can be managed over time and can be expressed by a practitioner (e.g. physician, nurse), patient, or related person.'
vs argo/uscore:
Problem 'The patients problems as identified by the provider(s). Items on the provider’s problem list'
They seem to say the same thing to me.
Grahame Grieve (May 22 2018 at 08:05):
can something be both a problem and a concern? because the definitions are unclear about that
Grahame Grieve (May 22 2018 at 08:05):
the expansion in US core is even more confusing given the additional 2 codes in the expansion
Robert McClure (May 22 2018 at 14:42):
@Eric Haas By "unimplementable" I mean any implementation spec (what would you call the Cerner web site I'm pointing to?) references a code system that does not exist anywhere. Even given the totally loose requirement of the identifier only needs to be a uri and not resolve, a value set is expected to return a real value set resource and this one can't because the code system fails at the url location specified. So yes, this is technnically unimplementable. I'm am a bit fed up with model requirements being held to task but terminology is always "ah, come on - it's good enough, right?"
Eric Haas (May 22 2018 at 15:55):
Like I said before - we published the IG at the same time as R3 and C'est Voila Problem category showed up in both. Based on my experience sitting in on many PC discussions, I think we will be hard pressed to come up with better definitions. MIchelle was instrumental in helping us navigate this tricky area and she should be consulted when she returns.
Eric Haas (May 22 2018 at 15:59):
@Robert McClure The CodeSystem is indeed present as I point out. You have to go in the way back machine to remember that it was represented as a ValueSet resource. However, there is a bad link.
Rob Hausam (May 22 2018 at 16:02):
yes, it does exist and we should fix the bad link as Grahame offered
Michael Lawley (May 23 2018 at 05:30):
As I read it, the difference between a problem and a concern is about where it is recorded, not about the underlying condition
Jenni Syed (May 23 2018 at 16:38):
concerns are not something that would normally be recorded as "problems" in a system. EG: I'm worried about my financial status is a concern
Jenni Syed (May 23 2018 at 16:39):
that category primarily came from the "health concern" ONC requirement, IIRC
Robert McClure (May 30 2018 at 13:42):
@Eric Haas Doing a bit of old-thread review I've noted you pointed out something I completely forgot - that in the olden-days code systems were value sets. So I stand corrected, the content was there, just not as we (me) now expect. And we are again reminded how fragile uri that look like url's are...
Last updated: Apr 12 2022 at 19:14 UTC
