Condition category - supposed usage?

I’m trying to work out whether the ‘category’ attribute of Condition would be a suitable place for me to add an extension (let’s not get into the purpose of that particular extension right now though, it is a bit off-topic though I might come back on that in another thread).
But I don’t really get the meaning of the attribute, either from its description or from its value set expansion {problem-list-item | encounter-diagnosis}, which also seems to have changed quite a bit from DSTU2 {complaint | symptom | finding | diagnosis}
Can someone point me to a place where the semantics of this attribute is discussed more in detail?

(BTW, when I model, “category” is an attribute name to be avoided, it can mean anything and everything without qualification…)


This is probably an item to take up on, though you may find a past trail in Patient Care Committee minutes and gForge Tasks. Like this search, for instance:[]=tracker&search[]=docman&search[]=news&search[]=frs&search[]=scmsvn&search_by=keywords&search_attach=1&startdate=&enddate=&do_search=Submit

1 Like

Thanks a lot for the link Grahame, some useful discussions there although it makes me suspect that the current value set expansion
might be erroneous? (I can’t see any mention at all about the current enumeration, rather much centering on {complaint | symptom | finding | diagnosis}.
I might turn to FHIRChat as well (have account but no activity so far)

This change was made following the Sept 2016 work group meeting as a result of a combination of trackers all related to Condition.category and the Condition scope, overall. Meeting minutes (which include links to related trackers, such as GF#10091) include:

In general, across many resources in FHIR, the category element has cardinality of 0…* with an example binding strength. Essentially, category is any tag, there can be multiple axes of categorization, etc.

For example, in US-Core (formally known as Argonaut) IG, we chose to have categories for diagnosis (point in time), problem list (provider-maintained concerns), and “other” health concerns (e.g. from patient or related persons).

Thank you very much for the links Michelle!
Yes the extensibility of the value set is what I intended to make use of, I just wasn’t entirely sure about the supposed semantic meaning of this attribute. ‘Type’ and ‘category’ are the kind of attribute names that can, like you say, usually just be tags even of multiple axes, and I wanted to research some more in this case. The links provided by you and Grahame and your feedback are indeed useful to me.