Stream: implementers
Topic: QuestionnaireResponse.item.concept needed?
Marc de Graauw (Dec 02 2016 at 15:45):
Questionnaire.item has concept 0..1, which may store a code for the question - i.e. LOINC Eye opening in the GCS example. QuestionnaireResponse.item does not have a concept 0..1 child, meaning the answer itself can be sent with a LOINC code (see GCS example), but the question cannot be coded. When I request a (set of) Observations in say a DiagnosticReport, I can get coded Observations and coded answers. With QR the only way to get the codes for the questions would be to resolve linkId and get the corresponding Questionnaire.item.concept. Wouldn't it be better to have an optional QuestionnaireResponse.item.concept so QR can be more self-contained?
Grahame Grieve (Dec 02 2016 at 19:20):
so this is the same as 'must it be possible to interpret the QR without knowing the Q?"
Marc de Graauw (Dec 02 2016 at 22:17):
Not "must", but could be convenient. We're gonna use QR to send data to the National Cancer Registry. Items are semantically defined by rheir associated codings; so why not carry those? I think it is along the lines of Q. item. definition, which if I remember your talk at DevDays well is there for similar purposes: aggregate or query sets of data which are meaningful to healthcare pro's.
Grahame Grieve (Dec 03 2016 at 01:16):
I think you want the DataElement link - which is in boyth QR and Q
Marc de Graauw (Dec 03 2016 at 08:00):
I want that too., yes, and I'm gonna use it. But, when I profile a Diagnostic Report (as I'm doing), I need the Structure Definition to validate the DR, but not to interpret it: the semantics are right there, in Observation.code. With QR, I need another thing (Q or DataElement) to look up tbe semantics. Why not model at least those answers which are Observations just as has been done for Observations, so with QR.item.concept? I can (redundantly) repeat Q.item.text in QR to make it more readable and self-contained, why not item.concept?
Grahame Grieve (Dec 03 2016 at 20:07):
given the requirements stated for .text, you'd appear to have a point
Marc de Graauw (Dec 04 2016 at 10:41):
Yes, I think I do. A very ancient (2011) piece which you might recognize starts with: "Resources for Healthcare (RFH) defines a set of "resources" that represent granular clinical concepts. The resources can be managed in isolation, or aggregated into complex documents." W/o QR.item.concept I'd say QuestionnaireResponse violates the first clause of the second sentence.
I'll make a GForge item then.
Marc de Graauw (Dec 05 2016 at 09:23):
Added GF#12429
Last updated: Apr 12 2022 at 19:14 UTC
