oliveregger
commented
3 years ago just a small comment: it is not an associated xml transform but the $extract operation with a provided Structure Map (which is a representation of the FHIR Mapping Language).
i would hope that not too may hiddens items need to be put in the questionnaire because they could be directly represented in the StructureMap (e.g. if you know the answer item with linkid is '1234' for a specific observation you can describe/define the appropriate observation resource elements directly in the StructureMap)
also the linkage can between the resources can be created directly within the StructureMap.
Review SDC data extraction methods and implement appropriate method. It may be that Definition-based extraction is too cumbersome. Discussion with @oliveregger suggests that Structure Map may be easier along with an associated XML transform.
We may need to introduce (many) hidden items into the questionnaire in order to correctly populate all the appropriate elements of the Composition, etc. It isn't clear how to create multiple resources rather than just a single one.
I've started a discussion on extraction at https://chat.fhir.org/#narrow/stream/179255-questionnaire/topic/SDC.20Extraction.20examples/near/224794152. Hopefully there will be some useful feedback.