Stream: questionnaire
Topic: subjectType in SDC
Paul Lynch (Jun 17 2020 at 23:38):
In the non-SDC Questionnaire, subjectType is 0..*. Why did SDC make that be 1..1? We have a Questionnaire whose subjectType includes "Patient" and "Person". That is a made-up example, but I would think there would be cases where you would not want to limit the usefulness of a Questionnaire to a single subjectType.
Lloyd McKenzie (Jun 18 2020 at 01:36):
Person isn't allowed to be a subject
Lloyd McKenzie (Jun 18 2020 at 01:37):
We had no use-case where multiple subject types for a single questionnaire made sense. If you can come up with a real one, submit a change request...
Paul Lynch (Jun 18 2020 at 13:08):
I don't have a real EHR database, so it I can only come with imagined use cases, like Patient and Person. Is that not realistic, e.g. for a weight & height panel? Also, I just noticed that LOINC's Questionnaires (e.g. https://fhir.loinc.org/Questionnaire/55757-9) do not specify a subjectType, which means they aren't usable in SDC apps, since SDC requires one. That I think, is a real use case, though on the other end of the cardinality. I will submit a change request.
Paul Lynch (Jun 18 2020 at 15:40):
Brian Postlethwaite (Jun 22 2020 at 01:20):
Patient and RelatedPerson could be quite reasonable - and possibly even Practitioner.
Another alternative could be Encounter or EpisodeOfCare
Brian Postlethwaite (Jun 22 2020 at 01:20):
I can very easily imagine forms where those are the same definition.
Paul Lynch (Jun 22 2020 at 14:13):
Lloyd McKenzie said:
Person isn't allowed to be a subject
I don't understand the purpose of that restriction. What if some organization wanted survey random people, perhaps in some neighborhood, asking them SDOH questions (or COVID-19 questions)? What would they use for a subject to represent the person being surveyed? The person is not really a patient, though maybe that resource could be used anyway.
Yunwei Wang (Jun 22 2020 at 15:53):
What is the purpose for subjectType. Does it set the scope for expression (FHIRPath for example) used in items?
Paul Lynch (Jun 22 2020 at 16:35):
I think its purpose is that a Questionnaire designer might wish to ensure that QuestionnaireResponse.subject is of a particular type. ("Might wish to" is currently "has to" in SDC).
Lloyd McKenzie (Jun 22 2020 at 19:32):
Those random people would be Patient - because the context you're dealing with them in is "potential recipients of care".
Lloyd McKenzie (Jun 22 2020 at 19:33):
Person is only used as a linking resource - to indicate that multiple other resources are the same individual and to allow maintenance of shared demographics.
Brian Postlethwaite (Jun 23 2020 at 04:57):
They might be RelatedPersons... (don't totally disagree on Person there, but think still should be able to)
Brian Postlethwaite (Jun 23 2020 at 04:58):
Could be a feedback form.
Lloyd McKenzie (Jun 23 2020 at 04:59):
RelatedPerson would each have to be tied to a specific Patient. Don't think you could do that with random people...
Brian Postlethwaite (Jun 23 2020 at 05:00):
The review form could be checking the quality of the resource data.
Last updated: Apr 12 2022 at 19:14 UTC
