some of us have been looking at the idea that a ‘report’ should be a new content type in openEHR, that can include other content of different types, and in different ways.
We have published a wiki page with the initial notes on the idea, and a (possibly) novel approach to solving the quoting/inclusion problem, which is to serialise included content into the report. We think the report type (the name may need to be reconsidered) would apply to discharge summaries, general summaries, referrals, and possibly care plans.
Comments and reactions welcome on the wiki page or here.
Thinking about the hierarchy, at which level will be a Report be? Below compo? Below entry? Structure? Representation?
OT: many asked me this and didn’t had a good answer. Do we have a pattern to model questionnaires? Some require to define questions, and the answer type in most cases is boolean, or coded text (multiple choice), and answers might be 0..* (more than one answer for the same question is valid).
There are several kinds of context archetypes/templates and their meta-data are used for:
de novo data - re-used data
step in the clinical treatment model (observation, assessment/inference, planning, ordering, execution)
kind of interface it is designed for (data presentation on a screen, data capture, database store/retrieve, CDSS, …
Each Template needs to capture all this and is a Composition.
All these contexts are characteristics of a Composition in the end.
Questionnaires are in essence a tool that classifies information.
And sometimes it transforms a set of responses into an aggregated value/code
The questionnaire can be treated as any classification, meaning we need to de fine inclusion and exclusion criteria,
and possible results per question can be a quantitative result (number, PQ, code), or a semi-quantitative result (high, low), or a qualitative result (present/ not present).
Semi-Qualitative results need, inclusion/exclusion criteria and a definition of what the norm/population is is about (females, children, etc.)
Maybe we just need a subclass of Folder that contains a “Locatable” (0..*) attribute. Shouldn’t break things too much. Maybe going the FHIR way is also feasible: Create a new class with a Folder and a container of locatables that can be pointed by the Folder.
Besides specific ways to model questionnaires, my questions is if our openEHR clinical modelers have a pattern to represent questionnaires using the openEHR information model.
Thinking about the hierarchy, at which level will be a Report be? Below compo? Below entry? Structure? Representation?
It is probably a COMPOSITION or specialised form of COMPOSITION - see the class DOCUMENT_COMPOSITION on the wiki page, which is one approach. So the main content is inside (some sort of) COMPOSITION. That's my early thoughts on a model, others may have better ideas.
OT: many asked me this and didn't had a good answer. Do we have a pattern to model questionnaires? Some require to define questions, and the answer type in most cases is boolean, or coded text (multiple choice), and answers might be 0..* (more than one answer for the same question is valid).
Having looked at questionnaires a few times in the past, I would say they are a different pattern, which we also need to elucidate. Ian and others have worked on this in the past and would be able to state what the current flavour is. We can start a separate thread if you like.
The current thinking, at least by the group mentioned on the is that we need to deal with ‘quoted’ or ‘included’ content; also we want to be able to include things like query results and probably other on-the-fly generated content. If we make it a subclass of FOLDER, that would imply that a report (for example a discharge summary) is a collection of content in the FOLDER object, plus, the items referenced by the FOLDER? I’m not sure I understand this. - thomas