This post is a continuation of the conversation I initiated on: https://groups.google.com/a/openmrs.org/forum/#!topic/dev/bwPjAw0zeh0
Basically, i’m trying to list down our plans for representing OpenMRS encounters using FHIR. In the previous email (see link) we discussed how to manage requests for Encounters and visits, and what gets returned each time. Now, lets try to build on that to see how different observations will be grouped together under encounters.
Lets say someone is using OpenMRS, and wants to return a bunch of obs grouped together under an encounter. How will we represent these Obs ? Some of them might be allergies, others may be conditions or diagnosis etc. Basically, what we could do is,
A. Figure out which obs can be categorized as an allergy (recommend using concept class, or if that’s missing, then maybe a concept defined as a global property)
B. Next, check which other obs can be grouped under whatever other FHIR resources (such as diagnosis or condition). If any obs that the OpenMRS FHIR module can support exists, then they would be identified based on concept class or a global property, and new FHIR resources created.
C. In each case, we have different strategies on how to model allergies, diagnosis or conditions etc. based on the underlying data model. Each of these resources will be built based on the appropriate strategy selected based on the underlying OpenMRS data model.
D. All other obs that didn’t fall into any of the FHIR categories supported by OpenMRS be represented as plain old FHIR observations
Thoughts ? comments ? glaring mistakes ?