Adds ResearchStudy aggregation Observations

NIH-NCPI / ncpi-fhir-ig

🔥 NCPI FHIR Implementation Guide

https://nih-ncpi.github.io/ncpi-fhir-ig/

Creative Commons Zero v1.0 Universal

5 stars 2 forks source link

Adds ResearchStudy aggregation Observations #37

Closed bwalsh closed 2 years ago

bwalsh commented 2 years ago

This PR contains the definitions and sample instances to create Observations on a ResearchStudy containing aggregations [Consortium, StudyCohort, Summary, CohortCount, SampleCount, Participant, StorageSize]

Thanks @torstees

bwalsh commented 2 years ago

This looks good to me, except that ncbi url. I'm not sure what the right URL should be:

https://ncpi-fhir.github.io is the prefix that our profiles currently have, as far as I'm aware, however, I think that is out of date after the somewhat recent change in the NCPI group's name. So, maybe it should be nih-ncpi.github.io/fhir? Regardless, that doesn't currently exist, but seems like one we might be able to control via github unless I'm misunderstanding how the github.io subdomains work.

Github pages follow the following: https://username.github.io/repository So, let's go with: https://nih-ncpi.github.io/ncpi-fhir-ig I'll do this for now, and we can refactor.

liberaliscomputing commented 2 years ago

@RobertJCarroll, @allisonheath, and @torstees, I want to open up a discussion about this PR since:

I am not sure whether using Observation for research metadata is within right scope and boundaries; and
If I understand correctly, component should be part of value[x] while this approach doesn't populate value[x]

I am down for the idea of not creating extensions when not necessary, but the above are my concerns.

RobertJCarroll commented 2 years ago

Can you point to the documentation that indicates it is outside of scope? Based on the inclusion of "group" as an acceptable subject reference type, it seems reasonable. Subject: "The patient, or group of patients, location, or device this observation is about and into whose record the observation is placed. "

Can you show us the documentation that describes this limitation? Components are separate from the value of the core observation. The only thing I can find connecting them is Obs-7 that says: "If Observation.code is the same as an Observation.component.code then the value element associated with the code SHALL NOT be present", which doesn't apply as far as I can tell as our components are not using the same code as the observation.

bwalsh commented 2 years ago

One piece additional evidence supporting the idea of using observations to report on ResearchStudies. This term exists to define terms that apply to a "cohort"

From https://terminology.hl7.org/2.1.0/CodeSystem-v3-ObservationValue.html

bwalsh commented 2 years ago

@liberaliscomputing @RobertJCarroll @torstees : thoughts? can we merge?