Stream: terminology
Topic: CodeSystem that's a subset of another CodeSystem
Brendan Keeler (Feb 20 2022 at 16:14):
Is it possible for a CodeSystem to be a subset of another CodeSystem? Or would that only be expressed as a ValueSet in FHIR?
Brendan Keeler (Feb 20 2022 at 16:16):
Thinking specifically about NCIT
http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl
and NCPDP
https://datascience.cancer.gov/resources/cancer-vocabulary/ncpdp-terminology
Lloyd McKenzie (Feb 20 2022 at 17:01):
Either there's one overall namespace (CodeSystem) of which a particular subset of codes is a ValueSet (e.g. LOINC answers within LOINC) or you have a code system that is constructed by defining a grammar that combines codes from other code systems (e.g. IETF language codes). In the first case, it's possible that the overall coordinating organization assigns responsibility to manage the subsets to distinct organizations. There are also situations where one organization defines codes that correspond to codes from lots of other code systems (e.g. NCI metathesaurus).
Michael Lawley (Feb 20 2022 at 20:09):
It is not uncommon for ontologies to be built in such a way - e.g., HPO is built from many sub-ontologies and each could reasonably be a code system in their own right. See https://pubmed.ncbi.nlm.nih.gov/32308861/ for a discussion of the issues this causes when trying to represent them as FHIR CodeSystem resources.
Last updated: Apr 12 2022 at 19:14 UTC
