FHIR Chat · Terminology Capabilities and defaults · terminology

Stream: terminology

Topic: Terminology Capabilities and defaults


view this post on Zulip Michael Lawley (Aug 13 2020 at 07:01):

The current structure of TerminologyCapabilities appears to be skewed to statements about (logical) instances of CodeSystems with respect to a URI and version.
Due to the optionality of most elements, though, it appears possible to make more general statements about behaviour by omitting the version and/or URI.
However, it is not clear how to interpret multiple statements that effectively overlap.

For example, the following is legal:

{
  "resourceType": "TerminologyCapabilities",
...
  "codeSystem": [{
    "version": [{
      "property": ["parent", "child", "status"]
    }]
  }, {
    "uri": "http://snomed.info/sct",
    "version": [{
      "property": ["moduleId", "effectiveTime", "active"]
    }]
  }, {
    "uri": "http://snomed.info/sct",
    "version": [{
      "code": "http://snomed.info/sct/32506021000036107",
      "property": ["hasArtgId"]
    }]
  }]

and could be considered as a way to say that the properties "parent", "child", and "status" are supported for _all_ CodeSystems, with "moduleId", "effectiveTime", and "active" being _additionally_ supported for SNOMED CT, and "hasArtgId" further supported for SNOMED CT-AU.

It is not clear if this strategy (to avoid repeatedly and redundantly say the same thing for many instances) is a legitimate strategy, nor whether we would want to push such a semantics onto the consumers of the TerminologyCapabilities resource?

view this post on Zulip Grahame Grieve (Aug 13 2020 at 11:19):

I'm not sure


Last updated: Apr 12 2022 at 19:14 UTC