FHIR Chat · Semantic bridge with SNOMED · implementers

Stream: implementers

Topic: Semantic bridge with SNOMED


view this post on Zulip nicola (RIO/SS) (Nov 27 2016 at 18:09):

Is there any work to map FHIR resources and elements to SNOMED concepts? Like FHIR/Patient - Patient (person) SCTID: 116154003, FHIR/Patient.name - Patient name (observable entity) SCTID: 371484003...

view this post on Zulip Grahame Grieve (Nov 27 2016 at 18:10):

yes. see https://onfhir.hl7.org/2016/11/03/summary-report-from-snomed-meeting/

view this post on Zulip nicola (RIO/SS) (Nov 27 2016 at 18:21):

Nice. May be add this information into StructureDefinition? Also map and sync with SNOMED all FHIR ValueSets/CodeSystems. The idea of unified ontology looks like amazing future and could be used in semantic/logical mapping as basis. We could start from snomed driven extensions to reduce duplication in extensions, if concept already in SNOMED. For example define race extension by snomed - Race (observable entity) SCTID: 103579009

view this post on Zulip Grahame Grieve (Nov 27 2016 at 18:24):

you should find what's already done in structure definitions - check, say, Condition

view this post on Zulip nicola (RIO/SS) (Nov 27 2016 at 18:25):

Yes, that's exactly that. We could help with it :)

view this post on Zulip Grahame Grieve (Nov 27 2016 at 18:26):

as for race.... tough subject. SNOMED CT doesn't have enough definitions to make concepts interoperable. Race is a great example, since it means such wildly different things depending on context. But more importantly, SCT has very poor coverage

view this post on Zulip Grahame Grieve (Nov 27 2016 at 18:26):

@Linda Bird is the person to help

view this post on Zulip nicola (RIO/SS) (Nov 27 2016 at 18:58):

https://gist.github.com/niquola/06a59517ee409077f9462900fb01f4b9 just played little bit with mapping

view this post on Zulip Grahame Grieve (Nov 27 2016 at 19:02):

comments:
- the mapping from Patient to 116154003 is only true if patient is a human
- technically, it's not true then because of the the difference between record/real world entity, but that's probably not something that matters
- the others except for the general practitioner mapping are probably ok. the generalPractitioner mapping is wrong because the SCT code is 'patient has a GP caring for them' not 'the identity of the practitioner caring for them'


Last updated: Apr 12 2022 at 19:14 UTC