OHDSI Home | Forums | Wiki | Github

ICD10CM within concept_ancestor


(Steve Patterson ) #1

@DTorok @Christian_Reich

We are currently building out our i2b2 web app instance, which is pointed to our OMOP DB (vocab version = v5.0 07-SEP-18 and cdm version = v.5.2). Ideally, we would source the Condition/Diagnoses Ontology from OMOP concept and concept_ancestor and display the ICD10CM hierarchy. Are there plans to incorporate ICD10CM and other OMOP Non-standard concepts into concept_ancestor where the hierarchies are known/structured?

(Anna Ostropolets) #2

While you’re waiting for these two gents answer:
I do not think that non-standard vocabularies hierarchy will be addded. First of all, because we promote using standard vocabularies and their hierarchies to browse and group concepts. Say, if you need diabetes mellitus type 2, just go for it in SNOMED, which will give you not only E codes from ICD10CM, but also the appropriate codes from O group (the exact reason why we use hierarchy!). Secondly, ancestor contains all the vocabs and we can’t afford expanding it endlessly with non-standard vocabularies. And lastly, ICD10 doesn’t have a real hierarchy, but rather a general grouping that you can easily re-create by playing with the codes themselves.

(Dmytry Dymshyts) #3

or you can use concept_relationship table for this purpose, each ICD10CM concept has it’s ancestors connected via “Is a” relationship.

(Christian Reich) #4


As @aostropolets said. ICD-10 is not used to create standard concepts. They are all mapped over to SNOMED. Only the standard SNOMED concepts form the hierarchy. And: that hierarchy sucks. One-parent only, and designed for reporting and billing, not analytics that follows medical or biological principles.

But: To be nice to people like you we also added Isa and Subsumes relationships between ICD10 which you may utilize just like CONCEPT_ANCESTOR records.

But again, we’d rather you wouldn’t.