For multipel values, the decision is basically to put one value into the person table and another in as an observation. The Themis convention includes mention of how to deal with values that are not in the domain_id ‘Race’….basically IIUC get it added to that vocabulary and do not use other vocabularies with Observation domains.
For a single value that means "multiple race", looking more closely at that vocabulary, there are 53 items. Included are non-standard concepts “Other Race” omop id 8522, and “Unknown” 8552. None for anything like multiple, and even then these two are non-standard. If we find such a concept in our data, we'll need to have it added to the Race vocabulary as described in doc.s linked from the Themis recommendation.
Possibly of interest to folks in the Vocabulary and FHIR groups too: @stephanieshong @ShahimEssaid
OHDSI thread here: https://forums.ohdsi.org/t/dealing-with-multiple-races-and-other-exceptions/20091/81
Themis decision and recommendation is here: https://github.com/OHDSI/Themis/issues/71
For multipel values, the decision is basically to put one value into the person table and another in as an observation. The Themis convention includes mention of how to deal with values that are not in the domain_id ‘Race’….basically IIUC get it added to that vocabulary and do not use other vocabularies with Observation domains.
For a single value that means "multiple race", looking more closely at that vocabulary, there are 53 items. Included are non-standard concepts “Other Race” omop id 8522, and “Unknown” 8552. None for anything like multiple, and even then these two are non-standard. If we find such a concept in our data, we'll need to have it added to the Race vocabulary as described in doc.s linked from the Themis recommendation.
Possibly of interest to folks in the Vocabulary and FHIR groups too: @stephanieshong @ShahimEssaid