OHDSI Home | Forums | Wiki | Github

Achilles - Issue with Ethnicity


We ran Achilles at our site and encountered the below error

9 ERROR 5-Number of persons by ethnicity; 3 concepts in data are not correct vocabulary (CMS Ethnicity)

However, when we inspected our ethnicity concept id values, we see that all 5 codes that we used in ethnicity concept id as shown below are from Race and other domains.


In that case, shouldn’t the error message show 5 concepts instead of 3 concepts (in incorrect mapped). Am I interpreting the error message incorrectly?


As discussed in the previous discussion, right now Ethnicity is defined as “Hispanic” and “Not hispanic”, and race is a whole hierarchy were the ethnicities are descendants of the races. This is an artifact of the US system and needs to be fixed. Unless you have a analytic use case I would not worry about that too much.

@Akshay, we found 30 ethnicity values in our source data that map to race codes. And a number of detailed ethnicities that we simply rolled up to the less granular ethnicity concept (Hispanic, Not Hispanic.

@Christian_Reich. I see two important use-cases for expanding the race and ethnicity standard concepts in OMOP - one is to ensure that source systems that use the expanded CDC value sets can retain the same level of granularity in OMOP, the other is to enable more robust research on disparities in health.