OHDSI Home | Forums | Wiki | Github

Have PheKB phenotypes been implemented on the OHDSI/OMOP framework?

We are wondering if the PheKB phenotypes, popularized by the eMERGE consortium, have been implemented on the OHDSI/OMOP framework. If so, we were wondering if the scripts are shareable?

Mary,

I know that some OHDSI filks are looking into measures calculations against OMOP v5. I’m not sure if these are Emerge.

I’ve just started digging into an approach to implement both Emerge and NQF measures against OMOP v5. It’s very early phase planning at this point.

Bill

@Mary_Regina_Boland, this was coincidentally just a topic of conversation at Tuesday’s all group meeting. Converting PheKB phenotypes to OHDSI has not been done as of yet but we encourage interested folks to go to CIRCE online and test building a few out through the UI. Any phenotypes built in CIRCE are easily shareable across the community.

Jon

Great thanks @jon_duke and @wstephens for your fast response, we are going to look into it

Hi @wstephens - are you still working on implementing NQF measures or Emerge phenotypes on OMOP v5?

Here’s the PheKB t2dm cohort definition I came up with based on the algorithm:
http://www.ohdsi.org/web/atlas/#/cohortdefinition/5751

One thing to note however: it seems that PheKB is about identifying cases, and not identifying when the person entered and exited the cohort. So, some additional work had to be performed to take the ‘spirit’ of what the PheKB algorithm was doing to infer a start date. In the case of the above phenotype, we are looking for either a T2 Diagnosis, medication, or lab result as the cohort start date, applying additional rules to those dates to see if they qualify as a start date, and then use the person’s observation period end date as the cohort end (we are asserting that once diagnosed with T2 you remain a T2 patient).

-Chris

t