OHDSI Home | Forums | Wiki | Github

Is "subject_id" the same as "person_id" in cohorts?

Dear all, I’m new to OHDSI and would like to know if there is a way to check if a cohort that I have created with ATLAS contains the same patients that are in another cohort that I have previously created with SAS.
When I look at my cohort in PostgreSQL Studio I can see the variable “subject_id”. Is this the same ID as the variable “person_id” in the dataset from which the cohort was selected? Thanks for your help!

In practice, subject_id = person_id. This is because the OHDSI software that are based on the cohort framework have been written for a patient/person as the unit of analysis.

Please review these posts for relevant past discussions. Bottom line, not many in the community have pushed for use cases where a care_site or provider_id is the unit of analysis. In those use cases, provider_id/care_site_id may become the subject_id. To allow for this interchangability, the term “subject_id” was used. Some have proposed to eliminate this ambiguity and use person_id uniformly.

Thank you for the explanation. Very helpful! :slight_smile:

t