Cohort
The COHORT table contains records of subjects that satisfy a given set of criteria for a duration of time. The definition of the cohort is contained within the COHORT_DEFINITION table. Cohorts can be constructed of patients (Persons), Providers or Visits.
More details about cohorts in The Book of OHDSI: https://ohdsi.github.io/TheBookOfOhdsi/Cohorts.html#Cohorts
Note: the structure of this table is missing in the reference scripts https://github.com/OHDSI/CommonDataModel/tree/v5.3.1_fixes/PostgreSQL
Field | Required | Type | FK Table | Description |
---|---|---|---|---|
cohort_definition_id | Yes | bigint | COHORT_DEFINITION | A foreign key to the record containing relevant Cohort Definition information. |
subject_id | Yes | bigint | Any | A foreign key to the subject in the Cohort. For e.g., records from Person, Provider, Visit Occurrence, Visit Detail, or Care Site. |
cohort_start_date | Yes | date | The date when the Cohort Definition criteria for the Subject first match. | |
cohort_end_date | Yes | date | The date when the Cohort Definition criteria for the Subject no longer match or the Cohort membership was terminated. |
Conventions
- The core of a Cohort is the unifying definition or feature of the Cohort.
This is captured in the
cohort_definition_id
. For example, Cohorts can include patients diagnosed with a specific condition, patients exposed to a particular drug, or Providers who have performed a specific Procedure. - Cohort records must have a Start Date.
- Cohort records must have an End Date, but may be set to Start Date or could have applied a censored date using the Observation Period Start Date.
- Cohort records must contain a Subject Id, which can refer to any table. The
Cohort Definition will define the type of Subject through the
subject_concept_id
. Common domains for Subject are the tables PERSON, PROVIDER, VISIT_OCCURRENCE, VISIT_DETAIL, and CARE_SITE. - A Subject can belong (or not belong) to a Cohort at any moment in time.
- A Subject can only have one record in the COHORT table for any moment of time, i.e. it is not possible for a Person to contain multiple records indicating cohort membership that are overlapping in time.