Cohort Reconciliation and Membership Overlap — cohort_overlap • omopduckdb

Compares participants included in Cohort A only, Cohort B only, or both, computing Jaccard similarity indices and event rate deltas.

Usage

cohort_overlap(
  con = NULL,
  cohort_a,
  cohort_b,
  by = "site",
  person_col = "subject_id",
  outcome_col = NULL
)

Arguments

con: Optional DuckDB DBI connection (if cohort_a/b are table names).
cohort_a: First cohort data frame or table name.
cohort_b: Second cohort data frame or table name.
by: Optional site column to stratify by (default 'site').
person_col: Person identifier column name (default 'subject_id').
outcome_col: Optional binary outcome column name to compare event rates.

Value

A data frame with columns site, n_cohort_a, n_cohort_b, both_count, a_only_count, b_only_count, jaccard_index, concordance_rate.