Combining Cohorts

library(omock)
library(CohortConstructor)
library(CohortCharacteristics)
library(ggplot2)

For this example we’ll use the Eunomia synthetic data from the omock package.

cdm <- mockCdmFromDataset(datasetName = "GiBleed", source = "duckdb")

Let’s start by creating two drug cohorts, one for users of diclofenac and another for users of acetaminophen.

cdm$medications <- conceptCohort(cdm = cdm, 
                                 conceptSet = list("diclofenac" = 1124300,
                                                   "acetaminophen" = 1127433), 
                                 name = "medications")
cohortCount(cdm$medications)

To check whether there is an overlap between records in both cohorts using the function intersectCohorts().

cdm$medintersect <- intersectCohorts(
  cohort = cdm$medications,
  name = "medintersect"
)

cohortCount(cdm$medintersect)

There are 6 individuals who had overlapping records in the diclofenac and acetaminophen cohorts.

We can choose the number of days between cohort entries using the gap argument.

cdm$medintersect <- intersectCohorts(
  cohort = cdm$medications,
  gap = 365,
  name = "medintersect"
)

cohortCount(cdm$medintersect)

There are 94 individuals who had overlapping records (within 365 days) in the diclofenac and acetaminophen cohorts.

We can also combine different cohorts using the function unionCohorts().

cdm$medunion <- unionCohorts(
  cohort = cdm$medications,
  name = "medunion"
)

cohortCount(cdm$medunion)

We have now created a new cohort which includes individuals in either the diclofenac cohort or the acetaminophen cohort.

You can keep the original cohorts in the new table if you use the argument keepOriginalCohorts = TRUE.

cdm$medunion <- unionCohorts(
  cohort = cdm$medications,
  name = "medunion",
  keepOriginalCohorts = TRUE
)

cohortCount(cdm$medunion)

You can also choose the number of days between two subsequent cohort entries to be merged using the gap argument.

cdm$medunion <- unionCohorts(
  cohort = cdm$medications,
  name = "medunion",
  gap = 365,
  keepOriginalCohorts = TRUE
)

cohortCount(cdm$medunion)