OHDSI Home | Forums | Wiki | Github

Outcomes count in Cohort Method Data

Researchers
#1

Hi all,

the cohortMethodData provides a number of outcomes (e.g. if use summary(cmData) command).

The number of outcomes provided by cohortMethodData is very different to the initial counts we have in outcome cohort. I assume that cohortMethodData includes only those outcomes that are presented in target or comparator cohorts. Is that correct or there might be some other reasons?

BR,

Yuriy

#2

That’s correct. The outcome count is after:

  • Restricting to people in the target and comparator cohort
    • After imposing additional CohortMethod criteria as specified by the user, such as imposing a washout window, removing subjects in both cohorts, removing people who have the outcome prior to the index date, etc.
    • After any matching / trimming
  • Restricting to the user-defined time-at-risk
  • If using a survival model: restricted to the first outcome per person

So the counts will likely be very different from the raw cohort counts.

t