Be notified of new releases
Create your free GitHub account today to subscribe to this repository for new releases and build software alongside 28 million developers.Sign up
In this release, the workflow has been redesigned. Most importantly, the function
createStudyPopulation has been introduced where most filtering of the study population will be performed. All filtering steps are now optional. The main reason for this is better connection to external cohort definition tools like Circe, which usually are used to perform most of the necessary filtering steps. Another reason is increased transparency. At each point in the analysis you can now call on these three functions:
getAttritionTablewill show how many subjects remain after the various filtering steps (starting with the count as found in the original cohorts in the database.
drawAttritionDiagramis similar to
getAttritionTable, but outputs a graph.
insertDbCohortswill insert the (filtered) cohorts back into the database, for example for use in external cohort characterization tools.
Some minor changes:
- Less use of ff to store data: simpler code
- Allow same person to appear in both cohorts
- Allow same person to have multiple index dates
- Can still filter on new use, washout period, and one cohort only during loading for efficiency (e.g. when using drug_era directly)
See the package manual and vignettes for full details on how to use this new version.
This new version is not backwards compatible. If you need to use the old version, you can call:
install.packages("devtools") library(devtools) install_github("ohdsi/CohortMethod", ref = "v1.1.4")