Causal inference packages in R for RWE and Observational data
A curated list of top Causal inference packages in R for Real World Evidence and Observational data.
ADataScience
9/2/20241 min read
In this resources you may find a curated list of R packages for RWE and Observational data. The main causal inference field covered :
Inverse probability weighting
Propensity scores
Matching implementation
Weighting / Matching diagnostics
Instrumental variables
Mendelian Randomization
Here is the curated list :
As you can see the list contains multiple types of packages for multiple approaches. From aproaches such as IPW, inverse probability weighting (ipw, causal weight and twang) to instrumental variable packages such as ivreg and ivpack. Packages mediation and mediator are great to complement with mediation analysis. Then we have a top level package for Mendelian Randomization (MendelianRandomization).
Cobalt is fantastic to generate balance tables / plots for covariates preprocessed through subclassification using propensity scores. PSAgraphics is fantastic for propensity based methods diagnostics and in combination with MatchIt a great complemnt.
CBPS and ebal are definitelly my go to's. CBPS - Covariate balancing propensity score is its main use case and its very advanced and validated in this field. For ebal - a state of the art implementation of entropy based balacing.
All in all, this list provides in my opinion the most advanced and very importantly most validated packages for R Causal inference in RWE / Observational data scenario.
By ADataScience