gesttools

Provides a series of general purpose tools to perform g-estimation using the methods described in Sjolander and Vansteelandt (2016) doi:10.1515/em-2015-0005 and Dukes and Vansteelandt doi:10.1093/aje/kwx347. The package allows for g-estimation in a wide variety of circumstances, including an end of study or time-varying outcome, and an exposure that is a binary, continuous, or a categorical variable with three or more categories. The package also supports g-estimation with time-varying causal effects and effect modification by a confounding variable.