/tmleplus

R functions to calculate effects from tmle, ltmle, survtmle, and SuperLearner fit objects.

Primary LanguageR

tmleplus

A collection of functions to calculate additional effects and standard errors from from tlverse's survtmle objects.

Installation:

devtools::install_github("hoffmakl/tmleplus")

Current capabilities:

surv_tidy()

This function takes a fit object from the survtmle package and uses the influence curves to return a tidied data frame row with the difference, standard error, 95% confidence intervals, and p-value.

surv_contrast()

This function takes two fit objects from the survtmle package and uses the influence curves to return a tidied data frame row with the difference, standard error, 95% confidence intervals, and p-value for one survtmle exposure estimate in comparison to a reference exposure estimate. This should be used when the exposure of interest is multinomial.

surv_eff_mod()

This function takes two fit objects from the survtmle package and uses the influence curves to calculate the difference and standard errors between two marginal cumulative incidences, stratified by an effect modification variable.

surv_hists()

This function takes a fit object from the survtmle package and returns four histograms to aid in model diagnostics: the predicted probability of treatment, predicted probabilities of censoring, predicted probabilities of the outcome, and influence curves.