Non-Parametric Causal Effects of Feasible Interventions Based on
Modified Treatment Policies
Non-parametric estimators for casual effects based on longitudinal modified treatment
policies as described in Diaz, Williams, and Hoffman (), traditional point treatment,
and traditional longitudinal effects. Continuous, binary, and categorical treatments are allowed as well are
censored outcomes. The treatment mechanism is estimated via a density ratio classification procedure
irrespective of treatment variable type. For both continuous and binary outcomes, additive treatment effects
can be calculated and relative risks and odds ratios may be calculated for binary outcomes. Estimation is enhanced
using the Super Learner from 'sl3' available for download from GitHub using 'remotes::install_github("tlverse/[email protected]")'.