medoutcon: Nonparametric efficient causal mediation analysis with machine learning in R


The medoutcon R package provides facilities for efficient estimation of path-specific (in)direct effects that measure the impact of a treatment variable on an outcome variable, through a direct path and an indirect path (through a set of mediators only). In the presence of an intermediate mediator-outcome confounder, itself affected by the treatment, these correspond to the interventional (in)direct effects described by Dı́az et al. (2020), though similar (yet less general) effect definitions and/or estimation strategies have appeared in VanderWeele, Vansteelandt, and Robins (2014), Rudolph et al. (2017), Zheng and van der Laan (2017), and Benkeser and Ran (2021). When no intermediate confounders are present, these effect definitions simplify to the well-studied natural (in)direct effects, and our estimators are analogs of those formulated by Zheng and van der Laan (2012). Both an efficient one-step bias-corrected estimator with cross-fitting (Pfanzagl and Wefelmeyer 1985; Zheng and van der Laan 2011; Chernozhukov et al. 2018) and a cross-validated targeted minimum loss estimator (TMLE) (van der Laan and Rose 2011; Zheng and van der Laan 2011) are made available. medoutcon integrates with the sl3 R package (Coyle et al. 2021) to leverage statistical machine learning in the estimation procedure.

In Journal of Open Source Software
Nima Hejazi
Nima Hejazi
NSF Postdoctoral Research Fellow in Biostatistics

My research interests lie at the intersection of causal inference and machine learning, especially as applied to the statistical analysis of complex data from observational studies and experiments in the biomedical and health sciences.