Mark van der Laan Ph.D.
Turing Institute
Maya Petersen M.D. Ph.D. and Mark van der Laan Ph.D.
Mathematical Sciences Research Institute
Kara Rudolph, Ph.D., Nima Hejazi, Ph.D., Nick Williams, and Ivan Diaz, Ph.D.
Causal mediation analysis can provide a mechanistic understanding of how an exposure impacts an outcome. However, rapid methodologic developments and few formal courses present challenges to implementation. Beginning with an overview of classical direct and indirect effects, this workshop will present recent advances that overcome limitations of previous methods, allowing for: (i) continuous exposures, (ii) multiple, non-independent mediators, and (iii) effects identifiable in the presence of intermediate confounders affected by exposure. We will walk through how to choose the most appropriate mediation effect given one’s scientific question, including the assumptions necessary for a causal interpretation, and how to estimate the effect using flexible, robust estimators. We will emphasize answering substantive epidemiological questions from real-world studies. Multiply robust, nonparametric estimators of these causal effects, and free and open-source R packages (crumble) for their application, will be introduced. To ensure translation to real-world data analysis, this workshop will incorporate hands-on R programming to allow participants to practice implementing the statistical tools presented. It is recommended that participants have working knowledge of causal inference, including counterfactuals and identification (linking the causal effect to a parameter estimable from the observed data distribution). Familiarity with the R programming language is recommended.
Mark van der Laan, Ph.D.
This podcast dives into Targeted Learning as Professor Mark van der Laan shares his expertise in this episode of Causal Inference. Discover how his research is revolutionizing statistical methods to unravel causal relationships from complex, high-dimensional data across various study designs.
Nick Williams, Kara Rudolph, Ph.D., and Ivan Diaz, Ph.D.
We tend to be most familiar with estimating the effects of binary treatments or exposures. The classic average treatment effect (ATE), risk difference, risk ratio, and odds ratio are all examples of this. However, the exposure may be more complicated than a simple binary variable. For example, there may be multiple exposures or multiple components of an exposure, and it is most relevant to consider intervening on them jointly. In addition, sometimes exposures are continuous, and one would like to have an easy-to-interpret causal effect. In this workshop, we will walk through how to define causal effects (what are called causal estimands) for categorical, continuous, and multiple exposures. This includes: causal effects for: 1) static interventions (like the ATE), 2) dynamic treatment regimes, 3) modified treatment policies (interventions that depend on the natural value of treatment, e.g., shift interventions), and 4) stochastic interventions. We will introduce the lmtp R package for estimating these causal effects in both point-treatment and longitudinal studies. The estimators we discuss will include doubly robust, data-adaptive options. In summary, we will cover how to formulate and estimate a very general, flexible set of causal effects to best match one’s research question. This workshop incorporates hands-on R programming to allow participants to practice implementing the statistical tools presented. We will work through multiple real-world data examples together.