Causal Inference with Missing Exposures and Missing Outcomes

TL;DR

This paper extends causal inference methods to handle missing data in exposures, outcomes, and baseline measures, using counterfactual strata effects and TMLE with Super Learner, demonstrated through a TB study in Uganda.

Contribution

It introduces a framework for causal effect estimation with missing exposures and baseline outcomes, expanding existing methods to more complex missing data scenarios.

Findings

Extended causal models for missing exposures and baseline outcomes.

Demonstrated identification results for complex missing data settings.

Applied TMLE with Super Learner to real-world public health data.

Abstract

Missing data are ubiquitous in public health research. When estimating causal effects, there are well-established methods to address bias to due missing outcomes. Commonly, causal estimands are defined under hypothetical interventions to "set" the exposure and to prevent missingness. We demonstrate how this framework can be extended to missing exposures. We further extend this framework to incorporate missingness on the baseline outcome, which induces missingness on the population of interest. To do so, we highlight the use of Counterfactual Strata Effects: causal estimands where the focus population is subject to missingness and/or impacted by the exposure. Our work is motivated by SEARCH-TB's investigation of the effect of alcohol consumption on the risk of incident tuberculosis (TB) infection in rural Uganda. This study posed several real-world challenges: confounding, missingness on the exposure (alcohol use), missingness on the baseline outcome (defining who was at-risk of TB and, thus, in the focus population), and missingness on the outcome at follow-up (capturing who acquired TB). We present a series of causal models and identification results to demonstrate the handling of missingness in these settings. We highlight the use of TMLE with Super Learner and the real-world consequences of our approach.