Abstract:Large observational databases are often subject to missing data. As such, methods for causal inference must simultaneously handle confounding and missingness; surprisingly little work has been done at this intersection. Motivated by this, we propose an efficient and robust estimator of the causal average treatment effect from cohort studies when confounders are missing at random. The approach is based on a novel factorization of the likelihood that, unlike alternative methods, facilitates flexible modelling of nuisance functions (e.g., with state‐of‐the‐art machine learning methods) while maintaining nominal convergence rates of the final estimators. Simulated data, derived from an electronic health record‐based study of the long‐term effects of bariatric surgery on weight outcomes, verify the robustness properties of the proposed estimators in finite samples. Our approach may serve as a theoretical benchmark against which ad hoc methods may be assessed. Résumé Les grandes bases de données observationnelles sont souvent confrontées au problème des données manquantes. De ce fait, les méthodes d'inférence causale doivent gérer à la fois les facteurs de confusion et les données manquantes, un domaine étonnamment peu exploré. Pour répondre à ce défi, les auteurs de cet article proposent un estimateur efficace et robuste de l'effet causal moyen du traitement. Cet estimateur est spécialement conçu pour les études de cohortes où les facteurs de confusion sont manquants de manière aléatoire. L'approche proposée s'appuie sur une nouvelle factorisation de la vraisemblance. Contrairement à d'autres méthodes, celle‐ci permet de modéliser les fonctions de nuisance de manière flexible, notamment grâce à des techniques avancées d'apprentissage automatique, et ce tout en maintenant les taux de convergence attendus des estimateurs finaux. Des simulations basées sur des dossiers médicaux électroniques concernant les effets à long terme de la chirurgie bariatrique sur le poids confirment la robustesse des estimateurs dans des échantillons finis. Cette approche pourrait servir de référence théorique pour évaluer d'autres méthodes ad‐hoc.

Nonparametric Estimation of Path-specific Effects in Presence of Nonignorable Missing Covariates

On semiparametric estimation of a path-specific effect in the presence of mediator-outcome confounding

Semiparametric Inference for Estimating Equations with Nonignorably Missing Covariates

Identifiability and Estimation of Causal Mediation Effects with Missing Data

A Novel Path-Specific Effect Statistic for Identifying the Differential Specific Paths in Systems Epidemiology.

Nonparametric estimators of interventional (transported) direct and indirect effects that accommodate multiple mediators and multiple intermediate confounders

Identification, Semiparametric Efficiency, and Quadruply Robust Estimation in Mediation Analysis with Treatment-Induced Confounding

Post-Treatment Confounding in Causal Mediation Studies: A Cutting-Edge Problem and A Novel Solution via Sensitivity Analysis

Nonparametric Estimation of Mediation Effects with A General Treatment

A Bayesian nonparametric approach for causal mediation with a post-treatment confounder

Semiparametric Estimation for the Non-Mixture Cure Model in Case-Cohort and Nested Case-Control Studies

Bayesian Sparse Mediation Analysis with Targeted Penalization of Natural Indirect Effects

Semiparametric Estimation of Treatment Effects in Observational Studies with Heterogeneous Partial Interference

Efficient Algorithms for Covariate Analysis with Dynamic Data Using Nonlinear Mixed-Effects Model.

Non-linear Mediation Analysis with High-dimensional Mediators whose Causal Structure is Unknown

Mediation pathway selection with unmeasured mediator-outcome confounding

A Quick and Accurate Method for the Estimation of Covariate Effects Based on Empirical Bayes Estimates in Mixed-Effects Modeling: Correction of Bias Due to Shrinkage.

Robust causal inference for point exposures with missing confounders

Robust inference for causal mediation analysis of recurrent event data

A Semiparametric Approach to Model Effect Modification