A Bayesian nonparametric approach to mediation and spillover effects with multiple mediators in cluster-randomized trials

2024-11-05

AI Summary

In cluster-randomized trials (CRTs) with intra-cluster interference and multiple unstructured mediators, conventional causal mediation methods fail to disentangle individual-level direct and indirect effects from intra-cluster spillover mediation effects. Method: We formally define the “intra-cluster spillover mediation effect” and propose a Bayesian nonparametric framework based on nested dependent Dirichlet process mixtures (NDDP-Mixture), enabling flexible modeling of multilevel outcomes and mediator surfaces. Our approach integrates causal diagrams with theory for identification under multiple mediators and interference. Contribution/Results: Simulation studies demonstrate substantial gains in estimation accuracy and robustness over parametric Bayesian alternatives across diverse interference patterns and heterogeneity structures. Applied to a real CRT, our method successfully decomposes mediation pathways—including their spillover contributions—yielding interpretable, policy-relevant causal mechanisms. The framework provides a generalizable tool for causal mediation analysis in complex, interference-prone settings.

Technology Category

• Reasoning under Uncertainty: Causality
• Machine Learning: Causal Learning
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Social Networks and Social Media: Analysis of web-mediated communications, communities, and crowds

Abstract

Cluster randomized trials (CRTs) with multiple unstructured mediators present significant methodological challenges for causal inference due to within-cluster correlation, interference among units, and the complexity introduced by multiple mediators. Existing causal mediation methods often fall short in simultaneously addressing these complexities, particularly in disentangling mediator-specific effects under interference that are central to studying complex mechanisms. To address this gap, we propose new causal estimands for spillover mediation effects that differentiate the roles of each individual's own mediator and the spillover effects resulting from interactions among individuals within the same cluster. We establish identification results for each estimand and, to flexibly model the complex data structures inherent in CRTs, we develop a new Bayesian nonparametric prior -- the Nested Dependent Dirichlet Process Mixture -- designed for flexibly capture the outcome and mediator surfaces at different levels. We conduct extensive simulations across various scenarios to evaluate the frequentist performance of our methods, compare them with a Bayesian parametric counterpart and illustrate our new methods in an analysis of a completed CRT.

Problem

• Addressing causal inference challenges in cluster-randomized trials with multiple mediators
• Disentangling mediator-specific effects under interference within clusters
• Proposing Bayesian nonparametric methods for flexible modeling of complex data structures

Innovation

• Bayesian nonparametric approach for mediation effects
• Nested Dependent Dirichlet Process Mixture prior
• New causal estimands for spillover mediation