Primary Submission Category: Applications in Physical Sciences, Engineering, Environment and Miscellaneous Applications
Stochastic interventions for studying the health effects of environmental mixtures
Authors: Zhuochao Huang, Antonelli Joseph,
Presenting Author: Zhuochao Huang*
Evaluating the causal health effects of multivariate continuous exposure
mixtures, such as air pollutants, is a critical public health challenge. A primary
obstacle is the frequent violation of the positivity assumption, which renders the
effects of standard deterministic interventions unidentified without unreliable
extrapolation. In this paper, we develop a novel causal inference framework to
address this challenge. We extend exponential tilting to multivariate exposures
and address the critical question of how to compare different intervention
directions fairly. This establishes a systematic framework for defining and
evaluating various policy-relevant causal estimands, allowing researchers to
address diverse scientific questions. We develop numerous methodological
advancements, including efficient one-step estimation strategies, a Riemannian
BFGS algorithm to solve a constrained manifold optimization problem,
semiparametric efficiency bounds for causal estimands, minimax rates for
estimators, and establishing asymptotic normality. We demonstrate our
framework’s utility by applying it to a large-scale, real-world environmental
health dataset to identify the optimal strategy for reducing adverse health
outcomes associated with a PM2.5 chemical mixture.