Leveraging the Causal Effects of Stochastic Interventions to Evaluate Vaccine Efficacy in Two-phase Trials

Abstract

Causal inference has traditionally focused on the effects of static interventions, under which the magnitude of the treatment is set to a fixed, prespecified value for each unit. The evaluation of such interventions faces a host of issues, among them non-identification, violations of the assumption of positivity, and inefficiency. Stochastic interventions provide a promising solution to these fundamental issues by allowing for the target parameter to be defined as the mean counterfactual outcome under a hypothetically shifted version of the observed exposure distribution. Despite the promise of such approaches, real data analyses are often further complicated by economic constraints, such as when the primary variable of interest is far more expensive to collect than auxiliary covariates. Two-phase sampling schemes are often used to bypass such limitations – unfortunately, their use produces side effects that require further adjustment when formal statistical inference is the principal goal of a study. We present a novel approach for use in such settings: augmented targeted minimum loss and one-step estimators for the causal effects of stochastic interventions, with guarantees of consistency, efficiency, and multiple robustness even in the presence of two-phase sampling. We further propose a technique that utilizes the estimated causal effects of stochastic interventions to construct a nonparametric working marginal structural model to summarize the effect of shifting an exposure variable on the outcome of interest, analogous to a dose-response analysis. Using data from the recent HVTN 505 HIV vaccine efficacy trial, we demonstrate this technique by assessing the effects of changes in post-vaccination immunogenicity on HIV-1 acquisition across a range of possible shifts, outlining a highly interpretable variable importance measure for ranking multiple immune responses based on their utility as immunogenicity study endpoints in future HIV-1 vaccine trials. This is joint work with David Benkeser, Mark van der Laan, Peter Gilbert, and Holly Janes, recently published at https://doi.org/10.1111/biom.13375.

Date
Wed, Dec 16, 2020 11:45 AM
Event
Methods for the Thorny Challenges of Real Studies, Society for Epidemiologic Research Meeting
Location
Boston, Massachusetts, United States (remote due to COVID-19)
Nima Hejazi
Nima Hejazi
Assistant Professor of Biostatistics

My research lies at the intersection of causal inference and machine learning, developing flexible methodology for statistical inference tailored to modern experiments and observational studies in the biomedical and public health sciences.