Optimal-Design Domain-Adaptation for Exposure Prediction in Two-Stage Epidemiological Studies

TL;DR

This paper introduces a novel method combining optimal design and domain adaptation to improve exposure effect estimation in two-stage epidemiological studies, reducing bias and increasing accuracy.

Contribution

It proposes a new estimator that leverages importance weighting and domain adaptation to enhance imputation and effect estimation in two-stage studies.

Findings

More accurate exposure effect estimates than current methods

Smaller PM effect estimates on hyperglycemia risk with tighter confidence intervals

Effective sharing of information improves health effect estimation

Abstract

In the first stage of a two-stage study, the researcher uses a statistical model to impute the unobserved exposures. In the second stage, imputed exposures serve as covariates in epidemiological models. Imputation error in the first stage operate as measurement errors in the second stage, and thus bias exposure effect estimates. This study aims to improve the estimation of exposure effects by sharing information between the first and second stage. At the heart of our estimator is the observation that not all second-stage observations are equally important to impute. We thus borrow ideas from the optimal-experimental-design theory, to identify individuals of higher importance. We then improve the imputation of these individuals using ideas from the machine-learning literature of domain-adaptation. Our simulations confirm that the exposure effect estimates are more accurate than the current best practice. An empirical demonstration yields smaller estimates of PM effect on hyperglycemia risk, with tighter confidence bands. Sharing information between environmental scientist and epidemiologist improves health effect estimates. Our estimator is a principled approach for harnessing this information exchange, and may be applied to any two stage study.