Estimating a Causal Exposure-Response Function with a Continuous Error-Prone Exposure: A Study of Fine Particulate Matter and All-Cause Mortality

• Activity Number: 133 - Statistical Issues in Environmental Epidemiology and Pharmacoepidemiology
• Type: Contributed
• Date/Time: Monday, August 9, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Kevin Patrick Josey* and Priyanka deSouza and Xiao Wu and Rachel Nethery and Danielle Braun and Francesca Dominici
• Companies: Department of Biostatistics, Harvard TH Chan School of Public Health and Department of Urban Studies and Planning, Massachusetts Institute of Technology and Department of Biostatistics, Harvard TH Chan School of Public Health and Department of Biostatistics, Harvard T.H. Chan School of Public Health and Harvard T.H. Chan School of Public Health, Dana-Farber Cancer Institute and Harvard University
• Keywords: Air Pollution; Environmental Epidemiology; Causal Inference; Measurement Error; Bayesian Inference; Exposure-Response

Abstract

Numerous studies have examined the associations between long-term exposure to fine particulate matter (PM2.5) and adverse health outcomes. However, most studies employ predicted PM2.5 concentrations, which are subject to measurement error. Previous approaches for exposure measurement error correction have either been applied in non-causal settings or have focused on a categorical exposure only. Moreover, most have failed to account for uncertainty induced by error correction procedures when fitting an exposure-response function (ERF). To remedy these deficiencies, we develop a Bayesian framework that combines regression calibration techniques with estimation of a causal ERF. We demonstrate how the different components produced by the measurement error correction steps can be seamlessly integrated into the causal ERF estimator. This approach also properly propagates the exposure measurement error uncertainty to yield accurate standard error estimates. We assess the robustness of our proposed approach in an extensive simulation study. We then apply our methodology to estimate the effects of PM2.5 on all-cause mortality among Medicare enrollees in the northeastern United States.