How Untested Modeling Assumptions Influence the U.S. EPA's Estimates of Population-Level Ozone Exposure Risk

Abstract: In recent reviews of the National Ambient Air Quality Standards (NAAQS) for ozone, the U.S. Environmental Protection Agency (U.S. EPA) has presented estimates of the health risks associated with ozone exposure. One way in which the U.S. EPA calculates population-level ozone risk estimates is through a simulation model that calculates ozone exposures and the resulting lung function decrements for a simulated population. This simulation model includes several random error terms to capture inter- and intra-individual variability in responsiveness to ozone exposure. In this manuscript we undertake a sensitivity analysis examining the influence of untested assumptions about these error terms. We show that ad hoc bounds imposed on the error terms and the frequency of redrawing the intra-individual error terms have a strong influence on the population-level ozone exposure risk reported by the U.S. EPA.