Expert Judgment on the Mortality Impact of Ambient Fine Particulate Matter in the U.S.

ISEE-570 Objective: The US Environmental Protection Agency is required to quantify the health and economic benefits of large-scale proposed air quality regulations. Appropriate characterization of uncertainty in these analyses is critical because mortality-related benefits represent the vast majority of the monetized benefits reported in EPA's analyses of proposed air regulations. Benefit calculations based on published risk estimates from individual epidemiological studies and their reported standard errors may not capture the full impact of the uncertainties associated with these estimates. Our objective is to present the findings from a multiyear expert judgment study that comprehensively characterized uncertainties in estimates of mortality reductions associated with decreases in fine particulate matter (PM2.5) in the United States. Materials and Methods: A structured protocol was developed to elicit qualitative and quantitative judgments regarding the influence of uncertainties regarding confounding, effect modification, exposure misclassification, thresholds, and other issues on the quantification of mortality benefits. Experts were identified and selected based on peer nominations. The protocol was used to conduct structured interviews of 12 leading experts, each of whom developed a probabilistic uncertainty distribution that reflected his own judgments regarding the quantitative influence of the broad array of uncertainties in the concentration-response relationship. Results: The resulting distributions suggested both potentially larger central estimates of mortality reductions for decreases in long-term PM2.5 exposure in the United States and a wider distribution of uncertainty than has typically been employed in health benefits analyses performed by US EPA. Conclusions: Expert judgment elicited through a structured interview process can provide a more complete accounting of uncertainties and/or biases in estimates of key health-related inputs to regulatory benefits models.