A probabilistic characterization of the relationship between fine particulate matter and mortality: elicitation of European experts

Abstract

In support of an assessment of the mortality impacts of the Kuwait Oil Fires we interviewed six European experts in epidemiology and toxicology using formal procedures for elicitation of expert judgment. While the primary focus of the elicitations was to characterize the public health impacts of the fires, the experts provided quantitative estimates of the mortality impacts of hypothetical changes in the levels of ambient fine particulate matter (PM2.5) in both the United States and Europe. Uncertainty was assessed by asking each expert to provide the 5th, 25th, 50th, 75th, and 90th percentiles of their subjective cumulative probability density function for each quantity of interest. The results suggest that many regulatory risk assessments underestimate the impacts of PM2.5 mortality; confirm that only a small fraction of the mortality impact occurs within the first few months after exposure; and indicate that it may be important to better address the differential toxicities of particles from various source classes. By providing quantitative estimates of the uncertainty in current estimates of PM2.5 mortality risks, the study facilitates structured analysis of the value of further research on PM2.5 and its impacts.