Development of a Multi Pollutant Model to Assess Air Pollution Association with Human Health Effects

Abstract

- A number of methodologies have been developed for investigation of the association between human health and exposure to a single pollutant [e.g., 1]. However, as pollutants are correlated, and the joint effect of pollutants is of high interest, work continues on development for multiple pollutant models. In this work, we discuss a method using Thin Plate Splines (TPS) to simultaneously model both PM 2.5 (particulate matter less than 2.5 μm in aerodynamic diameter) and O 3 (ozone) in association with human mortality. The results are compared to effect estimates obtained from single pollutant models. We find similar temporal trends in the estimates, with large movements in both PM 2.5 and O 3 being captured in the TPS estimates. The estimated errors for the TPS method are larger than the individual models combined and produce risks that are comparable but slightly elevated.