The Application of the Benchmark Approach to the Study of the Joint Effect of Mixtures of Air Pollutants

• Activity Number: 383
• Type: Contributed
• Date/Time: Wednesday, August 1, 2007 : 8:30 AM to 10:20 AM
• Sponsor: Section on Risk Analysis
• Author(s): David Svendsgaard*+
• Companies: U.S. Environmental Protection Agency
• Keywords: benchmark risk assessment ; chemical mixtures ; air pollutants

Abstract

The aim is to estimate the joint effect of the mixture of five air pollutants (PM10 , O 3 , SO2 , NO2 and CO) for time series mortality/air pollution data sets of large cities in the U.S. In a single city, some of these pollutants can be highly correlated, making it difficult to separate the effects of individual chemicals. We estimate the expected the daily death rates due to meteorology and time components each city using the General Additive Model. Using the Benchmark approach in (Budtz-Jorgensen et al. 2001 Biometrics 57, 698-706), a measure of adjusted death rates is calculated, and the unusually high daily rates are identified. This measure is the dependent variable used for various nonlinear chemical mixture models (Greco et al.1988 Proceedings of the Biopharm. Sect. of ASA). The data are pooled to improve the separation of each pollutant's effects by reducing the colinearity.