Methods for Spatially Varying Measurement Error in Air Pollution Epidemiology

• Activity Number: 250
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Stacey E. Alexeeff*+ and Alexandros Gryparis and Joel Schwartz and Brent A. Coull
• Companies: Harvard School of Public Health and Harvard School of Public Health and Harvard School of Public Health and Harvard School of Public Health
• Keywords: measurement error ; spatio-temporal ; air pollution ; land use regression ; environment ; epidemiology

Abstract

Land use regression models can improve exposure assessment for traffic-related air pollutants, especially compared to previous approaches using central monitoring sites. However, a health effect analysis that uses predicted values from an exposure model results in exposure misclassification because the predicted exposures are not the true exposures. Thus, the health model can be viewed as containing a covariate with measurement error. Previous work has explored methods that account for the measurement error induced by smoothing models in a purely spatial setting. We now examine the case where the exposure model fit may vary by location in spatio-temporal settings. We extend existing regression calibration correction techniques to account for the correlations of site-specific measurement error. We apply the proposed methods to an analysis of air pollution and birthweight in Boston.