Consequences of kriging and land use regression for PM2.5 predictions in epidemiologic analyses: insights into spatial variability using high-resolution satellite data

Journal of Exposure Science and Environmental Epidemiology (2015) 25, 138144

& 2015 Nature America, Inc. All rights reserved 1559-0631/15

http://www.nature.com/jes

Web End =www.nature.com/jes

ORIGINAL ARTICLE

Consequences of kriging and land use regression for PM2.5 predictions in epidemiologic analyses: insights into spatial variability using high-resolution satellite data

Stacey E. Alexeeff1,2, Joel Schwartz3, Itai Kloog3,4, Alexandra Chudnovsky3, Petros Koutrakis3 and Brent A. Coull1

Many epidemiological studies use predicted air pollution exposures as surrogates for true air pollution levels. These predicted exposures contain exposure measurement error, yet simulation studies have typically found negligible bias in resulting health effect estimates. However, previous studies typically assumed a statistical spatial model for air pollution exposure, which may be oversimplied. We address this shortcoming by assuming a realistic, complex exposure surface derived from ne-scale(1 km 1 km) remote-sensing satellite data. Using simulation, we evaluate the accuracy of epidemiological health effect estimates

in linear and logistic regression when using spatial air pollution predictions from kriging and land use regression models. We examined chronic (long-term) and acute (short-term) exposure to air pollution. Results varied substantially across different scenarios. Exposure models with low out-of-sample R2 yielded severe biases in the health effect estimates of some models, ranging from 60% upward bias to 70% downward bias. One land use regression exposure model with 40.9 out-of-sample R2 yielded upward biases up to 13% for acute health effect estimates. Almost all models drastically underestimated the SEs. Land use regression models performed better in chronic effect simulations. These results can help researchers when interpreting health effect estimates in these types of studies.

Journal of Exposure Science and Environmental Epidemiology (2015) 25, 138144; doi:http://dx.doi.org/10.1038/jes.2014.40

Web End =10.1038/jes.2014.40 ; published online 4 June 2014

Keywords: air pollution; kriging; land use regression; measurement error; PM2.5; spatial models

INTRODUCTIONThere is strong epidemiological evidence that both short-term and long-term exposures to air pollution are related to cardiovascular morbidity and mortality.1 In particular, much of the air pollution research shows that exposure to ambient particulate matter (PM) with aerodynamic diameter r2.5 mg/m3 (PM2.5) is associated with many adverse cardiovascular outcomes. In addition, ambient levels of PM2.5 often vary within a given city or region, and trafc sources may contribute to this variation.2,3 However, levels of PM2.5 are typically measured only at a small number of...