Background: The pathogenesis of childhood asthma remains poorly understood, despite intense epidemiologic investigation of individual-level risk factors. Studies of individual characteristics do not explain the large variation in rates between communities, suggesting that factors in the social and biophysical environment contribute to asthma and wheeze. Application of powerful new theories and methods from spatial multilevel modeling will advance knowledge about the causes of asthma by examining individual and contextual risk factors simultaneously. Research Questions: Three research questions that remain largely unanswered in the childhood asthma literature and in our Children's Health Study (CHS) motivate this proposal: First, do social, economic, and demographic variables and indicators of stress confound or modify the relationship between air pollution and the incidence of wheeze or asthma? Second, at what individual or spatial levels (family, neighborhood, school, or community) do these potential confounders and effect modifiers operate? Third, what will exclusion of one or more of these variables or levels do to the size and significance of the air pollution-asthma association? Methods: The CHS is an ideal resource for examining social, spatial, and contextual risk factors for asthma at various levels. Extensive characterization of contextual social environments will be integrated with indicators of stress to test the statistical associations with incident wheeze, using a novel multilevel Cox regression model. Simulations will be conducted to assess the influence of excluding one or more levels of influence on the asthma-air pollution association. These methods will focus initially on incident wheeze with the expectation that similar methods will also apply to incident asthma when the data become available in the parent study. Expected Benefits: This program will introduce modern multilevel spatial modeling to the field of asthma research. A specific contribution to the parent study will be made through inclusion of socioeconomic and stress variables that may confound or modify the asthma-air pollution association. The enhanced profiling of family, neighborhood, school and community environments will contribute to understanding whether these risk factors directly induce asthma onset or interact to worsen the effects of air pollution. Our focus on confounding and effect modification by contextual variables could have widespread applicability to future asthma studies and more generally to other health outcomes with high potential for interaction between the social and environmental context.
Showing the most recent 10 out of 20 publications
