A Statistical Study of Effects of Air Pollution on Children's Health.

摘要

This research incorporates two approaches to the statistical analysis of ambient air pollution on children's health. The first approach is an analysis of health and covariate data from the National Health Interview Survey and pollution data from the Environmental Protection Agency. Extensive exploration of the relationships between the health outcomes asthma and respiratory allergies, covariates, and annual concentrations of TSP, SO2, NO2, CO, and O3 does not reveal a consistent connection between ambient air pollution concentrations and either health outcome. Limitations of the data are discussed. Due to the complex survey design of the National Health Interview Survey, design-based analysis is the gold standard in the cases when the entire data set is used, however in cases when the original data size is severely reduced, alternative methods of analysis may be preferred. In order to compare the results of these two approaches, several different procedures are explored and the estimates of effect size and standard error are compared. The second approach to analysis involves combining the results of published studies. There have been numerous studies seeking to establish an association between air pollution and children's adverse health outcomes, and the ultimate findings are often varied. The conflicting results of these studies lead naturally to a novel application of statistical meta-analysis whose primary objective is to integrate or synthesize the findings from independent and comparable studies. We conduct a meta-analysis focusing on the association between children's (binary) health outcomes (such as cough and respiratory symptoms) and four pollutants: PM10, NO2, SO2 , and O3. While we find a statistically significant association in the case of every pollutant, in the cases of PM10, NO2 , and SO2, there is heterogeneity among the estimated effect sizes. We have explored the techniques of meta-regression by incorporating distinct study features to meaningfully explain the heterogeneity.