Particulate air pollution, social confounders, and mortality in small areas of an industrial city

Scientists and policymakers have shown growing interest in the health effects of chronic air pollution exposure. In this study, we use geostatistical techniques in combination with small-area data to address a central research question: "Does chronic exposure to particulate air pollution significantly associate with mortality when the effects of other social, demographic, and lifestyle confounders are taken into account?" Our analysis relies on age-standardized mortality ratios for census tracts (CTs) of Hamilton (average population of 3419 persons), social and demographic data from the 1991 Census of Canada, smoking variables extracted from secondary surveys, and total suspended particulate (TSP) data from 23 monitoring stations operated by the Ministry of the Environment. Air pollution data are interpolated with a geostatistical procedure known as "kriging". This method translates fixed-site pollution monitoring observations into a continuous surface, which was overlaid onto the population-weighted centroids of the CTs. Our results show substantively large and statistically significant health effects for women and men. Evaluated over the inter-quartile range of the data, we found the relative risk of premature mortality for TSP exposure to be 1.19 (95% CI: 1.13-1.26) for women and 1.30 (95% CI: 1.24-1.37) for men. We also tested associations with cardio-respiratory and cancer mortality. We found positive, significant associations between particulate exposure and these causes of death in most models. Inclusion of socioeconomic, demographic, and lifestyle reduced but did not eliminate the health effects of exposure to particulate air pollution. Overall our results suggest that intra-urban variations in particulate air pollution significantly associate with premature, all-cause, cardio-respiratory, and cancer mortality in small areas of Hamilton.