Assessment of sociodemographic and geographic disparities in cancer risk from air toxics in South Carolina.

ABSTRACT

Populations of color and low-income communities are often disproportionately burdened by exposures to various environmental contaminants, including air pollution. Some air pollutants have carcinogenic properties that are particularly problematic in South Carolina (SC), a state that consistently has high rates of cancer mortality for all sites. The purpose of this study was to assess cancer risk disparities in SC by linking risk estimates from the U.S. Environmental Protection Agency's 2005 National Air Toxics Assessment (NATA) with sociodemographic data from the 2000 US Census Bureau. Specifically, NATA risk data for varying risk categories were linked by tract ID and analyzed with sociodemographic variables from the 2000 census using R. The average change in cancer risk from all sources by sociodemographic variable was quantified using multiple linear regression models. Spatial methods were further employed using ArcGIS 10 to assess the distribution of all source risk and percent non-white at each census tract level. The relative risk (RR) estimates of the proportion of high cancer risk tracts (defined as the top 10% of cancer risk in SC) and their respective 95% confidence intervals (CIs) were calculated between the first and latter three quartiles defined by sociodemographic factors, while the variance in the percentage of high cancer risk between quartile groups was tested using Pearson's chi-square. The average total cancer risk for SC was 26.8 people/million (ppl/million). The risk from on-road sources was approximately 5.8 ppl/million, higher than the risk from major, area, and non-road sources (1.8, 2.6, and 1.3 ppl/million), respectively. Based on our findings, addressing on-road sources may decrease the disproportionate cancer risk burden among low-income populations and communities of color in SC.