Associations between measures of socioeconomic position and chronic nitrogen dioxide exposure in Worcester,...

Census block-group-specific predicted outdoor nitrogen dioxide (NO(2); a marker of traffic pollution) levels and four census block group socioeconomic position (SEP) measures were used to evaluate whether chronic exposures to traffic-related air pollutants are higher in areas with lower SEP, after controlling for spatial autocorrelation in mixed models. NO(2) levels were predicted using a geographic information system (GIS)-based spatiotemporal model that was validated with measured NO(2) concentrations. The GIS-based model predicted weekly NO(2) concentrations with high accuracy (slope of 0.98 from regression of held-out observations on predictions) and precision (cross-validation mean absolute error of 2.2 ppb). The model performed well in both rural and urban areas and warm and cold seasons. Estimated mean block group NO(2) concentrations were significantly negatively associated with median household income, and positively associated with poverty, crowding, and low educational attainment rates after controlling for spatial autocorrelation. Results indicated that a standard deviation (3.5 ppb) increase in block group NO(2) concentrations was associated with a $9090 decrease in median household income. Results suggest that on average those with lower SEP experience higher chronic exposure to outdoor NO(2).