Urban sprawl and cancer mortality in the United States

Background: Urban sprawl is a known correlate of obesity and has the potential to influence cancer mortality via direct and indirect effects on obesity, access to health services, physical activity, and transportation choices. Purpose: To examine the cross-sectional associations between cancer mortality, urban sprawl, obesity, and covariates at the county level in the United States. Methods: Data include SEER/NCI county (n = 936) cancer mortality estimates (2002-2006) for sixteen cancer sites, obesity prevalence (2007), urban sprawl index (2001), and additional covariates (2001-2006). The sprawl index was based on 22 variables grouped into four factors/dimensions of sprawl at the county level and ranged from 62 (most sprawling) to 478 (Least Sprawling). Results: Analyses of main effects indicated that cancer mortality was inversely associated with urban sprawl for liver and stomach and positively associated with sprawl for prostate, leukemia, non-Hodgkin's lymphoma, brain, melanoma, and lung, cancer in males. After adjustment for obesity, positive associations remained for brain, melanoma, lung, and liver cancer. Similar results were observed in females along with inverse associations between sprawl and cancer of the uterus. Analyses of two-way interactions suggest that some of these main effects are modified by pair-wise associations between race, availability of medical care and smoking characteristics. Conclusions: Ecological analyses of associations between urban sprawl and cancer mortality could help identify targets for public health interventions and potentially provide estimates of the potential magnitude of risk related to residence in urban vs. suburban areas.