Geographic Variations of Colorectal and Breast Cancer Late-Stage Diagnosis and the Effects of Neighborhood-Level Factors

Purpose: The purpose of this study was to examine the geographic variations of late-stage diagnosis in colorectal cancer (CRC) and breast cancer as well as to investigate the effects of 3 neighborhood-level factors—socioeconomic deprivation, urban/rural residence, and spatial accessibility to health care—on the late-stage risks. Methods: This study used population-based South Dakota cancer registry data from 2001 to 2012. A total of 4,878 CRC cases and 6,418 breast cancer cases were included in the analyses. Two-level logistic regression models were used to analyze the risk of late-stage CRC and breast cancer. Findings: For CRC, there was a small geographic variation across census tracts in late-stage diagnosis, and residing in isolated small rural areas was significantly associated with late-stage risk. However, this association became nonsignificant after adjusting for census-tract level socioeconomic deprivation. Socioeconomic deprivation was an independent predictor of CRC late-stage risk, and it explained the elevated risk among American Indians. No relationship was found between spatial accessibility and CRC late-stage risk. For breast cancer, no geographic variation in the late-stage diagnosis was observed across census tracts, and none of the 3 neighborhood-level factors was significantly associated with late-stage risk. Conclusions: Results suggested that socioeconomic deprivation, rather than spatial accessibility, contributed to CRC late-stage risks in South Dakota as a rural state. CRC intervention programs could be developed to target isolated small rural areas, socioeconomically disadvantaged areas, as well as American Indians residing in these areas.