Longitudinal, population-based study of racial/ethnic differences in colorectal cancer survival: impact of neighborhood socioeconomic status, treatment and comorbidity.

BACKGROUND: Colorectal cancer, if detected early, has greater than 90% 5-year survival. However, survival has been shown to vary across racial/ethnic groups in the United States, despite the availability of early detection methods. METHODS: This study evaluated the joint effects of sociodemographic factors, tumor characteristics, census-based socioeconomic status (SES), treatment, and comorbidities on survival after colorectal cancer among and within racial/ethnic groups, using the SEER-Medicare database for patients diagnosed in 1992-1996, and followed through 1999. RESULTS: Unadjusted colorectal cancer-specific mortality rates were higher among Blacks and Hispanic males than whites (relative rates (95% confidence intervals) = 1.34 (1.26-1.42) and 1.16 (1.04-1.29), respectively), and lower among Japanese (0.78 (0.70-0.88)). These patterns were evident for all-cause mortality, although the magnitude of the disparity was larger for colorectal cancer mortality. Adjustment for stage accounted for the higher rate among Hispanic males and most of the lower rate among Japanese. Among Blacks, stage and SES accounted for about half of the higher rate relative to Whites, and within stage III colon and stages II/III rectal cancer, SES completely accounted for the small differentials in survival between Blacks and Whites. Comorbidity did not appear to explain the Black-White differentials in colorectal-specific nor all-cause mortality, beyond stage, and treatment (surgery, radiation, chemotherapy) explained a very small proportion of the Black-White difference. The fully-adjusted relative mortality rates comparing Blacks to Whites was 1.14 (1.09-1.20) for all-cause mortality and 1.21 (1.14-1.29) for colorectal cancer specific mortality. The sociodemographic, tumor, and treatment characteristics also had different impacts on mortality within racial/ethnic groups. CONCLUSION: In this comprehensive analysis, race/ethnic-specific models revealed differential effects of covariates on survival after colorectal cancer within each group, suggesting that different strategies may be necessary to improve survival in each group. Among Blacks, half of the differential in survival after colorectal cancer was primarily attributable to stage and SES, but differences in survival between Blacks and Whites remain unexplained with the data available in this comprehensive, population-based, analysis.