PURPOSE: The purpose of this analysis is to describe factors associated with colorectal surveillance following diagnosis and treatment of nonmetastatic colorectal cancer. METHODS: Subjects were identified as part of the HMO Cancer Research Network's study of colorectal cancer survivors. To be eligible for the main study, patients had to be part of the staff model components of health maintenance organizations in southeastern Michigan and Minnesota. Using computerized databases, individuals were identified who were 40 years or older with incident nonmetastatic colorectal cancer diagnosed between January 1, 1990, and December 31, 2000. Using data current through 2002, we analyzed the cohort using chi-square test statistics, life tables, and Cox proportional hazards models to understand variations in posttreatment surveillance practices. Subjects were followed up from date of diagnosis to date of recurrence, death, disenrollment from the health plan, or loss to follow-up, which ever came first. We assessed factors associated with colorectal surveillance at 1, 3, and 5 years after treatment. We also included an analysis comparing those who received an exam and those who didn't regardless of exam timing. RESULTS: A total of 908 patients were eligible for the main study. Of these, we excluded subjects who were not white or African American (n = 27), resulting in an analytic sample of 881 (97% of the eligible cohort). Twenty-five percent of subjects were African American, 43% were female, and 48% were aged 70 years or older. The proportion who received an exam at 1 year was 18%, at 3 years was 60%, and at 5 years was 67%. Chi-square tests showed that African Americans were statistically significantly less likely than whites to receive an exam at all three time points. The Cox proportional hazards model for examinations regardless of timing and adjusted for confounders showed that African Americans were still less likely than whites to receive an exam (hazard ratio = 0.62; 95% confidence interval [CI] = 0.51 to 0.75). The same trend in undersurveillance was also observed for those 80 years of age or older at diagnosis, with an adjusted hazard ratio of 0.39 (95% CI = 0.26 to 0.57). CONCLUSION: Our data indicate that colorectal cancer survivors who are African American or aged 80 years or more at diagnosis are less likely to receive posttreatment colorectal surveillance. Whether these differences are due to system or patient level barriers needs further study.