Patients with ulcerative colitis have a high risk of colon cancer that increases with duration of disease. Annual colonoscopy with biopsies for 99 patients with pancolitis and a mean duration of disease at entry of 17 years was evaluated. The outcomes of mucosal dysplasia or cancer were analyzed to determine the annual risk (hazard rate) of high-grade dysplasia or colon cancer. The hazard rates were used to recommend screening intervals. The delay in the diagnosis of cancer is minimized for any given number of tests when the time between tests is inversely proportional to the square root of the hazard rate. Since the hazard rate for cancer increases with duration of disease, intervals for screening tests should not be uniform. For a range of reasonable marginal benefits from additional testing, the number of tests recommended is less than the number of tests done in fixed-interval screening programs. Scheduling screening tests using patient-specific hazard rates to minimize the delay in the diagnosis of cancer reduces the number of colonoscopies, associated cost, and morbidity.