Aspirin may decrease colorectal cancer incidence, but its role as an adjunct to or substitute for screening has not been evaluated. To examine the potential cost-effectiveness of aspirin chemoprophylaxis in relation to screening. Markov model. Literature on colorectal cancer epidemiology, screening, costs, and aspirin chemoprevention (1980-1999). General U.S. population. 50 to 80 years of age. Third-party payer. Aspirin therapy in patients screened with sigmoidoscopy every 5 years and fecal occult blood testing every year (FS/FOBT) or colonoscopy every 10 years (COLO). Discounted cost per life-year gained. When a 30% reduction in colorectal cancer risk was assumed, aspirin increased costs and decreased life-years because of related complications as an adjunct to FS/FOBT and cost $149 161 per life-year gained as an adjunct to COLO. In patients already taking aspirin, screening with FS/FOBT or COLO cost less than $31 000 per life-year gained. Cost-effectiveness estimates depended highly on the magnitude of colorectal cancer risk reduction with aspirin, aspirin-related complication rates, and the screening adherence rate in the population. However, when the model's inputs were varied over wide ranges, aspirin chemoprophylaxis remained generally non-cost-effective for patients who adhere to screening. In patients undergoing colorectal cancer screening, aspirin use should not be based on potential chemoprevention. Aspirin chemoprophylaxis alone cannot be considered a substitute for colorectal cancer screening. Public policy should focus on improving screening adherence, even in patients who are already taking aspirin.