There is no national screening programme for colorectal cancer in the UK despite the fact that the annual death toll from this disease exceeds that of breast and cervical cancer. Faecal occult blood testing (FOBT) is under evaluation for screening, but screening by sigmoidoscopy is not considered viable. This situation contrasts with the USA where both annual FOBT and screening by flexible sigmoidoscopy every 3 to 5 years are recommended from 50 years old. We seek to demonstrate that most of the benefit from the US screening policy would accrue from a single flexible sigmoidoscopy examination at age 55 to 60 years with appropriate colonoscopic surveillance for the 3% to 5% found to have high-risk adenomas (> or = 1 cm or villous histology). If applied nationally, this screening regimen could prevent about 5500 colorectal cancer cases and 3500 deaths in the UK each year, thus saving 40,000 years of life. We estimate that there would be little net cost to the National Health Service because savings obtained from treating fewer patients would largely offset the cost of screening. We recommend that a randomised trial to evaluate screening by single flexible sigmoidoscopy should start without delay. Such a trial would involve about 120,000 participants, and 15 years of follow-up would be required to obtain a clear answer on mortality, although information on incidence reduction would be available sooner.