Preventive programs for colorectal cancer (CRC) should ideally start w
ith an assessment of risk of the subject. Choice of screening procedur
es should be personalized for those of above average risk, i.e. those
with a familial incidence of CRC, personal history of colorectal neopl
asia or inflammatory bowel disease. For average-risk subjects, case-co
ntrol studies indicate that screening by sigmoidoscopy seems capable o
f reducing risk of death from cancer within reach by about 60%. The fi
rst reported randomized controlled study of occult blood testing has s
hown a 33% reduction in mortality from CRC (all sites). Provided that
additional controlled, population-based trials nearing completion also
show a significant reduction in CRC mortality, then large-scale commu
nity-based screening programs for average-risk subjects will be justif
iable. The value of colonoscopic screening at a community level is unc
lear; it seems likely that compliance will be very low and cost effect
iveness unsatisfactory. No controlled trials using colonoscopy as the
primary screening method have been conducted. Those with no obvious ri
sk factors apart from age over 50 years may reasonably be screened by
annual faecal occult blood testing (preferably with new, more sensitiv
e tests) and 5-yearly sigmoidoscopy (flexible if possible).