The sensitivity of Hemoccult testing (HO) in a population-based screening for colorectal cancer was evaluated. HO sensitivity estimates were calculated as the ratio of screen-detected HO-positives to total screen or interval-detected cancers. Sensitivity was estimated according to 1, 2, and 3-year rescreening intervals. Corresponding estimates are 69.4%, 61.8% and 57.7%, respectively. No significant correlation was observed between HO sensitivity and other variables, such as calendar period, tumor stage or site, or patient age or sex. HO sensitivity as estimated in the present experience is consistent with other reports of population-based screenings. Screening every year would achieve too limited an increase in sensitivity, compared to biennal screening, to be worth the difficulties of doubling organizational efforts and costs.