The authors report the results of a risk questionnaire (RQ) used in a population-based screening program for colorectal cancer. The positive predictive value (PPV) for cancer or adenoma was evaluated for the Hemoccult test (HO) and for all RQ items (symptoms, personal and familial risk) in 8,114 cases, by univariate and multivariate analysis. A significant correlation with the presence of cancer or adenoma was observed for HO-positive tests, whereas a significant correlation was absent for most RQ variables. The use of an RQ in screening practice is disregarded since it does not improve the rate of cancer detection. Moreover, the increase in the detection rate of HO-negative adenomas does not justify the high rate (0.18) of screening responders selected for endoscopic diagnostic workup, a figure which influences negatively the overall compliance to screening.