A questionnaire to detect persons at high risk for colorectal cancer was used in conjunction with fecal occult blood tests in a 3-year screening program in Northern Israel. Screening was offered to 2,590 persons over 40 years of age and accepted by 1,797 (compliance of 69.4%). In the subsequent 2 years, occult blood testing (Hemoccult II) was offered to those who had had negative tests (compliance rate of 99.6% and 100%). Six hundred and thirty persons (35.1%) had risk factors according to the questionnaire, and 195 of them underwent colonoscopy, with a predictive value of 15.9% for a neoplastic lesion. The Hemoccult II test was positive in 71 participants (4.0%) of whom 67 were investigated with a similar predictive value for neoplastic lesion (16.4%). In the second and third annual screening, the fecal occult blood test was positive in 29 (2.6%) and 27 (2.5%), and had a two and three times higher predictive value for neoplastic lesions, respectively. This was accompanied by a decrease in the cost of discovery. In all three stages, an adenomatous polyp was found in 48, and cancer in 10 participants (2.6% and 0.5% of the 1,797 original participants). Although a questionnaire may be fruitful in colorectal cancer screening, the higher number of participants subjected to further examinations makes this approach very expensive. The annual stool examination for occult blood has a higher predictive value for colonic neoplasm and a lower cost than a one stage, broader population based, study.