BACKGROUND AND OBJECTIVE: Some studies suggest that immunochemical fecal occult blood tests (FOBTs) and HemoQuant are more efficient at detecting fecal occult blood than the commonly employed Hemoccult II (guaiac) test. We undertook this study to determine whether an immunochemical test either alone or in combination with a guaiac test gives efficiency superior to the Hemoccult II test in predicting significant gastrointestinal tract disease. DESIGN: Criterion standard, prospective, blinded. SETTING: Referral population of ambulatory patients at an institutional and a private hospital. PATIENTS: Eight-one patients referred to a gastroenterologist and in whom colonoscopy was indicated. INTERVENTIONS: While on a restricted diet, patients made preparations for FOBTs from three consecutive bowel movements. Patients then underwent colonoscopy. Polyps larger than 1 cm in size, carcinoma of the colon, peptic ulcers, gastric erosions, and angiodysplasia were considered to be likely causes of occult gastrointestinal tract bleeding. MAIN OUTCOME MEASURES: Using colonoscopy results as the reference standard, sensitivity, specificity, and positive and negative predictive values for each of eight tests or pair of tests were compared with those of Hemoccult II. RESULTS: Of 81 patients, 10 had significant lower gastrointestinal tract lesions and six had significant upper gastrointestinal tract lesions. Hemoccult SENSA, Heme-Select, and FECA-EIA were shown to be more sensitive than Hemoccult II but slightly less specific. Paired tests showed less efficiency than Hemoccult II alone. CONCLUSIONS: We did not find an ideal test or pair of tests; however, Hemoccult SENSA exhibited higher sensitivity than Hemoccult II and many other tests. The sensitivity, specificity, and positive predictive values of many of the FOBTs were believed to be low. We recommend that physicians consider FOBTs only as adjuncts to history and physical examination findings in deciding how to proceed in diagnosing gastrointestinal tract disease.