Purpose: The increasing demand for venous duplex scans despite the relative rarity of detecting acute deep venous thrombosis (DVT) prompted us to review our experience with this diagnostic method. Methods: We retrospectively analyzed the results and indications of 2993 lower extremity venous duplex scans performed between July 1, 1992, and June 30, 1994, at our institution. The indication for the study and the results were prospectively recorded in a computerized data bank. The indications for these studies were leg pain (34%), leg swelling (24%), surveillance for DVT in a patient at high risk (23%), searching for a source of pulmonary embolism (14%), follow-up of previously diagnosed DVT (3%), and other indications (i.e., varicose veins, venous ulcer, 2%). Results: Overall, 74.1% of all scans were completely normal, and only 13.1% detected acute proximal (popliteal vein or higher) DVT. Scans performed for surveillance (87.3% normal) or source of pulmonary embolism (79.6% normal) were significantly more likely to be normal than when performed for any other indication (p < 0.01). When leg edema or calf tenderness was present, the incidence of acute DVT was significantly greater for all indications (p < 0.0001). Conclusion: The high percentage of normal venous scans implies that this diagnostic method is being inappropriately used. In the current climate of cost containment our data suggest that indications for venous duplex scans must be better defined and that improved education for referring physicians is needed.