We reviewed the indications for and results of 788 consecutive upper gastrointestinal radiographs (UGIs) performed for ambulatory patients. Sixty-three percent of tests were ordered for the evaluation of abdominal pain, dyspepsia, or esophageal reflux. Of these tests, only 4.8% yielded results of major clinical importance to patient management. The yield for patients greater than 50 years of age was greater than for patients less than 50, 6.9 versus 3.0% (p = 0.04). There was a significant increase in yield with increasing age (chi trend = 11.6, p less than 0.001). Among patients with an indication of esophageal reflux alone (n = 62), there were no patients younger than age 60 with a test result that would significantly affect therapy or outcome. Among patients evaluated for fecal occult blood or weight loss (n = 120), 11.7% of tests ordered showed a finding of major clinical importance. In this group, the yield was higher in those greater than or equal to 50 years of age than in those less than 50, 14.7 versus 6.7%, (p = 0.2). These results indicate that UGIs ordered to evaluate pain or symptoms of esophageal reflux in the absence of bleeding or weight loss rarely yield results that significantly influence therapy. Such patients may be best served by an initial trial of empiric therapy or some other test. The UGI has greatest value when indications for it include bleeding or weight loss.