The purpose of our study is to demonstrate the value of CT in the emergency department (ED) for patients with non-traumatic abdominal pain. Between August 1998 and April 1999, 536 consecutive patients with non-traumatic abdominal pain were entered into our study. Using a computer order entry system, physicians were asked to identify: (a) their most likely diagnosis; (b) their level of certainty in their diagnosis; (c) if they thought CT would be normal or abnormal; (d) their treatment plan (prior to knowledge of the CT results); and (e) their role in deciding to order CT. This information was correlated with each patient's post-CT diagnosis and subsequent management. Pre- and post-CT diagnoses were concordant in 200 of 536 (37%) patients. The physicians' certainty in the accuracy of their pre-CT diagnosis was less than high in 88% of patients. Prior to CT, the management plan included hospital admission for 402 patients. Following CT, only 312 patients were actually admitted; thus, the net impact of performing CT was to obviate the need for hospital admission in 90 of 536 (17%) of patients with abdominal pain. Prior to CT, 67 of 536 (13%) of all patients would have undergone immediate surgery; however, following CT only 25 (5%) actually required immediate surgery. Among patients with the four most common pre-CT diagnoses (appendicitis, abscess, diverticulitis, and urinary tract stones) CT had the greatest impact on hospital admission and surgical management for patients with suspected appendicitis. For patients with suspected appendicitis, CT reduced the hospital admission rate in 28% (26 of 91) of patients and changed the surgical management in 40% (39 of 91) of patients. Our study demonstrates the advantage of performing abdominal CT in the ED for patients with non-traumatic abdominal pain.