The use and quality of computed tomography (CT) in the evaluation of blunt abdominal trauma has been assessed by a retrospective review of 122 patients referred in the first 8 years following the installation of CT of the body. The referral patterns were analysed, and the quality of the images was assessed, with particular reference to unsatisfactory (unacceptable or poor) images. Referrals increased over the 8 years following installation. A large proportion of these examinations were performed outside normal working hours (58/122, 48%). In 64 out of 122 (52%) of examinations the images were considered of unsatisfactory quality with inadequate use of intravenous contrast medium cited as the most common cause for unsatisfactory examinations. Image quality did not improve with time. Image quality was unsatisfactory in a larger proportion of those examinations performed outside normal working hours (35/58, 60%) than those during the routine working day (29/64, 45%), although this difference did not reach statistical significance (0.1 greater than p greater than 0.05). Fewer unsatisfactory examinations were produced by a system dedicated to body CT (31/74, 42%) than by a neuroradiological system (33/48, 69%) (p less than 0.01). The proportion of unsatisfactory abdominal examinations was lower when the CT study concentrated on the abdomen (26/65, 40%) than when multiple regions (head, chest, abdomen, etc.) were examined (38/57, 67%) (p less than 0.01). This audit revealed many unsatisfactory examinations. In order to complete the audit loop we have addressed many of the contributory factors. A new CT system now provides images of high quality and is used for all abdominal work. Guidelines, which describe an optimized technique, have been issued to radiographers and radiologists. In particular, we have emphasized that there is no place for a poorly supervised "quick look" at the abdomen following a cranial or thoracic study.