A prospective study was performed to evaluate the effect of using a special requisition form on the utilization of lumbosacral spine radiography for patients presenting in the emergency room with acute lower back complaints. Over a 1-year period, emergency room house officers were encouraged to complete a special form that listed only three acceptable indications for lumbosacral spine radiographs: history of trauma, evidence of focal neurologic abnormality, and "other." Neurologic abnormalities included hypesthesia, hyperesthesia or anesthesia of lumbar or sacral dermatomes, weakness or hyporeflexia of the lower extremities, and bladder or bowel incontinence. If the indication "other" was chosen, radiographs were done only if the form contained both a one- to two-sentence history and the signed approval of a supervising attending physician. The number and results of lumbosacral spine examinations were compared with those from the previous year, which served as a control. In the control year 1443 examinations were performed, and in the experimental year only 759 were done (a decrease of 47%). The percentage of patients with vertebral fractures increased from 5.1% to 5.8% in the experimental year if only the detection of new fractures was considered positive, and from 9.1% to 13.4% if the detection of fractures of all ages was regarded as significant. The use of the special requisition form appears to be a simple and effective means of reducing unnecessary lumbosacral radiography in the emergency room setting.