776 zyxwvutsrqpo ACADEMIC EMERGENCY MEDICINE AUG 1997 VOL zy 4/NO 8 Ankle Radiograph Utilization after Learning a Decision Rule: A 12-month Follow-up zyxw P. zyxwvutsrqpon Richard Verbeek, MD, Ian G. Stiell, MD. MSc, Guy Hebert, MD, Cathy Sellens, MD I ABSTRACT Objective: To test whether the reduction in ankle radiograph ordering was sustained during a 12-month period after a formal trial to introduce the Ottawa ankle rules. Methods: A before -after clinical trial of ankle radiograph ordering practice was performed in a university- based ED. All 1,884 (947 “during intervention,” 937 “postintervention”) adults seen with acute ankle injuries during 2 12-month trial periods were evaluated. The behavioral intervention was the teaching of the Ottawa ankle rules and feedback of compliance with the rules during the intervention period. No further education about the ankle rules or feedback regarding compliance occurred during the postintervention year. Physicians were unaware of any postintervention surveillance. The primary outcome was the proportion of eligible pa- tients referred for an ankle radiograph during the intervention and postintervention periods. Results: During the intervention period (January 1 -December 31, 1993), the proportion of patients who received an ankle radiograph [609 x-rayed of 947 patients seen (64.3%; 95% CI 61.2-67.4%)] did not differ from the proportion who received an x-ray in the postintervention period (January I-December 31, 1994) [583 x-rayed of 937 patients seen (62.2%; 95% CI 59.1-65.3%), p zyxw = 0.65, power > 0.80 to detect a 10% increase in the radiograph ordering rate]. There was also no difference in the radiograph ordering rate in the first 3 months of the postintervention period compared with the last 3 months of the postintervention period (68.8% vs 64.7%, respectively,, p > 0.30). Conclusions: Compliance with the Ottawa ankle rules was sustained during a 12-month postintervention surveillance period when physicians did not know they were being observed. Physicians will continue to use a simple clinical guideline once it has been learned. Key words: ankle injuries; radiography; emergency medicine; decision making; Ottawa ankle rules. Acad. Emerg. Med. 1997; 4:776-779. ....................................................................................................................................... I Ankle injuries are a common cause for ED visits. Al- though physicians traditionally order ankle radiographs for most of these patients,’-5 few patients will be shown to have a clinically significant Recently, the Ot- tawa ankle rules were developedg.10 and validated’’ as being highly sensitive in identifying clinically significant ankle fractures.” In a multicenter trial, the Ottawa ankle rules were introduced successfully to a wide variety of com- munity and teaching hospitals and resulted in an overjall 26.4% relative reduction in ankle radiograph ordering.” As ................................................................ From rhe University of Toronto. Sunnybrook Health Science Centre, Toronto, Ontario, Canada, Department of Emergency Services (PRV) and Emergency Medicine Residency Program zyxwvutsr (GH, CS); and the Uni- versity of Ottawa, Ottawa. Ontario, Canada, Division of Emergency Medicine (IGS) and Clinical Epidemiology Unit (IGS). Received: September IO. 1996; revision received: December zyxwvuts 10. 1996; accepted: December 22, 1996; updated: February 5, 1997. Prior presentation: SAEM annual meeting, Denvec CO. May I996. Address for correspondence and reprints: Dr. P. Richard Verbeek, BG- 15, Department of Emergency Services. 2075 Bayview Avenue, Sunny- brook Health Science Centre. Toronto, ON, M4N 3MS. Canada. Fax: 416-480-4704: e-mail: Kverbeek@‘utoronto.cn one of the participating hospitals, we achieved a 25.5% relative reduction in ankle film ordering. These rules have been accepted as an example of a rigorously derived, sim- ple, easily followed guideline that offers a high standard of comparison in the clinical guidelines m0~ernent.I~ However, publishing guidelines falls far short of the effort required to convince physicians to use them.” Sim- ply publishing guidelines may have little effect on phy- sician behavior even in circumstances where physicians are fully aware of them.I6 We sought to determine whether a group of physicians continued to show a reduction in ankle film ordering over a 12-month period following the end of the formal trial when no explicit monitoring of behavior occurred. I METHODS ...................................................................... Study Design: A before-after clinical trial of ankle ra- diograph ordering rates was performed in a university- based ED. The rate of ankle film ordering during a clinical trial to introduce the Ottawa ankle rules (before) was com- pared with the ordering rate during the year following completion of the trial (after). The study was approved by