Clinical Decision Rules for Knee Radiographs
Clinical Decision Rules for Knee Radiographs
ABSTRACT & COMMENTARY
Synopsis: Each decision rule using historical and physical criteria to evaluate extremity injuries had 85% sensitivity and about 50% specificity. Because neither decision rule is 100% sensitive, the rules must be further refined.
Source: Richman PB, et al. J Emerg Med 1997;15:459-463.
Cost-effective ordering of radiographs in evaluating extremity injuries requires that physicians be able to use historical and physical criteria to assess the likelihood of a fracture. While no one expects these criteria to be perfectly accurate, much attention has been focused on which findings suggest a fracture strongly enough to warrant x-rays.
Stiell and colleagues met with great success in defining the "Ottawa ankle rules," which, if applied correctly, may save millions of dollars in unnecessary ankle x-rays.1
In this article, Richman and colleagues examine the performance of two published decision rules for limiting radiographs in patients with acute knee injuries. The Bauer rule suggests that all knee fractures would correctly be imaged if clinicians rely on the presence of effusion, ecchymosis, or inability to bear weight.2 The Stiell rule yields perfect sensitivity by limiting radiographs to patients with tenderness at the fibular head, isolated patellar tenderness, inability to flex 90°, inability to bear weight (both immediately and in the ED), or those patients at least 55 years of age.
To validate these rules, Richman et al collected data on 391 consecutive patients presenting with acute knee injuries. Elements of the history and examination were included that were not components of either decision rule. Radiographs were ordered at the discretion of the examining physician without specific application of either rule. Follow-up calls were made to patients not undergoing radiography at the time of the initial visit.
Of the 351 evaluable patients, 26 (7%) had clinically significant fractures. Each decision rule, had it been applied in determining the need for radiographs, would have missed four of these fractures. Thus, each rule had 85% sensitivity and about 50% specificity. Richman et al conclude that because neither decision rule is 100% sensitive, the rules must be further refined.
COMMENT BY DAVID J. KARRAS, MD
Stiell and Bauer each reported that their respective decision rules had perfect sensitivity.1,2 In fact, Stiell et al validated their rule in more than 1000 patients, which confirmed that the rules detected every patient with a fracture. Why then do Richman et al find the rules to be less than perfect? The most likely answer is that these physicians were not trained to apply any specific decision rule when determining the need for radiographs. While this is technically a weakness of the study, few practicing physicians are likely to receive such special training before using a decision rule. This limitation represents a plausible real-world scenario in which physicians have never closely examined the rules they attempt to apply. While perfect sensitivity in detecting fractures may be a necessary prerequisite for a prediction rule in an "in vitro" setting, it may be an unrealistic goal for the "in vivo" practice of medicine.
To me, 85% sensitivity is good. The less-than-perfect quality of the rules mandates that I continue to use some judgment and arrange for follow-up if I don't suspect a fracture. It also speaks against allowing insurers to adopt either decision rule and use it to deny payment for radiography in cases that do not meet specific criteria. (Dr. Karras is Associate Professor of Medicine, Temple University School of Medicine, and Director of Emergency Medicine Research, Temple University Hospital, Philadelphia, PA.)
References
1. Stiell IG, et al. Ann Emerg Med 1995;26:405-413.
2. Bauer SJ, et al. J Emerg Med 1995;13:611-615.
Subscribe Now for Access
You have reached your article limit for the month. We hope you found our articles both enjoyable and insightful. For information on new subscriptions, product trials, alternative billing arrangements or group and site discounts please call 800-688-2421. We look forward to having you as a long-term member of the Relias Media community.