Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
QUESTION: In emergency departments patients with an acute ankle sprain, are the Ottawa ankle rules accurate for excluding fractures of the ankle and midfoot?
Studies were identified by searching 6 databases, reviewing reference lists, and contacting authors.
Studies published in any language were selected if they evaluated the Ottawa ankle rules for diagnosing fractures of the ankle or midfoot and reported sufficient data to calculate the false positive and false negative rates.
Data were extracted on sample size, demographic characteristics of the patients, setting, type of Ottawa ankle rules used, study quality, and outcomes. Main outcomes included sensitivity, specificity, and negative likelihood ratios. After inspection of the receiver operating characteristics curve, sensitivities but not specificities were pooled using the bootstrap.
27 (15 581 patients) of 32 studies that met the selection criteria were included in the meta-analysis. 47 of 15 581 patients (0.3%) had a false negative result. The table shows the pooled sensitivities, distribution of specificities, pooled negative likelihood ratios, and the probability of a fracture after a negative test result (assuming a 15% prevalence of fractures) stratified by several characteristics.
In patients presenting to emergency departments with an acute ankle sprain, the Ottawa ankle rules are accurate for excluding fractures of the ankle and midfoot.
We know more about how well the Ottawa ankle rules rule out clinically important fractures of the ankle and midfoot than just about any other diagnostic test.1 Dozens of rigorous studies evaluating its performance, and now this systematic review by Bachmann et al, have been published. The rules are exceptionally simple, relying on just 2 observations: local tenderness and the ability to walk. They are designed to be very sensitive for use in ruling out clinically important injuries, and they do so with a high degree of accuracy. In 27 studies, only 47 of 15 581 patients (0.3%) had a false negative result. False negative rates differed little according to age of patient, prevalence of fracture, or methodological strength of the study.
However, the Ottawa ankle rules, like other fine clinical prediction rules, come up short in the real world of clinical decision making. They are not followed, even when actively promoted.2 Are there good reasons for this or are we clinicians just backsliders? As with any guideline, there may be the usual grousing about loss of clinical autonomy and “cookbook medicine.” But just behind the scene, and perhaps not mentioned because of political correctness, is bound to be socially instilled reluctance to miss clinically important findings (Alvan Feinstein’s “chagrin factor”) and fear of malpractice. The value judgment to follow a decision rule depends on the social context over and above raw, compelling facts. How small must the false negative rate be to be small enough? For reason to prevail, this question needs to be argued out in society and the courts and not just in doctors’ offices and emergency departments.
The Ottawa ankle rules are an extraordinarily sound basis for evaluation of ankle injuries, and the results should be taken seriously when deciding whether to recommend ankle radiography. But the rules should be treated as providing valuable information, not decisions. Methodologists may call them “decision rules,” but clinicians, insurers, and lawyers should treat them as guidelines and not be constrained by the results. Patient issues other than fracture/not fracture are important too, and social baggage surrounding clinical decision making cannot be discounted.
Source of funding: no external funding.
For correspondence: Dr L M Bachmann, Horten Centre, Zurich University, Zurich,