Use of radiography in acute ankle injuries: physicians' attitudes and practice.

OBJECTIVES

To examine the efficiency of the current use of radiography in patients with acute ankle injury. To study the judgements and attitudes of experienced clinicians in their use of ankle radiography and to thereby assess the potential for improved efficiency.

DESIGN

Two-stage study: retrospective chart review and prospective survey.

SETTING

Emergency departments of two adult teaching hospitals and one community hospital.

PARTICIPANTS

The records of 1831 adults presenting with acute blunt trauma to the ankle over 5 months were examined; another 732 patients were seen by 21 full-time emergency staff physicians over a subsequent 6-month period.

MEASURES AND MAIN RESULTS

Of the 1831 patients with an ankle injury in stage 1, 94.9% had had at least one radiographic series; the yield for clinically important fractures was 12.8%. In stage 2, experienced physicians predicted the probability of fracture to be 0% or 10% in 57.8% of cases. The kappa (kappa) level for interobserver agreement in 98 patients seen independently by two physicians was 0.55 (95% confidence interval [CI] 0.39 to 0.72). The area under the receiver operating characteristic curve for physicians' predicted probability was 0.88 (95% CI 0.84 to 0.92), reflecting good discrimination between fracture and nonfracture cases. Likelihood ratios for predicted probabilities ranged from 0.08 for the 0% level to 151 for the 100% level. The physicians indicated that they would feel comfortable or very comfortable in not ordering radiography in 45.9% of cases (kappa level 0.52; 95% CI 0.34 to 0.70).

CONCLUSIONS

Emergency physicians order radiography for most patients with ankle injury even though they can accurately discriminate between fracture and nonfracture cases and clearly expect most of the radiographs to give normal results. These findings suggest great potential for a more efficient use of radiography in patients with ankle injury, possibly through the use of guidelines.