A multicentre cross-sectional study to examine physicians' ability to rule out a distal radius fracture based on clinical findings

To study current use of radiography in patients with wrist trauma and examine physicians' ability to rule out a distal radius fracture based on their physical findings. We performed a multicentre cross-sectional observational study in five Emergency Departments (ED) between November 2010 and June 2014 and included all consecutive adult patients with wrist trauma. Physicians were asked to perform a standardized examination of the wrist and to subsequently indicate the probability of a distal radius fracture. The majority of the 924 included patients were referred for radiography (99.6 %). Of the 920 patients that were imaged, 402 (44 %) had sustained a distal radius fracture, 82 (9 %) an isolated carpal fracture and 12 (1 %) an isolated ulna fracture. Overall, physicians were able to accurately discriminate between patients with and without a distal radius fracture (area under the receiver operating characteristics curve: 0.87, 95 % CI 0.85-0.89). Physicians were absolutely certain of their clinical diagnosis in 180 patients (19 %), for whom they indicated either a 0 % or a 100 % probability. In these patients, physicians showed a 99 % sensitivity (95 % CI 98-100) and 67 % specificity (95 % CI 53-80) for predicting a distal radius fracture. Although physicians in the ED are able to accurately discriminate between patients with and without a distal radius fracture based on their physical findings, they were only completely certain of their diagnosis in 19 % of the patients. A validated clinical decision rule could reinforce physician's clinical judgment and support them in their decision not to routinely request radiography