Introduction We found variability in ordering radiographs for patients with elbow injury in our Emergency Department (ED). Review of the literature at the outset of our study provided few data to guide practice. It seemed possible to use clinical data, such as elbow extension to guide decision making. Recent large scale studies have suggested full elbow extension to be a reasonable rule out for fracture, although less sensitive in children. We aim to derive a maximally sensitive decision rule for elbow x-ray in adults and children.

Method We prospectively recruited patients attending the ED with elbow injury. Practitioners were advised to treat patients according to their usual practice, including applying their current criteria for selection for x-ray. Clinical variables were recorded, including site of tenderness and range of motion. Those not x-rayed were followed up by structured telephone interview at 1 week, and invited to return for re-evaluation if they had ongoing symptoms.

Results 467 patients were recruited over 2 years. 50.5% were male, and 26.3% were children (<16 years). 424 (90.8%) were x-rayed. 156 investigations (37% of radiographs performed) demonstrated an abnormality. Of those not x-rayed, 28 were followed up by telephone and had no problems, 13 could not be contacted (no return to our ED within 3 months) and one returned early and was x-rayed (no fracture).

We found elbow extension to have a sensitivity of 83% (95% CI 77% to 88%) and specificity 54% (49% to 60%) for abnormal x-ray in our whole dataset (76% sensitivity in children and 88% in adults).

Discussion We found elbow extension to be relatively insensitive as a rule out for fracture following elbow injury. We plan to use other clinical variables to derive a maximally sensitive decision rule for elbow x-ray after injury from these data.