We evaluated a list of high-yield criteria (HYL) as a means of reducing posttraumatic skull roentgenograms and as a diagnostic test for skull fractures. Physicians requested roentgenograms for 76.9% of head trauma patients in a four-month control period, but for only 46.1% during a four-month experimental period, a relative decrease of 40%. When the experiment ended, roentgenogram use increased to 72% of head trauma patients. The HYL accurately classified 92.5% of the patients with and 84.5% of the patients without pathological conditions. The HYL displayed substantial differences in sensitivity and specificity from HYLs used previously. Clinicians disliked the HYL, believing it was inaccurate and a monitoring device. We conclude that criteria lists can effectively decrease requests for roentgenograms and should be carefully evaluated as diagnostic tests for accuracy and adverse outcomes.
(JAMA 244:673-676, 1980)