The rate of clinically significant plain radiograph misinterpretation by faculty in an emergency medicine residency program.

Abstract

To estimate the rate of clinically significant discrepancies between radiograph interpretations by attending radiologists and emergency medicine (EM) faculty in 2 academic EDs, using a unique scoring system. A retrospective comparison of radiographic agreement between EM and radiology faculty members was performed. All plain films initially interpreted by EM faculty or by EM residents with immediate reinterpretation by EM faculty were subsequently reviewed by attending radiologists. All discrepancies between these readings were reported to the ED on the following day for review by an EM faculty member (usually different from the initial EM faculty reader) who determined the need for treatment or follow-up changes. A secondary chart review by a quality assurance faculty member determined whether radiographic findings not noted on the x-ray log were present on the ED record All discrepancies from February to June 1994 were reviewed. A severity score was assigned based on the following criteria. Q-0: There was no change in treatment or follow-up; or the initial interpretation by EM faculty was validated by repeat or additional views. Q-1: Discrepancy is minor. Q-2: Discrepancy is significant, with potential for injury or bad outcome. Q-3: Discrepancy is significant, with actual injury or bad outcome. Of 14,046 radiographic studies eligible for enrollment, there were 134 discrepancies (0.95%). Only 28 cases (0.2%) were found to be clinically significant. Of these, 25 were scored Q-1, 3 were scored Q-2, and 0 were scored Q-3. These clinically significant discrepancy rates were highest for the finger, skull, elbow, hand, and lumbar spine. Emergency medicine faculty provide highly accurate rates of plain radiograph interpretation, particularly when adjusted for clinical significance and actual impact on patient care.