Emergency Medicine Resident Interpretation of Pediatric Radiographs

Objective: To examine the concordance of pediatric radiograph interpretation between emergency medicine residents (EMRs) and radiologists.
Methods: A prospective, observational study was performed in a university pediatric ED with an annual census of 60,000 visits. Radiographs ordered by EMRs from December 1993 through October 1994 were initially interpreted solely by the EMR, with subsequent unmasked final review by attending radiology staff. Misinterpreted radiographs were placed into 3 categories. The groupings included overreads , underreads with no change in treatment, and underreads that required a change in treatment.
Results: A total of 415 radiographs were interpreted by PGY1–3 residents. Overall concordance was found for 371 radiographs (89.4%). There were 44 misinterpretations (10.6%), with 24 (5.78%) overreads, 13 (3.13%) underreads, and 7 (1.69%) underreads that required follow-up interventions. Misinterpretations were similar for the different levels of training:

The 5 most frequently ordered radiographs were chest (28%), ankle (7%), foot (6%), wrist (5%), and hand (5%). The most frequently misinterpreted radiographs were sinus, foot, shoulder, facial, and hand.
Conclusion: 89.4% of all the radiographs interpreted by PGY1–3 residents were read correctly. Only 1.69% of the misinterpreted radiographs led to a change in management. Level of training did not significantly correlate with radiograph misinterpretation rates.