Study objective: To determine whether radiologists' review of radiographs is unwarranted when emergency physicians are confident in their interpretations. Methods: This was a prospective, descriptive study. Treating emergency physicians at a high-volume, urban teaching hospital and a moderate-volume nonteaching hospital indicated their confidence or lack of confidence in their plain-film radiographic study interpretations. All radiographs were then reviewed by radiologists according to the preexisting practice of each hospital. A total of 16,410 emergency department radiographic studies were included, comprising consecutive patients at each hospital. Charts of all discordant readings in the confident group were reviewed and judged clinically significant if treatment was altered. Charges for radiologic review of the confident ED interpretations were calculated. Results: The ED physicians were confident in 9,599 sets of radiographs out of a total of 16,410 (58%). Discordant interpretation rates for the 'confident' and 'not confident' groups were 1.2% and 3.1%, respectively (difference, 1.9%; 95% confidence interval [CI] of the difference, 1.44% to 2.36%). Review of the 118 discordant interpretations in the confident group demonstrated that 11 were significant. Few management changes were made as a result of radiologists' review of these radiographs. Total radiology review charges for the confident group were $215,338. Therefore the average radiology charge for each significant discordant interpretation was $19,576. Conclusion: Of 9,599 sets of radiographs interpreted confidently by the emergency physicians in this study, there were 11 clinically significant discordant interpretations (0.1%). The standard practice of radiologists' review of all ED radiographs may not be justifiable.