Purpose or Case Report: To study the frequency of discrepancies between interpretations of radiologists at outside referring hospitals, overnight residents preliminarily interpreting imaging, and pediatric fellowship trained attending radiologists at a tertiary children’s hospital.
Methods & Materials: We retrospectively reviewed all requests for second opinion interpretations of plain film and computed tomography (CT) imaging at our institution in 2016 that occurred between the hours of 4:30pm to 6:00am, when on-call radiology residents provide preliminary interpretation. Our hospital is a pediatric, tertiary referral, level I trauma center with multiple transplantation services and approximately 80,000 emergency department visits each year. Patients are frequently transferred to our emergency department and inpatient services from a large catchment area extending to Northern California, Nevada, and Arizona, and often arrive with outside hospital diagnostic imaging and radiologist interpretations. Using the RADPEER Scoring System (Table 1), we compared the interpretations of the outside hospital radiologist, overnight radiology resident, and attending pediatric radiologist.
Results: There were 158 requests for second opinion interpretations, of which 149 were CT scans and 9 were plain films. The overnight radiology resident and attending pediatric radiologist concurred in their interpretations (RADPEER score 1) in 145 studies (92%), had discrepancies that were not ordinarily expected to be made (RADPEER 2) in 12 studies (8%), and had a discrepancy that should have been made and was likely clinically significant (RADPEER 3b) in 1 study. Of the 12 RADPEER 2 studies, 9 were likely clinically significant (2b). The radiologist at the outside referring hospital and attending pediatric radiologist concurred in their interpretations (RADPEER score 1) in 146 studies (92%), had discrepancies that were not ordinarily expected to be made and were likely clinically significant (RADPEER 2b) in 11 studies (7%), and had a discrepancy that should have been made but was likely not clinically significant (RADPEER 3a) in 1 study.
Conclusions: There was a high degree of concordance between interpretations provided by outside hospitals, overnight radiology residents, and attending pediatric radiologists at our institution.