Trainee radiologist reports as a source of confirmation bias in radiology

• Published in: Clinical radiology 2018-12, Vol.73 (12), p.1052-1055
• Main Authors: Nanapragasam, A., Bhatnagar, P., Birchall, D.
• Format: Article
• Language: English
• ISSN: 0009-9260; 1365-229X
• DOI: 10.1016/j.crad.2018.08.003

To assess and quantify the relationship between trainee reporting and radiology errors. A retrospective analysis of the 100 most recent cases reviewed by a discrepancy forum in a tertiary neuroradiology service was performed. Data on the time of the scan and the presence of a trainee report were collected, with comparison being made between the cohort of erroneous reports and the overall service. Although out-of-hours imaging only constituted 18% of the overall service, 36% of erroneous reports originated from scans performed out-of-hours. Eighteen percent of scans were first reported by a trainee and then checked by a consultant, with the remaining 82% being solely reported by the consultant. Despite this, 52% of errors were from consultant-checked trainee reports. Although out-of-hours imaging has long been associated with error, this study identifies consultant checking of trainee reports as another error-associated reporting context. This is likely to relate to confirmation bias, in which the pre-existing trainee report may result in inattentional blindness on the part of the checking consultant. Awareness of this phenomenon is important for the reduction of error in this specific and widely underestimated reporting context. •A pre-existing trainee report is associated with a higher incidence of error.•Confirmation bias of a trainee report may explain the higher discrepancy rates.•Reporting errors are more common with imaging performed in the out-of-hours period.