In July 2013, the institution initiated a structured reporting initiative for abdominal studies, with the goal of standardizing the language and follow-up recommendations for lesions in the liver, kidneys, pancreas, and adrenal glands by using a numerical code similar to the BI-RADS atlas for breasts, according to presenter Dr. Joe Wildenberg, PhD. The structured reporting template was required for all abdominal studies, including those read by residents or fellows on call without immediate feedback from attending radiologists.
Later, when they are overreading the study, the attendings are required to assess the trainee's evaluation of organs within the structured reporting framework, Wildenberg said. During implementation, the researchers realized that this approach could also enable attendings to provide feedback on trainee call errors that was more granular and easier to summarize than the current system (grades of agreement, minor change, or major change). Errors can be broken down by organ as well as type of error, Wildenberg said.
The researchers retrospectively analyzed all studies of the abdomen read on call by trainees between July 2013 and March 2015. They found that the trainee error rate decreased according to year of training, an improvement that was statistically significant.
"More errors were made in lesions of the liver and kidneys than pancreas and adrenal glands, though this is probably due to a higher prevalence of lesions in those organs," Wildenberg told AuntMinnie.com.
The use of structured reporting in radiology can improve the education and assessment of trainees, the group concluded.
"Nearly all training programs assess resident performance in a call setting in some manner; however, most do so without consistent recording of the organ involved or type of error made," Wildenberg said. "This study is a proof of concept for how data in a structured reporting format can be automatically extracted and categorized to provide a more granular assessment of trainee performance."
