Dorothy Amy Sippo MD, Presenter: Nothing to Disclose

Gorkem Sevinc, Abstract Co-Author: Co-founder, InSight Medical Technologies Officer, InSight Medical Technologies

John William Nance MD, Abstract Co-Author: Nothing to Disclose

Paul G. Nagy PhD, Abstract Co-Author: Nothing to Disclose

Brandyn D Lau MPH, Abstract Co-Author: Nothing to Disclose

Using our application, we found that 3% of cases selected for daily peer review contained clinically significant errors. These represent an opportunity for targeted quality improvement and continuing education. The fact that two thirds of clinically significant errors resulted from failure to identify a finding suggests that search patterns may be improved by identifying commonly missed findings.

The Joint Commission and American College of Radiology require radiologists to participate in peer review. We have developed an application to facilitate the peer review process at our academic medical center. The purpose of this study is to assess the frequency of clinically significant findings during the peer review process using our application.

We included all radiographic cases selected for peer review (first two cases each day) from January 2013 through March 2014. We also included all cases with independently identified discrepancies, cases where a second radiologist’s reading was different than the original radiologist. For each case, a reviewing radiologist could: (1) concur with the original radiologist; (2) disagree due to failure to identify a finding; (3) disagree with the interpretation of a finding. If the reviewing radiologist disagrees with the original radiologist, they must specify their findings and stratify the error as: (1) clinically significant; (2) not clinically significant. Reviews of outside imaging were excluded from analysis.

Of 4992 cases selected for peer review, the reviewing radiologist disagreed with the original radiologist in 368 (7.4%) cases. 150 cases contained clinically significant findings, resulting from 94 (62.7%) identification errors and 56 (37.3%) interpretation errors. Overall, 3% of peer reviewed cases contained clinically significant errors. Out of 90 studies with a reported discrepant finding, the reviewing radiologist identified 60 (66.7%) cases where the original radiologist failed to identify a finding, of which 43 (71.7%) were clinically significant. Among the remaining 30 (33.3%) cases with interpretation errors, 18 (60%) were clinically significant. Overall, 68% of discrepant findings were clinically significant. 

Sippo, D, Sevinc, G, Nance, J, Nagy, P, Lau, B, Using a Peer Review Application to Identify Clinically Significant Errors.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14045800.html