Hazel Janet Scott BA, Presenter: Nothing to Disclose

Alastair George Gale PhD, Abstract Co-Author: Nothing to Disclose

Robin Wilson FRCR, Abstract Co-Author: Nothing to Disclose

In the UK, radiologists have to read a minimum of 5,000 cases a year. However, it can be fatiguing to read many normal cases and needing to maintain a high sensitivity to detect the rare early malignant appearance. We examined whether fatigue was an issue during a reading session using a simulated screening task.

The performance of 250 breast screening radiologists on a set of 360 selected difficult normal, benign and malignant cases was examined. Participants read the case set in groups of 60 cases, each in a single session, with the case order randomized. Participants were free to read the cases at a speed which suited them. The timing of the case reporting process (i.e. identifying key mammographic features and determining a BIRADS type of classification) was recorded.

Participants varied in the overall speed with which they read each group of 60 difficult cases, taking between 60 -120 minutes. This is slower than in typical UK screening practice as in the current task they had to report much more detailed information. Errors (false positive and false negative decisions) were called as those cases where a participant’s classification decisions differed from a previously determined radiological ‘gold’ standard decision. Each set of 60 cases were then split into six groups of 10 for time series analysis and the errors within each 10 films examined. The number of errors were found to be related to time on task.

These data indicate that, in reading a series of cases, performance varies, indicating potential fatigue effects.

Vigilance is a potential source of error in reading films within a single session.

Scott, H, Gale, A, Wilson, R, Does Fatigue Contribute to Error Rate in Breast Screening?.  Radiological Society of North America 2006 Scientific Assembly and Annual Meeting, November 26 - December 1, 2006 ,Chicago IL. http://archive.rsna.org/2006/4432380.html