Radiologists who read fewer than 1,000 mammography studies per year may have worse diagnostic accuracy than those who read more, according to a new study published online June 4 in Radiology. The performance threshold is more than twice what is required in the U.S. for radiologists who read mammograms.
Researchers from the University of Sydney in Australia found that perceptual errors linked to poor performance in mammographic interpretation were influenced by the number of readings radiologists performed annually. Perceptual errors account for 60% of all diagnostic mistakes in radiology, the group noted. These errors arise from inadequate knowledge to be able to search an image properly, inability to recognize an abnormality, and faulty decision-making as to whether an abnormality found on an image is normal or a pathologic lesion.
"As long as radiologists interpret medical images, the human factor and its limitations on abnormality detection will continue to influence overall performance," lead author Mohammed Rawashdeh and colleagues wrote (Radiology, June 4, 2013).
Different countries set different levels for the minimum annual volume of readings that radiologists should perform, according to Rawashdeh's group: In the U.S., the number is 490 readings per year. In Australia and Canada it's 2,000 readings (although British Columbia requires 2,500 readings), and in the U.K. it is 5,000 readings.
For this study, 116 radiologists read 60 mammographic cases -- 20 cases with cancer and 40 cases with normal findings -- and scored their level of confidence in their interpretation on a scale of 1 to 5. The findings were analyzed using the jackknife free-response receiver operator characteristics (JAFROC) method, which addresses lesion location and multiplicity and is more representative of how radiologists actually interpret images in clinical practice, according to the authors.
Rawashdeh's group found that reader performance correlated in a statistically significant way with the number of years since qualification as a radiologist, the number of years reading mammograms, and the number of readings per year. Radiologists who read fewer than 1,000 mammograms per year had lower performance scores than those who read 1,000 or more, up to 5,000 per year, according to the team.
In addition, readers with an annual volume of fewer than 1,000 mammograms didn't necessarily get better over the course of their careers; instead, they showed reduced performance with increased years of reading mammograms.
"This inverse relationship between performance and years of experience has ... important implications for radiology, as a reasonable assumption that a low volume of readings can be compensated for by years of experience is unfounded," Rawashdeh and colleagues wrote.
The opposite was also true, the researchers discovered: Radiologists who read more than 5,000 mammography exams got better over time.
"The true discriminating factor that separates individuals performing at the highest levels from others is the ability to recognize what is normal," the group concluded.
