Certain patient characteristics, such as age and breast density, are known to contribute to mammographic accuracy, but it is not known how the characteristics of physicians trained for mammogram interpretation, such as experience or mammogram volume, can affect this. In the United States, the Mammography Quality Standards Act requires physicians to interpret at least 960 mammograms over a 2-year period--about 10 mammograms per week--to be qualified for this task. Physicians in the United Kingdom National Health Service Breast Screening Program must interpret 10 times that amount. However, the few studies that have evaluated the relationship between annual volume of mammogram interpretation and accuracy have had conflicting results.

To identify characteristics of U.S. physicians that are associated with mammographic accuracy, Rebecca Smith-Bindman, M.D., of the University of California, San Francisco, and colleagues modeled mammography sensitivity (rate of true-positive results) and specificity (rate of false-positive results) using data from the Breast Cancer Surveillance Consortium and the American Medical Association Master File.

The false-positive rate varied from 1% to 29% between individual physicians, but was lowest, in general, among physicians with the most experience. Physicians who had been practicing the longest, who interpreted 2,500 to 4,000 mammograms annually, and who emphasized screening--instead of diagnostic--mammography had lower false-positive rates.

The researchers estimated that, compared with physicians who met the minimum standards of the Mammography Quality Standards Act and who focus less on screening, physicians who interpret 2,500 to 4,000 mammograms annua
'"/>

Contact: Sarah L. Zielinski
jncimedia@oupjournals.org
301-841-1287
Journal of the National Cancer Institute
1-Mar-2005