screens, but a higher percentage of abnormal mammograms, biopsies, and false-positive findings. Therefore—if judged by rates of false positives—the apparent performance of a mammography service or individual radiologist would be influenced by the proportion of how many women are receiving their first mammogram.
A relationship between the volume of procedures performed and the outcome of those procedures has been established for many complex medical procedures, particularly in surgery and oncology.55 Many studies have suggested that the volume of mammograms read by a radiologist is correlated with accuracy, and mammography volume standards are mandated by federal law. However, relatively few studies have directly compared the number of mammograms read by a radiologist and the accuracy of their interpretations. The results of these are variable, and are shown in Table 3-4. The most comprehensive study to date was the analysis by Beam and his colleagues in 2003,12 which indicated that the volume of mammograms interpreted by a radiologist accounts for less than 2.5 percent of the variation. This means that more than 97 percent of inconsistency in interpretation is due to other factors.