False positive rates varied among diagnostic mammography facilities

Variability in the results of diagnostic mammograms was found to exist among facilities, according to the results of a cross-sectional survey of facilities in the Breast Cancer Surveillance Consortium. Specifically, the rate of false positives was almost 50% higher among facilities who reported recommending additional tests out of concern about malpractice.

Researchers conducted a survey of 45 facilities that performed diagnostic mammography. Facilities were surveyed about their structure, organization and interpretive processes.

From January 1998 to December 2005, measurements of false-positive rate, sensitivity, and likelihood of cancer among women referred for biopsy were taken. Thirty-two facilities performed diagnostic mammography; 28,100 diagnostic mammograms were performed. One hundred eighteen radiologists at the participating facilities completed a survey examining radiologist characteristics and perceptions.

For facilities, the overall mean false-positive rate was 6.5% (95% CI, 5.5%-7.4%), and mean sensitivity was 73.5% (95% CI, 67.1%-79.9%). Positive predictive value of biopsy recommendation was 33.8% (95% CI, 29.1%-38.5%).

In an unadjusted analysis, variability was observed across facilities for sensitivity (P=.006), false-positive rate (P<.001) and positive predictive value of biopsy recommendation (P<.001). However, when adjusted for patient and radiologist characteristics, variability in performance also persisted for false-positive rates (P<.001).

Facilities reporting that concerns about malpractice greatly increased recommendations for additional tests at the facility were linked to even higher false-positive rates (OR=1.48; 95% CI, 1.09-2.01).

Jackson SL. J Natl Cancer Inst. 2009;101:814-827.