A clinical algorithm for the management of abnormal mammograms. A community hospital's experience.

Mammography is an important tool in the early detection of breast cancer, but its use has been criticized for stimulating the performance of unnecessary breast biopsies. We retrospectively reviewed the results of breast biopsies preceded by abnormal mammograms at a community hospital for three 5-month periods--baseline, postintervention, and follow-up--to determine the effectiveness of algorithm-based care for patients with an abnormal mammogram. Cases in which there was a definite or implied recommendation for biopsy by a radiologist revealed a baseline positive predictive value of 4% (2/45), a postintervention positive predictive value of 21% (9/42), and a follow-up phase positive predictive value of 18% (5/28). A Fisher's exact test of the preintervention and postintervention positive predictive values after an abnormal mammogram with a "recommendation for biopsy" was significant (n = 87, P = .023). A Kruskal-Wallis analysis of variance to determine if there had been an increase in the mean lesion size of breast cancers detected over the 3 study periods was not significant. The results of this study suggest that developing a clinical algorithm under the leadership of an opinion leader combined with continuing medical education efforts may be efficacious in reducing the incidence of unnecessary surgical procedures.