Referral policy and positive predictive value of call for surgical biopsy in the florence breast cancer screening program

The authors evaluate the referral rates and the predictive value (PV) of call for diagnostic work up or biopsy in a population based breast cancer screening program in the 1979–1986 period. The presence of mammographic abnormalities either benign or suspicious proved to be the only reliable referral criterion (recall rate = 7.9%, recall PV = 3.8%, biopsy call PV = 39%). When mammography was normal the presence of breast complaints other than pain or of a radiologically dense breast were aspecific referral criteria, and since the latter was abandoned, (a) referral rates dropped especially in younger women, (b) referral or biopsy PV improved, whereas (c) cancer detection rate was almost unaffected. Younger age was associated with higher referral rates and with lower PV of both referral or biopsy. Whenever comparison with previous examination was possible, mammography accuracy was higher; this explains the decrease in referral rates and the increased predictivity observed at further rounds with respect to first screening round.