In recent years techniques for the selection of risk groups for high priority breast screening have been evaluated. We used data from a large multicentric case control study in Italy, recently published, to fit a reduced logistic model for predictive purposes, using variables which are easily collected. The predicted probability of being a case obtained through the logistic analysis was then used to classify the same set of data by calculating values of sensitivity and specificity at different cutoff points of the predicted probability. The results showed the low predictive power of the model. The selection of higher risk groups in a screening population program is discussed.