Figure 2

The classification tree graph. (legend) The classification tree obtained using the most predictive variables according to the logistic regression analysis, and depicting paths leading to higher/lower prescription probability. The yellow box explains how to read a classification tree graph. The root node shows that, in the whole sample, the most represented class is Statin- (57.7%). Moreover, it shows that the most important variable distinguishing the two classes is Dyslipidemia: when present, 71.3% of patients receive statins at discharge; when absent, only 35.7% (100–64.3) of patients undergo the prescription.