Background: The aim of this study was to compare the ability of three risk models to predict operative mortality after cardiac surgery.
Methods: Risk factors of 3111 patients (73% male, mean age 65.2 +/- 10.7 years) were derived from our institutional database at the Cardiac Surgery Department of the G.M. Lancisi Hospital, Ancona, Italy. The predicted mortality was derived from the Society of Thoracic Surgeons risk score (STS), the EuroSCORE (ES) and the Northern New England Cardiovascular Disease Study Group score (NE).
Results: The observed mortality in the myocardial revascularization population (1995 patients) was 2.2% (43 patients). The mean predicted mortality by STS, ES and NE was 1.9, 4.2 and 1.9%, respectively. The predictive ability of the models was measured by means of the ROC curve. Curves were respectively of 0.82, 0.77 and 0.78.
Conclusions: All tested models proved e good accuracy level but ES showed a constant overestimation of mortality at all risk levels.