or an evaluation of the results of cardiovascular surgery, crude operative mortality rate (ie, the number of deaths divided by the number of operations) is generally employed, although these raw outcome data do not reflect the preoperative condition of the individual patient. Therefore, to resolve that problem, risk stratification analysis was developed 1-3 for 2 major uses: one is estimating the risk of surgical mortality faced by an individual patient, as an aid to patients and physicians contemplating cardiovscular surgery, and the other is as a tool for measuring the quality of surgical care. Recently, the European System for Cardiac Operative Risk Evaluation (Euro SCORE) additive model was introduced, and it has global scores with simple and objective definitions provided at the data collection stage. 2 We used the EuroSCORE algorithm to assess the biennial trend towards a reduction in operative mortality rate in order to assess the quality of cardiac and thoracic aortic surgical care in our institution.
MethodsBetween August 1994 and December 2000, 803 consecutive patients had surgery for cardiac and thoracic aortic disease: 486 men (61%) and 317 women (39%), aged from 13 to 90 years old (mean, 63±12 years). All patients selected had to have undergone cardiovascular surgery with cardioCirculation Journal Vol.66, February 2002 pulmonary bypass, so those who had off-pump coronary artery bypass grafting (CABG) were excluded. The patient population was divided into 3 time periods: early (August 1994 through September 1996, n=260), middle (October 1996 through September 1998, n=259), and late (October 1998 through December 2000, n=284). The distribution of the surgery performed is shown in Table 1, and the data for the EuroSCORE risk factors, which was entered into a database, is shown in Table 2. Expected or predicted mortality was calculated for individual patients using the EuroSCORE additive model, arranged sequentially in order of predicted score. The study population was divided into 5 clinically relevant risk categories: 0-2% risk, 3-5% risk, 6-8% risk, 9-11% risk, and 12+% risk. Mortality was defined as death from any cause within 30 days of operation or within the same hospital admission, and expected mortality was compared with observed or actual mortality for each risk category.The continuous data were expressed as the mean ± one standard deviation, and categorical variables were expressed as percentages. Statistical analysis was conducted using StatView 5.0 (SAS Institute Inc, Cary, NC, USA). Comparison of 2 groups was performed for categorical variables with the chi-square test with 2 × 2 contingency tables or Fisher's exact test as appropriate. The receiver operating characteristic (ROC) curves were plotted to assess the discrimination abilities (accuracy) of the EuroSCORE. 4 The area under the ROC curve was calculated as an index for how well the model could discriminate between patients who lived and those who died. The discriminative power of the model is thought excellent if the area under ...