Prognostic index: prediction of operative mortality in geriatric patients by use of stepwise logistic regression analysis

The decision-making process with regard to surgery in geriatric patients is of great importance, requiring consideration of ethical, medicolegal, and economic factors in addition to the purely medical ones. An important component in the decision-making process is the precise knowledge of the expected mortality. Clearly, such knowledge helps the surgeon as well as the patient and family to make the final decision whether to operate. Twelve hundred major laparotomies in elderly patients were used as a data base. Thirty-six preoperative variants were noted for each patient as well as the occurrence of postoperative mortality or morbidity. All data were loaded into a 6,000 CDC computer and subjected to Multifactorial Stepwise Logistic Regression Analysis (MSLRA). The method allowed a choice of the 5 most significant factors in the prediction of mortality, as well as calculation of a numerical coefficient for every factor. A formula (MIX Formula) was devised to calculate the mortality in percentile terms using the factor coefficient. The results of this study show that the use of statistical multifactorial analysis permits an accurate preoperative prediction of postoperative mortality in geriatric candidates for major abdominal surgery. The accuracy of the prediction was confirmed by both retrospective and prospective analysis.