Fig. 1

Flow diagram detailing the patients included in the retrospective analysis. Anesthesia cases for non-surgical procedures, such as central venous catheter and endobronchial double lumen tube placements, and electroconvulsive therapy (ECT) were excluded. Only the first surgery/anesthesia was included in the analysis when multiple procedures were performed during the same hospital stay. Surgery/anesthesia cases were divided into the derivation cohort and the validation cohort. Demographic data, pattern of in-hospital death and logistic regression analysis to define preoperative risk factors for the postoperative mortality were obtained using the former cohort. The latter served to validation of the preoperative risk stratification formula