Derivation and Validation of a National Multicenter In-hospital Mortality Risk Stratification Model - the Ex-Care Model II
Derivation and Validation of a National Multicenter In-hospital Mortality Risk Stratification Model - the Ex-Care Model II
The goal of this study is to develop a novel surgical risk stratification tool designed to adults operated in Brazil. The main question it aims to answer is:
•Is the Extended Care (Ex-Care) II model a good tool to assess the risk of death among patients operated in Brazilian hospitals within 30 days after surgery?
Information of patients undergoing surgery in the participating hospitals over a period of 24 months will be analyzed to evaluate the relationship of some patients characteristics (called predictors) with the study outcome (probability of death).
The sample will be divided in two groups. The first, called derivation sample, will be used for the development of the Ex-Care II model. The second, called validation sample, will evaluate the performance of this new model.
From four predictors collected preoperatively ( American Society of Anesthesiology Physical Status (ASA-PS) classification, age, extent of surgery and its nature - elective vs urgency/emergency), through a multilevel logistic regression, a post-surgical in-hospital mortality prediction model will be developed, the Ex-Care II model. Discrimination and calibration measures will be used to determine the accuracy of the model. In addition, each predictor will have their odds ratio calculated for the study outcome. Once this model is developed, it will have its external validity tested in a second sample, using the same accuracy metrics.
Inclusion Criteria:
Exclusion Criteria: