The Ensemble-Based Risk Estimation System (ERES) is a machine learning model developed to predict postoperative mortality risk in patients undergoing cardiac surgery. Designed to enhance and potentially replace the traditional EuroSCORE system, this model specifically targets patients undergoing coronary artery bypass grafting and/or valve surgery.
The model is based on a retrospective analysis of 543 patients, utilizing preoperative clinical data. The dataset includes demographic information, clinical characteristics, and preoperative laboratory results of the patients.
ERES incorporates significant clinical parameters such as creatinine level, age, and left ventricular ejection fraction. The model utilizes 15 key features to demonstrate superior predictive performance.
ERES offers a more accurate estimation of mortality risks for patients.
📌 For a detailed understanding of the model’s application, please refer to the flowchart provided in this repository, which guides through the model's workflow and implementation steps.
The model has been evaluated using various machine learning algorithms and compared with the EuroSCORE I model.
📊 ERES has been assessed through calibration plots and ROC curve analysis, showcasing advanced prediction capabilities and suitability for clinical use.