Abstract
<title>Abstract</title> <p> <bold>Purpose</bold> : Postoperative pain following Coronary Artery Bypass Grafting (CABG) can delay recovery and reduce quality of life. Early identification of patients at risk of poor pain recovery may support personalized rehabilitation planning. This study aimed to develop machine learning models to predict postoperative pain improvement using preoperative patient characteristics. <bold>Methods</bold> : Data from 192 patients who underwent CABG surgery in six Palestinian hospitals were analyzed. Preoperative demographic, clinical, psy-chosocial, fatigue, and functional variables were used as predictors. Six supervised machine learning classifiers were developed and evaluated using stratified five-fold cross-validation and an independent test set. Model performance was assessed using accuracy, weighted F1-score, and the area under the receiver operating characteristic curve (ROC-AUC). <bold>Results</bold> : The Support Vector Classifier (SVC) achieved the best performance, with a test accuracy of 84.5%, a weighted F1-score of 84.4%, and an ROC-AUC of 91.4%. Preoperative pain severity, fatigue, transfer ability, smoking status, and functional independence were identified as the most influential predictors. 1 Conclusion: Machine learning models can accurately predict postoperative pain improvement after CABG using preoperative data alone. These findings highlight the potential of machine learning to support risk stratification and personalized postoperative rehabilitation planning. <bold>Conclusion</bold> : Machine learning models can accurately predict postoperative pain improvement after CABG using preoperative data alone. These findings highlight the potential of machine learning to support risk stratification and personalized postoperative rehabilitation planning. </p>