Predictive capabilities of postoperative complications in abdominal surgery: from diagnostic scoring systems to clinical decision support systems (literature review)

Postoperative specific and nonspecific complications in abdominal surgery remain a significant clinical and organizational problem, associated with increased mortality, prolonged hospital stay, and higher treatment costs. Despite advances in modern surgical technologies, their incidence has not significantly decreased over recent decades, indicating the need to improve approaches to risk assessment and management of complications. To analyze the role of modern diagnostic scales and existing clinical decision support systems in risk assessment and prevention of postoperative complications in abdominal surgery. A review of domestic and international publications was conducted, focusing on postoperative complications, surgical risk assessment scales, outcome registries, surgical safety checklists, and clinical decision support systems, including soft computing methods and machine learning approaches. Traditional clinical and scoring risk scales were shown to have limited predictive accuracy and are generally focused on individual complications or specific nosologies in abdominal surgery. Regression and machine-learning models demonstrate higher predictive accuracy; however, they require large datasets and substantial computational resources. The use of surgical registries and safety checklists potentially reduces complication rates, but their implementation is limited by organizational and methodological factors. Existing methods for predicting postoperative complications are fragmented. A relevant objective is the development of integrated modular clinical decision support systems that account for individual patient factors and improve the effectiveness of complication prevention.