Heliyon (Oct 2023)
Development of a prediction model for postoperative complications and economic burden analysis in older patients with hip fractures
Abstract
Purpose: The high rates of disability and mortality due to postoperative complications of hip fractures in the elderly, especially the oldest-old individuals, have become an increasingly serious global public health concern. This study aimed to establish a nomogram prediction model and analyze the economic burden to guide clinical decision-making and improve patient prognosis. Methods: Data of 514 patients aged over 80 years with hip fractures who received surgical treatment were retrospectively collected, and the patients were divided into training and validation cohorts. Independent risk factors for postoperative complications were identified based on logistic regression analysis, and a nomogram was constructed. The model was evaluated for its discrimination and consistency using receiver operating characteristic (ROC) curves and calibration curves, and for its clinical benefit using decision curve analysis (DCA). The economic burden was analyzed using propensity score matching (PSM). Results: The American Society of Anesthesiologists (ASA) classification ≥Ⅲ, anemia, male sex, diabetes mellitus, and the number of comorbidities were found to be independent risk factors for postoperative complications in oldest-old patients with hip fracture (all P < 0.05). The areas under the curve (AUC) of the nomogram prediction model for the training and validation cohorts were 0.743 and 0.767, respectively, indicating reliable discrimination. The calibration curves and DCA showed that the model has good consistency and high benefits. The direct economic burden of postoperative complications for the patients was US$1045.10. Conclusions: The nomogram model can accurately quantify the risk of postoperative complications among oldest-old patients with hip fractures and guide clinical professionals to implement early and targeted preventive treatment for high-risk patients.
