Development and Internal Validation of a Predictive Model for Postoperative Recovery Quality in Cardiovascular Surgery Patients Based on the QoR-15 Scale: A Retrospective Cohort Study

Objectives A retrospective analysis was conducted to evaluate postoperative recovery quality using the Quality of Recovery-15 (QoR-15) scale in patients undergoing cardiovascular surgery. The study aimed to examine the impact of various perioperative factors on recovery and to develop a predictive model. Methods This retrospective cohort study analyzed clinical data from the medical record system for patients who underwent cardiovascular surgery at a single tertiary care center between March 2020 and September 2022. A total of 198 patients were included in the final analysis after excluding 15 patients due to incomplete data or loss to follow-up. The variables gathered encompassed demographic information (gender and age), duration of postoperative follow-up, American Society of Anesthesiologists (ASA) classification, preoperative lactate levels, emergency surgical status, and whether cardiopulmonary bypass (CPB) was implemented. The modified Frailty Index (mFI) was calculated for each patient to assess baseline frailty. In addition, detailed surgical and perioperative data were recorded. Postoperative data and QoR-15 scores were also included. Univariate and multivariate logistic regression analyses were performed to develop and validate a predictive model. Results A total of 213 patients were included in this study, with 15 patients excluded, resulting in a total of 198 postoperative QoR-15 scores. Gender, ASA classification, preoperative lactate levels, follow-up time, and mFI were identified as independent predictors of excellent postoperative recovery (QoR-15 ≥ 120). The multivariate model showed good discrimination (AUC = 0.925; 95% CI: 0.884–0.966) and internal validation (bootstrap-corrected AUC = 0.901). The Hosmer-Lemeshow test confirmed good calibration ( p  = 0.394). Conclusion A simple model using five routinely available variables demonstrated strong performance in predicting recovery quality. This tool may aid clinicians in identifying patients at risk for poor postoperative outcomes, facilitating personalized perioperative strategies.