Multimodal Prediction Model for Postoperative Voiding Dysfunction After Pelvic Organ Prolapse Surgery

Abstract

Introduction and Hypothesis: Postoperative voiding dysfunction remains a clinically important complication after advanced pelvic organ prolapse (POP) surgery. We hypothesized that preoperative urodynamic parameters provide incremental predictive value beyond clinical characteristics and uroflowmetry. Methods: We conducted a retrospective cohort study of 283 women undergoing advanced POP surgery. Multivariable logistic regression models were constructed sequentially using clinical variables, uroflowmetry, and urodynamic parameters. Model performance was evaluated by discrimination, calibration, bootstrap internal validation, and decision-curve analysis. Results: Postoperative voiding dysfunction at 6 months occurred in 56 out of 283 patients (19.8%). Complete data for model comparison were available in 271 patients. Discrimination improved stepwise across models: clinical variables alone (areas under the curve [AUC] 0.68), plus uroflowmetry (AUC 0.72), plus urodynamics (AUC 0.78), and the combined model (AUC 0.82). Bootstrap validation demonstrated stable performance with good calibration (optimism-corrected slope 0.80; mean absolute error 0.036). Decision-curve analysis showed greater net clinical benefit for models incorporating urodynamic parameters. Conclusions: Preoperative urodynamic parameters substantially improve the prediction of postoperative voiding dysfunction after POP surgery. A multimodal model integrating clinical, functional, and physiological assessments provides the most accurate risk estimation and supports individualized perioperative management.