Development and Validation of A Risk Prediction Model for Recurrent Female Pelvic Organ Prolapse After Prolapse Surgery at The University of Gondar Comprehensive Specialized Hospital, Northwest Ethiopia

Abstract

Introduction: Recurrence of pelvic organ prolapse is one of the most common complications after prolapse surgery. In Ethiopia, the burden and prediction model of recurrence was not studied. So, this study stratified the risk of recurrence of pelvic organ prolapse based on baseline characteristics of patient factors. It helps the clinician to plan the most durable surgical procedures based on patient's risk factors. Objective: to develop and validate a risk prediction model for postoperative recurrence of pelvic organ prolapse at the University of Gondar Comprehensive Specialized Hospital, Northwest Ethiopia, from January 2019 to January 2023. Methods: A retrospective follow-up study was undertaken at Gondar Comprehensive Specialized Hospital on 611 patients who had received pelvic organ prolapse surgery between January 1, 2019, and June 30, 2023, to predict the recurrence risk of pelvic organ prolapse within one year of primary surgery. Data were gathered using the open-source freeware KoboToolBox. Finally, the data were analyzed with STATA version 17.0. Predictors were selected using Lasso regression, and nomograms were developed using multivariable logistic regression with a P-value < 0.05. The model's internally validated using bootstrapping, and its performance was assessed using discrimination and calibration plots. Result: In this investigation, the prevalence of recurrent pelvic organ prolapse was determined to be 19.6%. The woman's age, residency, occupation, marital status, parity, mode of delivery, duration of prolapse symptoms, presence of a decubitus ulcer, stage of POP, and presence of postoperative complications were all substantially linked with recurrent POP. With this predictor nomogram was constructed with a calibration test p-value of 0.861 and a discriminating power of 96.3% (95% CI: 93.77, 97.13). The model had sensitivity of 87.5%, specificity of 86.35%, positive predictive value of 61.46%, negative predictive value of 96.58%, and accuracy of 86.58%. The nomogram had a significant net benefit between the 0.2 and 0.8 threshold probabilities. Conclusion: The generated model has high performance and good calibration; it is used to predict the recurrence of pelvic organ prolapse following surgery. The model proved beneficial in clinical settings, as validated by a net benefit analysis. The clinicians can use it to stratify cases and select the most durable surgical techniques based on risks