PREDICTING THE RESULTS OF TREATMENT OF PRIMARY VESICOURETERAL REFLUX IN CHILDREN

Abstract

Purpose. Creation of a predictive model for calculating the effectiveness of endoscopic correction of vesicoureteral reflux by binary logistic regression.Materials and methods. The study included 150 children (240 renal units) aged 1 month to 7 years who underwent endoscopiccorrection of vesicoureteral reflux. As possible success predictors were evaluated reflux grade, ureteral diameter ratio, time of reflux at cystourethrography, the dilatation of the renal pelvis and ureter according to ultrasound.Results. Statistically significant differences were found in predicting the outcome of endoscopic correction of vesicoureteral reflux according to the selected prognostic parameters (p<0.001). To assess the diagnostic significance of predictors in predicting the outcome, the method of analyzing ROC curves was used. A predictive model was developed to determine the probability of the results of three attempts at endoscopic correction, depending on the selected predictors by binary logistic regression. The sensitivity and specificity of the models was over 90%. The resulting models are combined into a predictive calculator, where you can enter variables and generate an individual treatment outcome prediction.Conclusion. The resulting computational model is a potentially useful clinical tool in the choice of treatment for vesicoureteralreflux.