Development And Platform Application Of Predictive Model For Urinary Tract Infection In Neurogenic Bladder Patients

BackgroundPatients with neurogenic bladder(NGB)are prone to urinary tract infection(UTI).The prevention of UTI is of great significance to improvement of management outcomes.However,there are no tools which can predict NGB patients' risk of UTI,and no bladder management platform for NGB patients.Objective1.To establish a predictive model of NGB-UTI so as to screen patients with NGB for UTI.2.To establish an online NGB management platform for NGB patients.3.To evaluate the clinical value of prediction model for NGB-UTI with the help of NGB management platform.Method1.The data about patients with NGB treated in three hospitals from January 2012 to May 2019 were analyzed.Data collected from 2 of the hospitals were used as the training dataset to establish a predictive model for NGB-UTI,while those from the other hospital were as the validation dataset to verify the efficacy of the model.Univariate and multivariate analysis were performed on the training dataset to identify dependent risk factors of UTI.A predictive model for NGB-UTI was established with Software R.C-index,Calibration curve and decision curve analysis(DCA)were utilized to evaluate the performance of the predictive model.The threshold of high-risk population in the model was calculated based on the maximum value of Youden's index.2.With bladder management being the theme,the modules and content of the platform were preliminarily developed based on patients' needs and according to clinical practice guidelines/expert consensuses.Adjustment was made to the platform according to the results of expert consultation,forming NGB management platform.System usability scale(SUS)was used to evaluate the usability of the platform.3.NGB patients meeting the inclusion criteria were recruited online.They were assigned according to the results of the NGB-UTI predictive model into the high-risk group and the low-risk group.The incidence of UTI in the two groups was recorded after 16 weeks of NGB management platform usage,so as to investigate the clinical application effect of the predictive model for NGB-UTI.Result1.The incidence of UTI in the 337 patients with NGB was 45.7%.The C-index of the NGB-UTI predictive model which consisted of length of stay in hospital,urination mode,white blood cell(WBC)count,urine leukocyte(LEU)count and urine pH was 0.921(95%CI:0.87396-0.96804).Bootstrapping was repeated 1,000 times for internal cross-validation and the calibrated C-index was 0.8905541.The net benefit of the model was good in the range of DCA>0.02.The C-index of the validation dataset was 0.817.2.The NGB management platform consisted of 7 modules:The risk assessment,Health knowledge,Bladder management,Questionnaires investigation,Urination diary,Online consultation and user center.Nineteen patients evaluated the usability of the NGB management platform with SUS.The score ranged from 55 to 92.5 with an average of 80.39±10.00.3.The incidence of UTI was 14.5%among 55 patients with NGB after 16 weeks of NGB management platform usage.The number of UTI cases was 6 in the high-risk group and 2 in the low-risk group.The sensitivity,specificity,accuracy and Kappa value of the NGB-UTI predictive model was 75.0%,87.2%,85.4%and 0.515,respectively.ConclusionThe risk of UTI can be evaluated with the predictive model which consists of low-cost and easy-to-collect indicators.The NGB management platform has a good usability score.We explore the clinical effect of the NGB-UTI model by NGB management platform.The results show that the model has relatively high predictive value in clinical application.