Predicting Time to Reflux of Children With Antenatal Hydronephrosis: A Competing Risks Approach

• Main Authors: Maryam Nazemipour, Abdol-Mohammad Kajbafzadeh, Kazem Mohammad, Abbas Rahimi Foroushani, Mahmood Mahmoudi
• Format: Article
• Language: English
• Published: Tehran University of Medical Sciences 2017-09-01
• Series: Acta Medica Iranica
• Subjects: Competing risks; Trivariate weibull survival; Reflux; Hydronephrosis
• Online Access: https://acta.tums.ac.ir/index.php/acta/article/view/5653
• ISSN: 0044-6025; 1735-9694

The aim of this study was describing methodological aspects and applying a trivariate Weibull survival model using the competing risks concept to predict time to occurrence different types of reflux (unilateral (left, right) or bilateral) in children with antenatal hydronephrosis. Data from 333 children in Pediatric Urology Research Center of Children’s Hospital Medical Center, affiliated with Tehran University of Medical Sciences was used. The effect of some demographic and clinical factors on child’s reflux was studied. The assumption of independent between times of different types of reflux was evaluated. Of infants 80.5% were boy. The percentage of children experienced right, left and bilateral reflux or have been censored are 15.3%, 14.1%, 60.4% and 10.2% respectively. For the time of left reflux, variables, Week of diagnosis ANH, UC, UA, HUN, HN, APD_Right, Direction of ANH, CA19-9 baby, Urethra were significant. For the time of right reflux, variables, constipation, UC, UA, HUN, APD_Right, Direction and Severity of ANH, Bladder, and finally for the time of bilateral reflux, variables, Week of diagnosis ANH, Gender, UA, HUN, HN, APD_Left, Urethra, and Bladder were significant P<0.05. In the presence of competing risks, it is inappropriate to use the Kaplan-Meier method and standard Cox model which do not take competing risks into account. Trivariate Weibull survival model using competing risks not only is able to calculate the hazard rate of variables with different type of events but also it will be able to compare the hazard rate within the same type of event with different covariates.