Parametric and the Cox risk model in the analysis of factors affecting the time of diagnosis of retinopathy with patients type 2 diabetes

Abstract

Background: The aim of this study was to compare the effectiveness of Cox model and Exponential parametric, Weibull, Log Normal and Log Logistic models in evaluating factors affecting retinopathy diagnostic time in patients with type 2 diabetes. Methods: In this prospective historical study, 400 patients with type 2 diabetes without retinopathy referred to the Ophthalmology Clinic of Yazd Diabetes Research Center in 2008 were followed up for diagnosis of retinopathy by January 2013. Significant variables in the univariate model were introduced into the Cox multivariate and parametric models to determine the effective factors on the time of retinopathy diagnosis. The criterion for comparing the performance of the models was the Akaike’s criterion. All calculations were performed using R software and a significant level of 0.05 was considered. Resuls: The mean and median time of retinopathy diagnosis was 52.46 and 58 months, respectively. 3% of patients in less than one year and 16% of patients in less than two years of retinopathy were diagnosed. Conclusion: According to Akaike’s criterion, Cox model has the best fit in determining the factors affecting the time of retinopathy diagnosis.