Analyzing competing risks data using bivariate Weibull-geometric distribution

The motivation of this paper came from a study which was conducted to examine the effect of laser treatment in delaying the onset of blindness in patients with diabetic retinopathy. The data are competing risks data with two dependent competing causes of failures, and there are ties. In this paper we have used the bivariate Weibull-geometric (BWG) distribution to analyse this data set. It is well known that the Bayesian inference has certain advantages over the classical inference in certain cases. In this paper, first we develop the Bayesian inference of the unknown parameters of the BWG model, under a fairly flexible class of priors and analyse one real data set with ties to show the effectiveness of the model. Further, it is observed that the BWG can be used to analyse dependent competing risk data quite effectively when there are ties. The analysis of the above-mentioned competing risks data set indicates that the BWG is preferred compared to the MOBW in this case.