Bayesian competing risks analysis without data stratification

Background: The Cox Proportional Hazard (CPH) model is a commonly used method to study death as the outcome of therapeutic effect. However, death may occur due to other causes not as the direct therapeutic effect. Death due to other causes is known as competing risks (CR). The conventional strategy is to handle the CR inflated death data with the stratified CPH model. However, it reduces the power of the data analysis. It is important to establish a model that can handle the CR without stratifying the data. This paper is devoted to developing the CR model. Objective: Develop a statistical model that can handle death data analysis without stratifying in the presence of CR. Methods: Statistical methodology is established to deal with CR. The Bayesian Inference is considered to overcome the computational difficulties. The Markov chain Monte Carlo (MCMC) was used to run the model. Model is illustrated for lung adenocarcinoma patient's data. The arm wise comparison on adenocarcinoma treated patients is performed. The computation of the competing risk model is performed in open source software OpenBugs. Results: The proposed method helps to obtain the posterior estimates of different covariates with credible intervals. The posterior mean (SD) estimate helps to decide the best effective treatment through survival analysis in presences of CR. Conclusions: This work is helped to work with CR data without stratifying it. It provides flexibility to analyze death data with CR by Bayesian.