Survival analysis is widely applied to data that described by the length of time until the occurrence of an event under interest such as death or other important events. The purpose of this paper is to use the dynamic methodology which provides a flexible method, especially in the analysis of discrete survival time, to estimate the effect of covariate variables through time in the survival analysis on dialysis patients with kidney failure until death occurs. Where the estimations process is completely based on the Bayes approach by using two estimation methods: the maximum A Posterior (MAP) involved with Iteratively Weighted Kalman Filter Smoothing (IWKFS) and in combination with the Expectation Maximization (EM) algorithm. While the other method was represented by the Hybrid Markov Chain Monte Carlo method (HMCMC). Moreover, two hazard function models were considered in the comparison: the Logistic model and the discrete Cox model. Two criteria were used for comparisons Average Mean Square Error: AMSE and Cross Entropy Error: CEE. All these four combinations of methods were clarified via the discussion of the numerical results with their explanations. It can be noticed the superiority of HMCMC method through the two hazard models.
