Robust Procedure for Handling Censored Data in Clinical Trials

One of the difficult aspect in parameter estimation in survival analysis is the presence of censored values in survival data When patients survival time are measured in continuous time interval, the censored values continue to create discrepancies in estimations since the stochastics realization of censored values are masked. For this reason, appropriate distributions must be specified before maximum likelihood is used. An optimal approach that merges both censored values and uncensored values is appreciable since asymptotically distributions are not normal. Robust model and efficient algorithms developed to enhance optimal performance in estimation. Simulations show that the optimal robust model is maximized.