Empirical likelihood estimator of medical costs with censored data

The statistical analysis and inference of medical cost is often challenged by the fact that the survival times are censored. In this paper, we consider the case that medical cost history is available. An empirical likelihood procedure is proposed for making statistical inference on the mean cost when censoring is present, then we can easily get the confidence intervals for the parameters of interest at a given confidence level. In the simulation sdudies, we use different sample sizes to compare the finite sample performances of our procedure with some popular procedures in the literatures. We show that our procedure gives better confidential intervals for the mean of medical cost. The proposed procedure is also applied to a real dataset collected in a cardiology trial.