Graphical methods for investigating the finite-sample properties of confidence regions
Computational Statistics & Data Analysis
| Hi-index | 0.03 |
Quality-adjusted lifetime (QAL) has been playing an important role in clinical trials to evaluate treatments for chronic diseases. Much research has been focused on estimating the mean QAL when the data are subject to right censoring. The estimated mean QAL is often presented with a confidence interval, obtained by using either the available variance estimator or the bootstrapping method. However, no research has shown which method yields confidence intervals with better coverage probability and shorter length, especially when the sample size is not big. In this paper, we examine the confidence intervals obtained by some available methods for the mean QAL with censored data. Simulation studies are employed to compare the performance of these methods with various sample sizes and different censoring rates. Methods with the best performance will be identified. A data example from a breast cancer clinical trial study is used to illustrate the application of these methods.