本計畫擬針對二組具有共變?的右設限存活資?，研究其限制平均壽命(restricted mean lifetime)差?的估計。於97 ?國科會計畫，探討如何在共變?已知情形下，建構此一限制平均壽命差?的條件信賴區間，並且應用此ㄧ比較各劑?組與?劑?的條件信賴區間，鑑別共變?已知情形下的最低有效劑?。所以在本計畫中，我們考慮進一步研究二個限制平均壽命差?為共變?函?的信賴域，並且將之應用於最低有效劑?之鑑別，求出對應各種可能最低有效劑?的共變?範圍。引用實際資??明此方法的應用，並且以模擬方法研究其相關的覆蓋機?及信賴域的面積。此外，也研究應用此信賴域鑑別MED 的準確性。 In this project, we consider estimation of the difference of two restricted mean lifetimes with covariates for randomly right-censored survival data. We construct the conditional confidence interval for the difference given the covariates in the current project granted for 2008. We then apply the confidence intervals for comparing each treatment with zero-dose control to identify the minimum effective dose (MED). Therefore, in this project we further set up the confidence band for the difference as a function of covariates. We also discuss the application of the confidence band for identifying the MED. The implementation of the confidence bands will be illustrated using real data sets. The associated coverage probability and the expected area of the band will be investigated via a simulation study. We also investigate the accuracy of the MED identification based on the proposed confidence band.
