Semiparametric Proportional Mean Residual Life Model With Censoring Indicators Missing at Random
Semiparametric Proportional Mean Residual Life Model With Censoring Indicators Missing at Random
复制标题
DOI：
10.1080/03610926.2013.879894
复制
发表时间：
2015-12
期刊：
影响因子：
--
通讯作者：
Xiaolin Chen;Qihua Wang
中科院分区：
文献类型：
--
作者：
Xiaolin Chen;Qihua Wang
对于右删失生存数据，观察时间是生存时间还是删失时间的信息经常丢失。竞争性风险数据的情况就是如此。本文研究了比例平均剩余寿命模型下截尾指标随机缺失的右截尾生存数据的统计推断问题。提出了简单的和增广的逆概率加权估计方程方法，其中不遗漏概率和一些未知的条件期望值由核平滑技术估计。所有提出的估计的渐近性质的建立，而广泛的模拟研究表明，我们提出的方法表现良好的中等样本容量下。最后，所提出的方法被应用到第二阶段乳腺癌试验的数据集。
For right-censored survival data, the information that whether the observed time is survival or censoring time is frequently lost. This is the case for the competing risk data. In this article, we consider statistical inference for the right-censored survival data with censoring indicators missing at random under the proportional mean residual life model. Simple and augmented inverse probability weighted estimating equation approaches are developed, in which the nonmissingness probability and some unknown conditional expectations are estimated by the kernel smoothing technique. The asymptotic properties of all the proposed estimators are established, while extensive simulation studies demonstrate that our proposed methods perform well under the moderate sample size. At last, the proposed method is applied to a data set from a stage II breast cancer trial.