Quantile regression methods for left-truncated and right-censored data
Quantile regression methods for left-truncated and right-censored data
复制标题
DOI：
10.1080/00949655.2015.1016433
复制
发表时间：
2016-02
影响因子：
1.2
通讯作者：
Jung-Yu Cheng;Shujiao Huang;Shinn-Jia Tzeng
中科院分区：
文献类型：
--
作者：
Jung-Yu Cheng;Shujiao Huang;Shinn-Jia Tzeng
在后续研究中，由于流行的队列抽样，经常会遇到左截断右删失(LTRC)数据。由于生存时间分布的偏向性，分位数回归是一种替代Cox比例风险模型和加速失效时间模型进行生存分析的有效方法。本文将分位数回归模型应用于LTRC数据，建立了回归系数的无偏估计方程。所提出的估计方法使用截断的逆概率和截尾加权技术。由此得到的估计量是一致一致和渐近正态的。通过大量的仿真研究，对所提出的估计方法的有限样本性能进行了评估。最后，通过对实际数据的分析来说明我们所提出的估计方法。
Left-truncated and right-censored (LTRC) data are encountered frequently due to a prevalent cohort sampling in follow-up studies. Because of the skewness of the distribution of survival time, quantile regression is a useful alternative to the Cox's proportional hazards model and the accelerated failure time model for survival analysis. In this paper, we apply the quantile regression model to LTRC data and develops an unbiased estimating equation for regression coefficients. The proposed estimation methods use the inverse probabilities of truncation and censoring weighting technique. The resulting estimator is uniformly consistent and asymptotically normal. The finite-sample performance of the proposed estimation methods is also evaluated using extensive simulation studies. Finally, analysis of real data is presented to illustrate our proposed estimation methods.