Huang Yijian, Chen Ying Qing
Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA, USA.
Lifetime Data Anal. 2003 Sep;9(3):293-303. doi: 10.1023/a:1025892922453.
Recurrent event data typically exhibit the phenomenon of intra-individual correlation, owing to not only observed covariates but also random effects. In many applications, the population may be reasonably postulated as a heterogeneous mixture of individual renewal processes, and the inference of interest is the effect of individual-level covariates. In this article, we suggest and investigate a marginal proportional hazards model for gaps between recurrent events. A connection is established between observed gap times and clustered survival data with informative cluster size. We subsequently construct a novel and general inference procedure for the latter, based on a functional formulation of standard Cox regression. Large-sample theory is established for the proposed estimators. Numerical studies demonstrate that the procedure performs well with practical sample sizes. Application to the well-known bladder tumor data is given as an illustration.
复发事件数据通常表现出个体内相关性现象，这不仅归因于观测到的协变量，还归因于随机效应。在许多应用中，总体可合理地假定为个体更新过程的异质混合，而感兴趣的推断是个体水平协变量的效应。在本文中，我们提出并研究了一个用于复发事件间隔的边际比例风险模型。在观测到的间隔时间与具有信息性聚类大小的聚类生存数据之间建立了联系。随后，我们基于标准Cox回归的函数形式，为后者构建了一种新颖且通用的推断程序。为所提出的估计量建立了大样本理论。数值研究表明，该程序在实际样本量下表现良好。作为示例，给出了对著名的膀胱肿瘤数据的应用。
