Modelling Marginal Hazards in Multivariate Failure Time Data
研究了在个体聚集成簇且失效时间可能相关的样本中，估计边际分布相对风险参数的方法，提出了回归参数的估计方程和累积基线边际风险的估计量，并给出了大样本分布理论和小样本模拟结果。
SUMMARY We consider the problem of estimating relative risk parameters in the marginal distribution of failure times from a sample in which individuals are grouped into clusters and failure times for individuals within clusters may be dependent. We assume that the marginal hazard of failure for individuals within a cluster is specified by a relative risk regression model and the structure of dependence between individuals within a cluster is left unspecified. Estimating equations for the regression parameters and an estimator for the cumulative base-line marginal hazard are described. Large sample distribution theory based on a representation of the estimating functions as U-statistics is developed and results from a small simulation study are described. Two examples which originally motivated this work are presented.