区间删失复发事件数据混合效应模型

The Mixed Effects Model Based on Interval Censored Recurrent Event Data

通过向一般速率模型中引入随机效应项来研究区间删失复发事件数据的个体异质性,允许协变量对含有区间删失的复发事件数据产生影响。在给定随机因子和协变量的条件下建立基于区间删失复发事件数据的混合效应模型。利用极大似然方法估计此模型中所感兴趣的参数,证明此混合效应模型中参数估计的相合性与渐近正态性。通过大量的数值模拟试验研究混合效应模型和参数估计方法的有限样本性质,对一组真实膀胱癌数据的分析表明服用三胺硫磷可以降低膀胱癌的复发率。

In this paper, the random effect item was incorporated into the general rate model to deal with the individual heterogeneity of interval censored recurrent event data, and allowed covariates to affect interval censored recurrent event data, and a mixed effect model was proposed by given random factors and covariates. The model parameters were estimated under the maximum likelihood procedures, the large sample properties in terms of consistency and asymptotic normality properties of the proposed estimates. The finite sample properties of the mixed effects model and the parameters estimate method in this paper were investigated by a large number of numerical simulation experiments. An illustrative example from a bladder cancer study was provided to indicate that the rate of bladder cancer could reduce by taking thiotepa.