比例优势模型在多结局区间截断数据上的应用

探索比例优势模型在临床医学中常见的多结局区间截断数据中的应用.用条件的逻辑回归方法避免讨厌参数的估计,用牛顿-拉普森算法估计回归系数,用＂夹心方差＂估计量作为参数方差的估计.通过随机模型检验模型应用的有效性.

Applications of Proportional Odds Model to Multivariate Interval-Censored Data

Since the multiple failures situation is not uncommon in the clinical medicine,we explore the use of proportional odds model to the multivariate interval-censored data.The approach is based on the conditional logistic regression,which prevents the complications in the existence of nuisance parameters.The estimation of parameters is obtained by the Newton-Raphson algorithm.The sandwith estimator for the covariance is made according to the situation where there is correlation in the score statistic.Simulations are also presented to assess the accuracy of the procedure.