The distribution of unobserved heterogeneity in competing risks models

被引:1
|
作者
Lu, Yang [1 ]
机构
[1] Univ Paris 13, Sorbonne Paris Cite, CEPN, CNRS,UMR 7234, Paris, France
关键词
Dependent competing risks; Unobserved heterogeneity; Regular variation; Large duration survivor; DURATION; IDENTIFIABILITY; DEPENDENCE;
D O I
10.1007/s00362-017-0956-y
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
We show that in a large class of proportional hazard competing risks models, the distribution of the bivariate frailty among survivors converges to a limiting distribution. This generalizes the result of Abbring and van den Berg (Biometrika 94(1):87-99, 2007), who show that in a single spell duration model, the frailty distribution converges to the gamma distribution. The resulting limiting distribution has an interpretation in terms of partition of a gamma distribution, and allows for both positive and negative correlation between the survival variables. This result provides a natural and flexible specification for the frailty in competing risks models.
引用
收藏
页码:681 / 696
页数:16
相关论文
共 50 条