A generalized additive regression model for survival times
A generalized additive regression model for survival times
复制标题
生存时间的广义加性回归模型
DOI：
10.1214/aos/1013203457
复制
发表时间：
2001
影响因子：
4.5
通讯作者：
T. Scheike
中科院分区：
文献类型：
--
作者：
T. Scheike
We present a non-parametric survival model withtwo time-scales. The time-scales are equivalent up to a constant that varies over the subjects. Covariate effects are modelled linearly on eachtime scale by additive Aalen models. Estimators of the cumulative intensities on the two time-scales are suggested by solving approximate local maximum likelihood estimating equations. The local estimating equations necessitate only the choice of one bandwidth. The estimators are provided with large sample properties. The model is applied to data on patients with myocardial infarction, and used to describe the prognostic effect of covariates on the two time scales, time since myocardial infarction and age. 1. Introduction. In many bio-medical applications in survival analysis it is of interest to study the effect of covariates on various time-scales. We consider the situation where multiple time-scales are involved, and focus on the specific situation withtwo time-scales that are equivalent up to a constant for eachindividual suchas for example follow-up time and age. A class of models where it is relevant to consider multiple time-scales is the the three-state model known as the illness-death model, or the disability model, where the additional time-scale may be duration in the illness state of the model; see Keiding (1991) for a general discussion of these models. Oakes (1995) discussed how multiple time-scales may be combined into a single scale. Previous work has considered semi-parametric survival models where one time-scale is modelled parametrically and the other time-scale is considered non-parametric. An example of this type of analysis with three time-scales applied to diabetes patients can be found in Ramlau-Hansen et al. (1987). In th is paper we present a non-parametric regression approachwithtwo time-scales where each time-scale contribute additively to the mortality. The effect of covariates are modelled by additive Aalen models on eachtime-scale [Aalen (1980,1989,1993), McKeague (1988), Huffer and McKeague (1991)]. This allows covariates to have effects that vary on two different time-scales. In a motivating example we consider patients that experience myocardial infarction, and aim at predicting the intensity considering the two time-scales age and time since myocardial infarction. As an example, one of the covariates describes the heart function and is allowed to have an effect that varies
