一本书：survival analysis生存分析

我曾经问过理查德库克the difference between the logistic model and proportional hazzard model. He replied that it was all about time and logistic cannot catch the time.

xiexie!!!!!!!!!!

9# gooddrug

SAS 在生存分析上做得还可以，比较全面。原因无外乎是生存分析在医疗制药公众健康上用的较多，而这些行业也恰恰是SAS的重点服务之一。

楼主能多介绍一下本书吗，有点贵啊

14# zhucheng1980

这本书包括了不同的专家对生存分析的方方面面进行的讲解，其中包括PH model, parametric models on different distributions, competing risk, multiple events (recurrent data, cluster data), frailty model, and so on. Some of them involve comparatively cubesome theoretical illustrations while most of them are applied with many examples.
许多的章节是应用举例和理论相结合的。

14# zhucheng1980

I General methodology.
II Censored data and inference.
III Truncated data and inference.
IV Hazard rate estimation.
V Comparison of survival curves.
VI Competing risks and analysis.
VII Proportional hazards model and analysis.
VIII Accelerated models and analysis.
IX Frailty models and applications.
X Models and applications.
XI Multivariate survival data analysis.
XII Recurrent event data analysis.
XIII Current status data analysis.
XIV Disease progression analysis.
XV Gene expressions and analysis.
XVI Quality of life analysis.
XVII Flowgraph models and applications.
XVIII Repair models and analysis.

14# zhucheng1980

I General methodology.
II Censored data and inference.
III Truncated data and inference. phreg /counting process
IV Hazard rate estimation.phreg
V Comparison of survival curves. lifetest
VI Competing risks and analysis. phreg
VII Proportional hazards model and analysis. phreg
VIII Accelerated models and analysis. lifereg
IX Frailty models and applications.
X Models and applications.
XI Multivariate survival data analysis.
XII Recurrent event data analysis. phreg/counting process or gap time
XIII Current status data analysis.
XIV Disease progression analysis.
XV Gene expressions and analysis.
XVI Quality of life analysis.
XVII Flowgraph models and applications.
XVIII Repair models and analysis. reliability

以上列举一些SAS对应的procedures。欢迎指正。

1. 
   
2. XII Recurrent event data analysis. phreg/counting process or gap time

复制代码

More in details about recurrent event:

1. Anderson and Gill (AG) model: counting process +sandwich estimator
2. Wei, Lin, and Weissfeld (WLW)  model: stratified Cox + sandwich estimator
3. Prentice, Williams and Peterson (PWP) model
   (1) total time
   (2) gap time

all have been well applied in SAS.

感谢楼主分享这么多，向楼主致敬！谢谢

19# weihancool

关于Sandwich estimator的一点看法：

在SAS生存分析之全部procedures中，皆为独立地处理每个data line。换而言之，行和行之间是不相关的。但是，当一个ID贡献不只是一条记录时，这些记录之间就存在着相关性，而sandwich estimator正是用于处理这种相关性（correlation). 当然，此估计只局限在标准差之上，而系数值仍保持不变。
如果联想一下，比如repeated data 之GEE模型，不也正是用这种三明治式的估计来处理相关性吗？