标题: 求助：生存分析方法的选择？ [打印本页] 作者: shiyiming 时间: 2010-4-16 15:14 标题: 求助：生存分析方法的选择？ 回访的数据中既有到期仍存活（终检值？），又有失访，应该用哪种分析方法？
我用lifetest和phreg两种方法都试了，lifetest方法的logrank检验，P=0.4,phreg方法的P=0.2，虽然都没有统计学意义，但为何有差异？
谢谢！！！作者: shiyiming 时间: 2010-4-16 21:41 标题: Re: 求助：生存分析方法的选择？ When no ties exist in the survival times, the likelihood score test in the Cox regression analysis is equivalent to the log-rank test. on the other hand, it is no sense to compare the both if more than one covariate was involved in the PH model since log-rank test in proc lifetest cannot take any adjustment. Overall, if there are no ties and only one class variable included in your model, the result shoud be consistent.
JingJu作者: shiyiming 时间: 2010-4-18 21:03 标题: Re: 求助：生存分析方法的选择？ 我看不明白，能说的简单点吗？用中文好吗？