标题: 怎样用SAS PROCEDURE POWER 来计算生存数据的样本量 [打印本页] 作者: shiyiming 时间: 2009-8-25 09:13 标题: 怎样用SAS PROCEDURE POWER 来计算生存数据的样本量 在为生存数据做试验设计时,需要计算样本量的大小,下面是一些基本的假设:
试验包含两个组: Placebo group and Vitamin D group, each group has the 6 months dose, the study time point is : month 3, month 6, month 9.
Set alpha=0.05, power=0.8;
1)For Placebo group, assuming the hazard rate is 0.36 at each study point;
2)For Vitamin D group, assuming the hazard rate is 0.15 at each study point;
3) for both group, assuming the rate of losting to follow-up is 0.30 at each study point,
我打算选用SAS 的PROCEDURE POWER,但是只找到当失访率(lost to follow up at each treatment interval)服从指数分布的时候的OPTION(见下面的CODE),那么当失访率服从其他分布或者是一个稳定的数值时,还能使用POWER来计算吗? 另外,这是我初次使用POWER 程序,还请各位高手看一下下面的代码是否正确,谢谢。
proc power;
twosamplesurvival test=logrank
curve("Placebo") = (3 to 9 by 3): (0.64 0.4096 0.262144)
curve("Vitamin D") =(3 to 9 by 3): (0.85 0.7225 0.614125)
groupsurvival = "Placebo" | "Vitamin D"
curve("Placebo1")= 3: 0.30 /*assuming the loss curve is the exponential loss curve*/
curve("Vitamin D1")=3: 0.30
grouploss = "Placebo1" | "Vitamin D1"
accrualtime = 6
followuptime =3
alpha=0.05
sides=2
power = 0.8
npergroup = .;
run;作者: shiyiming 时间: 2009-8-26 03:26 标题: Re: 怎样用SAS PROCEDURE POWER 来计算生存数据的样本量 我们光顾着学习你的程序了,哪里还有什么指正?作者: shiyiming 时间: 2009-8-26 08:30 标题: Re: 怎样用SAS PROCEDURE POWER 来计算生存数据的样本量 我也是第一次用POWER 来求样本量,很多参数不明白,如果有朋友熟悉这个算法,请多多指教,我也会努力把它搞清楚,最后把正确的CODE给大家发上来,谢谢.