topic Re: PROC POWER for Cox regression in Statistical Procedures

PROC POWER for Cox regression

TomHsiung — Fri, 16 May 2025 14:33:52 GMT

Hello, guys

You suggestion is precious.

To estimate the sample size, we need a value of R^2. I don't know how do we get that R^2.

An example script is:

proc power;
   coxreg
      hazardratio = 1.6
      rsquare = 0.2
      stddev = 1.1
      power = 0.9
      eventstotal = .
   ;
run;

The definition of R^2 is:

Re: PROC POWER for Cox regression

Ksharp — Thu, 22 May 2025 00:59:59 GMT

I think it is just like R-Square of PROC REG and GLM .
r-square= 1 - uss(residual)/css(y) ;

https://support.sas.com/kb/60/162.html

Re: PROC POWER for Cox regression

SteveDenham — Wed, 21 May 2025 18:42:17 GMT

One nice feature of PROC POWER is its ability to give power or sample size estimates for a variety of values of the various parameters. In this case, you may want to look at several values for Rsq, to see how sensitive sample size (or power) is to this value. If it turns out that the result is not too sensitive to changes in Rsq, then you can use the standard formula that @Ksharp gave. If it is otherwise, then you need to look at other estimators for Rsq that reduce bias.

SteveDenham

Re: PROC POWER for Cox regression

TomHsiung — Sun, 25 May 2025 08:58:51 GMT

Thank you both @Ksharp and @SteveDenham I think I need to supplement my mathematical statistics basis before I can dig deeper into this issue.

Regards,