R help - Apr 2011 - Question abt power calculation in multistrata data using Hmisc or other package

Hi,
I have two unequal patient groups belonging to different
geographic locations, treated by different drug regimens and their
survival data. N1=150 and N2=60

In each group of data there are
three strata for one independent variable- treatment status (2=two
drugs administered, 1=only one conventionally used drug administered,
0=no drug administered), dead or not and months to death. In order to
calculate the statistical power necessary for carrying out this
analysis, i tried to find appropriate function in Hmisc package. But it
seems that it is meant for two groups such as reference/control versus
cases. 

This is the sample of data in each of the two groups:

Treatment_stat Death MonthsToEvent
1 0 60.0
2 1 55.3
0 1 36.7
0 1 39.9

number of cases in treatment strata : 
group 1: n12=110; n11=25;n10=15 (N2=150)
group 2: n22=43;n21=8;n20=9 (N2=60) 

Coxph
output indicates significant different survival in group 1 data but not
in group 2. How can one be sure that the lack of significance of
survival difference is true in group 2 and it is not because of fewer
number of cases in group 1 ?
Is there any function for this type of calculations in Hmisc or some other R
package.

-Vickie
 		 	   		  
	[[alternative HTML version deleted]]