sample size calculation randomizd controlled trial

[Sugi Jeyaganth]

Dear Statalisters,
I am using stata 13, I want to calculate randomized controlled trial
sample size.
we are expected Sample size between 80 to 120 patients
sample size must be calculated on the basis of the following parameters:
· Median Progression Free Survival control group: 9 months,
Median Progression Free Survival for treatment group: 12.6 months (40% improvement)
· 2 years accrual, 2 years Follow up
· Drop-out rate 10% in total (over 4 years)
· Power 80%
· 1-sided p=0.05

to calculate the sample size,
stpower cox,failprob(0.6) wdprob(0.10) p(0.8) al(0.05) onesided

I use above stata syntax , I am not sure this is correct, I do not include the
median progression free survival in the syntax.
please advice me , am I using the information correctly or is there are other stata command

Thanks
Kind regards
Sugi jey

[Marcos Almeida]

Hello Sugi,

Since you're using Stata 13, you may also prefer to browse the options on the drop-down menu. I believe this is the easiest way to delve into these commands.

Also, here you have the reference manual over sample size and power estimations: http://www.stata.com/manuals13/pss.pdf

I don't know what was the rationale for you to select a one-sided test. However, at least in my field and practice, two-sided tests are usually preferred.

Hopefully this helps.

Best,

Marcos

[Sugi Jeyaganth]

Thanks very much for your reply Dr.Marcos,
My researcher ask me for one-sided test.
also I was reading example from the LACHINE (1981,107), they also use one sided exponential test,
based on the difference between hazards, of the superiority of new treatment.
I did read the "pss"
I don't see any example using sample size calculation for the median survival.
from the theory i guess i have to use "stpower exponential",
i am don't sure i input my numbers correctly.

stpower exponential 0.4 , t(3.6) power(0.8) onesided aperiod(2) fperiod(2) losshaz(0.05 0.05) detail

Thanks
Sugi