0
$\begingroup$

We want to conduct a multi-center randomized clinical trial with two arms (2 treatments : a superiority trial) stratified on three centers (three sites).

The primary outcome variable is a continuous variable (presented as mean, sd). How can i estimate the sample size needed for this stratified trials (for each center)?

I found many packages in R that deal with sample size estimation, but i do not I did not find any that is suitable for this analysis.

Thank you

Improve this question
$\endgroup$
1
  • $\begingroup$ I am also looking for the answer to this problem. Have you found the answer yet? $\endgroup$
    – user124406
    Commented Jul 25, 2016 at 10:21

1 Answer 1

Reset to default
1
$\begingroup$

If you are assuming the treatment effect size and the sd of the outcome are the same in all three sites, then you don't need to account for center in your power calculation.

This doesn't mean that the stratified randomization isn't likely to be improving your power in actual practice. Stratified randomization, matched randomization, minimization, and other forms of restricted randomization all increase your power through balancing (among the treatment arms) the distribution of covariates that are predictive of the outcome. However in your case, you are already accounting for this benefit if you are assuming a common sd for the outcome within each site. The sensibility of that assumed sd is what will largely drive the sensibility of your power calculation.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Thank you. Please, could you provide me with a reference. I have been searching for this for about a week. $\endgroup$
    – user1594303
    Commented Jan 4, 2016 at 21:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.