Hi,

I am looking to power a sub-study within a larger 2 arm randomised study. The outcome is a continuous biomarker.

In the substudy, we will compare the treatment arm (group A) with 2 different subgroups of the control arm: 'early' responders (group B) and 'late' responders (group C). Thus making 2 comparisons ([A vs B] and [A vs C]). We expect the biomarker to be similar in the first comparison (A vs B) and different in the second (A vs C). Expected biomarker levels would be 0.15 (sd 0.05) in groups A and B and 0.25 (0.10) in group C.

Do you have any advice on what would be the best way to run a power calculation for this design in Stata?

Thank you for any help,
Megan