I'm looking to calculate the sample size needed for a three armed trial or the power detectable with the current sample size where a survival analysis will be the primary outcome. Can anyone advise how to do this or where to look to figure it out?
Another potential technique is to undertake the sample size calculation for the two arm likely to have the closest survival, but setting alpha at (say) 0.03 to allow for multiple testing. Then multiply the total sample size by 3/2.