I have to calculate sample size for adaptive clinical trial design (Multi arm multi stage MAMS) in which several interventions will be tested (superiority). This is a time-to-event trial in which I know that: - only 20% of the patients in placebo will present the outcome (e=1) at 7-days of follow-up - I hypothesized that the effect of having an increase of 70% the prob of having the outcome is an interesting effect (HR=1.7). However, an effect of 10% increased will be depreciable (H0=1.1)
With this information I can build the function (tite.mams in the MAMS R package) setting the beta=0.8 and the alpha=0.025. Doing that I have:
These means that I need 39 events in the placebo arm. Does it means that I need 39*100/20 patients on this arm?
Also, I am thiniking in adding new arms at the middle of the trial, do I have to count for the number of arms (including the potential addings) before sample size calculation?