rstudio gsDesign - boundaries and alpha spending functions

To whom it may concern

Randomized phase III trial with two arms (E: experimental and C: active control) to assess superiority of E over C on a time-to-event mortality endpoint.
The single stage design: assuming an annual event rate of 14.5% in arm C, an enrollment period of 22 months, and a minimum follow-up of 21 months, calculate the required sample size n to detect a hazard reduction of 15% in E with a power of 97%, α=0.025 one-sided.

I have to extend this previous single stage design to a four-stage GSD by adding three interim analyses on the primary endpoint with efficacy and (non-binding) futility stopping rules according to these assumptions:

  • The interim analyses will be carried out at 33%, 50% and 75% of the sample size.
  • On the z-scale the boundaries are (3.72, 3.09, 3.09, 1.96).
  • On the z-scale the boundaries are (-1.32, 0, 0.5, 1.96).
    I have to calculate:
  • the alpha-spent (αk) at each stage
  • the required sample size
    Teachers suggest to use function gsProbability and gsDesign in R.
    I don't know how to use boundaries on the z-scale ...can anybody help me?


This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.