Maxcombo.beta.n: The power for a vector of sample sizes

In lilywang1988/GSMC: Group Sequential Design for Maxcombo tests

Description

To obtain a spectrum of power for a vector of numbers of subjects (n) using Maxcombo.beta.n or events (d) using Maxcombo.beta.d.

Usage

Maxcombo.beta.n(Sigma1, mu1, z_alpha_vec, interim_vec, R, n_seq, n.rep = 5)

Maxcombo.beta.d(
  Sigma1,
  mu1,
  z_alpha_vec,
  interim_vec,
  R,
  d_seq,
  sum_D,
  n.rep = 5
)

Arguments

Sigma1	the correlation matrix under the alternative hypothesis.
mu1	the unit mu under the alternative hypothesis (the mean of the expectation of each subject scaled weighted log-rank test statistic, which can be approximated using the formula for E^* in Hasegawa 2014 paper. ).
z_alpha_vec	same as the one exported from Maxcombo.bd, which is the boundaries for ordered test statistics, its order should be consistent to the rows and columns in Sigma1.
interim_vec	the vector of the interims in each stages, not that it should be a repeat vector with same iterim values for all the test statitics at same stages.
R	end of the enrollment time, which is identical to R defined in other functions like I.1.
n_seq	the sequence of number of patients.
n.rep	number of repeats to take the median for output
d_seq	the sequence of number of expected events.
sum_D	same as the exported value from sample.size_FH, the summed D^* in Hasegawa (2014).

Author(s)

Lili Wang

See Also

Maxcombo.sz

Examples

Sigma0_v<-rep(0.5,6)
Sigma0<-matrix(1, ncol=4,nrow=4)
Sigma0[upper.tri(Sigma0)]<- Sigma0_v
Sigma0[lower.tri(Sigma0)]<- t(Sigma0)[lower.tri(Sigma0)]
Sigma0
alpha_stage <- c(0.01,0.025) # The error you would like to spend at the interim stage
zz <- Maxcombo.bd(Sigma0 = Sigma0,index=c(1,1,2,2),alpha_sp=alpha_stage)
zz$z_alpha # boundary value for each stage
zz$z_alpha_vec # boundary value for each test statistic corresponding 
to the index

lilywang1988/GSMC documentation built on March 9, 2021, 5:25 p.m.