sample.size_FH: Sample size calculation for Fleming-Harrington weighted...

In lilywang1988/GSMC: Group Sequential Design for Maxcombo tests

Description

Sample size calculation for Fleming-Harrington weighted log-rank tests This sample size calculation method was proposed by Hasegawa (2014). This function is to calculate the sample size for Fleming-Harrington weighted log-rank tests with piece-wise exponential distributed survival curves in described in Hasegawa(2014).

Usage

sample.size_FH(
  eps,
  p,
  b,
  tau,
  omega,
  lambda,
  lambda.trt,
  rho,
  gamma,
  alpha,
  beta
)

Arguments

eps	The change point, before which, the hazard ratio is 1, and after which, the hazard ratio is theta
p	Treatment assignment probability.
b	The number of subintervals per time unit.
tau	The end of the follow-up time in the study. Note that this is identical to T+τ in the paper from Hasegawa (2014).
omega	The minimum follow-up time for all the patients. Note that Hasegawa(2014) assumes that the accrual is uniform between time 0 and R, and there does not exist any censoring except for the administrative censoring at the ending time τ. Thus this value omega is equivalent to tau-R. Through our simulation tests, we found that this function is quite robust to violations of these assumptions: dropouts, different cenosring rates for two arms, and changing accrual rates.
lambda	The hazard for the control group.
lambda.trt	The hazard for the treatment group after time eps.
rho	The first parameter for Fleming Harrington weighted log-rank test:W(t)=S^ρ(t^-)(1-S(t^-))^γ.
gamma	The second parameter for Fleming Harrington weighted log-rank test:W(t)=S^ρ(t^-)(1-S(t^-))^γ.
alpha	Type I error.
beta	Type II error.

Value

n	The needed sample size.
n_event	The needed event numbers for both arms together.
E.star	The unit mean, correspoinding to E^* in Hasegawa(2014)
sum_D	The cumulative D, and ceiling(n*D) is quivalent to n_vent.

Note

This function is based on a R function from Dr. Ting Ye's paper : Ye, T., & Yu, M. (2018). A robust approach to sample size calculation in cancer immunotherapy trials with delayed treatment effect. Biometrics, 74(4), 1292-1300.

Author(s)

Lili Wang, Ting Ye

References

Ye, T., & Yu, M. (2018). A robust approach to sample size calculation in cancer immunotherapy trials with delayed treatment effect. Biometrics, 74(4), 1292-1300. Hasegawa, T. (2014). Sample size determination for the weighted log‐rank test with the Fleming–Harrington class of weights in cancer vaccine studies. Pharmaceutical statistics, 13(2), 128-135.

Examples

## Not run:  
# Example 1 from Hasegawa (2014)
p<-2/3
tau<-66
omega<-18
eps<-6
m1=21.7  #median survival time for placebo group
m2=25.8  # median survival time for treatment group
lambda<-log(2)/m1
lambda.trt<-log(2)*(m1-eps)/(m2-eps)/m1
theta=lambda.trt/lambda
alpha<-0.025
beta<-0.1
rho=0
gamma=1
b=30
sample.size_FH(eps,p,b,tau,omega,lambda,lambda.trt,rho, gamma,alpha,beta)$n
#1974, identical to the paper's report

## End(Not run)

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