cor.0: Predicted cross-test correlation

In lilywang1988/GSMC: Group Sequential Design for Maxcombo tests

Description

These two functions are to predict the correlation between two weighted log-rank tests at certain time t.star under either the null hypothesis (using cor.0) or the alternative hypothesis (using cor.1).

Usage

cor.0(rho1, gamma1, rho2, gamma2, lambda, R, p, t.star)

cor.1(rho1, gamma1, rho2, gamma2, lambda, theta, eps, R, p, t.star)

Arguments

rho1	First power parameters for the two Fleming-Harrington weights, defined for covariance calculation.
gamma1	Second power parameters for the two Fleming-Harrington weights, defined for covariance calculation.
rho2	First power parameters for the two Fleming-Harrington weights, defined for covariance calculation.
gamma2	Second power parameters for the two Fleming-Harrington weights, defined for covariance calculation.
lambda	Event hazard for the control arm.
R	End of the accrual period.
p	Treatment assignment probability.
t.star	Time point we pause the study to check the cumulative information under the null.
theta	Hazard ratio after the change point (before the change point HR should be 1).
eps	Change point.

Details

These two functions are designed to calculate the predicted correlation between the two weighted log-rank tests at time t.star under the two hypotheses. The null hypothesis is an exponential distribution for both the treatment and control arms with hazard lambda, while the alternative hypothesis has the control group following an exponential distribution with hazard lambda, and the treatment group following a piece-wise exponential distribution with hazard lambda before eps, but a hazard theta times lambda after eps.

Author(s)

Lili Wang.

References

Hasegawa, T. (2016). Group sequential monitoring based on the weighted log‐rank test statistic with the Fleming–Harrington class of weights in cancer vaccine studies. Pharmaceutical statistics, 15(5), 412-419.

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