schoenfeld: Number of events or power calculation with the 'Schnoenfeld...

In Knusprikus/BSSRed: Blinded Sample Size Reestimation For Event Driven Trials

Description

Calculates the number of events needed to achieve a targeted power (or the power given the number of events) in a two group scenario for all combinations of provided hazard ratios theta, balance between the groups k and the alpha niveaus alpha.

Usage

nschoenfeld(theta, k = 1, alpha = 0.05, beta = 0.2, alternative = "two-sided")

pschoenfeld(theta, nEvents, k = 1, alpha = 0.05, alternative = "two-sided")

Arguments

theta	vecor of hazard ratios.
k	positive vektor. Defines the relative group sizes with the relation k:1. If k=1, equal group sizes are assumed (Note: that the formula is symmetric around k=1).
alpha	number between 0 and 1 defining the deserved type I error rate. Default is 0.05.
beta	number between 0 and 1 defining the deserved type II error rate. Default is 0.1 resulting in a power of 1-0.1=0.9 (90%).
alternative	a string; either 'one-sided' or 'two-sided'. Default is 'two-sided'.
nEvent	vecor of number of events.

Value

A data frame containing the input values and calculated number of events needed to satisfy the given assumptions or the power achieved with the given number of events.

References

Schoenfeld, D. (1983). Sample-Size Formula for the Proportional-Hazards Regression Model. Biometrics, 39(2), 499-503. doi:10.2307/2531021

Examples

# for a assumed hazard ratio of theta = .7
# in a design with a treatment to control group allocation of 2:1 resulting in k = 2
# power of 1-beta=90%
# and a two sided significance niveau of alpha = .05
x <- nschoenfeld(theta = .7, k=2, alpha=.05, beta=.1, alternative = "two-sided")


# with 20 events less the power would be
pschoenfeld(theta=.7, nEvents=x$nEvents-20, k=2, alpha=.05, alternative = "two-sided")

Knusprikus/BSSRed documentation built on July 6, 2020, 11:02 p.m.