BioInfo.Power {ADCT}

Power calculation for Biomarker-Informed Design with Hierarchical Model

Description

Given the Biomarker-Informed design information, returns the overall power and probability of the arm is selected as the winner.

Usage

BioInfo.Power(uCtl, u0y, u0x, rhou, suy, sux, rho, sy, sx, Zalpha, N1, N, nArms, nSims)

Arguments

uCtl

mean value for the control group.

u0y

mean parameter of the group 1 for the parent model.

u0x

mean parameter of the group 2 for the parent model.

rhou

correlation coefficient between two groups for the parent model.

suy

standard deviation of the group 1 for the parent model.

sux

standard deviation of the group 2 for the parent model.

rho

correlation coefficient between two groups for the lower level model.

sy

standard deviation of the group 1 for the lower level model.

sx

standard deviation of the group 2 for the lower level model.

Zalpha

crtical point for rejection.

N1

sample size per group at interim analysis.

N

sample size per group at final analysis.

nArms

number of active groups.

nSims

number of simulation times.

Value

The evaluated power and probability of selecting the arm as the winner.

Author(s)

Yalin Zhu

References

Chang, M. (2014). Adaptive design theory and implementation using SAS and R. CRC Press.

Examples

## Determine critical value Zalpha for alpha (power) =0.025
u0y=c(0,0,0); u0x=c(0,0,0)
BioInfo.Power(uCtl=0, u0y, u0x, rhou=1, suy=0, sux=0, rho=1, sy=4, sx=4,
 Zalpha=2.772, N1=100, N=300, nArms=3, nSims=1000)
## Power simulation
u0y=c(1,0.5,0.2)
u0x=c(2,1,0.5)
BioInfo.Power(uCtl=0, u0y, u0x, rhou=0.2, suy=0.2, sux=0.2, rho=0.2, sy=4, sx=4,
 Zalpha=2.772, N1=100, N=300, nArms=3, nSims=500)