testing_and_armdropping {BayesianPlatformDesignTimeTrend}

testing_and_armdropping

Description

This function makes a decision on whether any active arm should be dropped based on posterior probability and return the vector of decision on each arm, the vector of active arms index and the number of arms left for further study.

Usage

testing_and_armdropping(
  K,
  armleft,
  post.prob.btcontrol,
  group,
  cutoffeff,
  cutoffful,
  treatmentindex,
  test.type
)

Arguments

K

A numeric value indicating the total number of arm at the beginning of trial including both control and treatment.

armleft

A numeric vector indicating the number of active arms before this interim analysis;

post.prob.btcontrol

A numeric vector of posterior probability of each treatment arm better than control

group

A numeric value. The current stage index.

cutoffeff

A numeric vector of the cutoff value at each stage for efficacy boundary.

cutoffful

A numeric vector of the cutoff value at each stage for futility boundary.

treatmentindex

A numeric vector of the current active treatment arm index

test.type

A character indicating which hypothesis testing we are use. "Oneside": H_0: \pi_k \leq \pi_0; H_0: \pi_k \> \pi_0 "Twoside": H_0: \pi_k \eq \pi_0; H_0: \pi_k \neq \pi_0

Value

A list of information including armleft: the number of active arms after this interim analysis; treatmentindex: the index vector of active arm after this interim analysis; stats3: the vector of conclusion on whether null hypothesis is rejected

Examples

testing_and_armdropping(
K = 4,
armleft = 4,
post.prob.btcontrol = c(0.5,0.99,0.02),
group = 3,
cutoffeff = c(1, 0.99, 0.975, 0.96, 0.95),
cutoffful = c(0, 0.01, 0.025, 0.04, 0.05),
treatmentindex = c(1,2,3),
test.type = "Oneside")

testing_and_armdropping(
K = 4,
armleft = 4,
post.prob.btcontrol = c(0.5,0.99,0.02),
group = 3,
cutoffeff = c(1, 0.99, 0.975, 0.96, 0.95),
cutoffful = c(0, 0.01, 0.025, 0.04, 0.05),
treatmentindex = c(1,2,3),
test.type = "Twoside")