pickwin: Two-Stage Randomized Phase II Designs
In raredd/desmon: Functions for design and monitoring of clinical trials

pickwin

R Documentation

Two-Stage Randomized Phase II Designs

Description

Calculates operating characteristics of pick the winner rules for two stage randomized phase II designs

Usage

pickwin(n1, n2, p1, p2, r1, r2, r3 = r1, r4 = r2)

Arguments

`n1`	first stage sample size (assumed to be the same for both treatments)
`n2`	addition accrual during the second stage (assumed to be the same for both treatments)
`p1`	True response probability on treatment 1
`p2`	True response probability on treatment 2
`r1`	The maximum number of responses to declare treatment 1 inactive in the first stage
`r2`	The maximum number of total responses over both stages to declare treatment 1 inactive
`r3`	The maximum number of responses to declare treatment 2 inactive in the first stage
`r4`	The maximum number of total responses over both stages to declare treatment 2 inactive

Details

Assumes that during the first stage of accrual n1 patients are randomized to each treatment (RX1 and RX2). If more than r1 (r3) responses are seen on treatment RX1 (RX2), then an additional n2 patients will be accrued to treatment RX1 (RX2). If more than r2 (r4) total responses are seen on RX1 (RX2), then RX1 (RX2) will be declared active.

Let X1 and X2 be the number of responses observed on treatments 1 and 2. This function computes and prints P(X1<=r2)=P(RX1 declared inactive), P(X2<=r4)=P(RX2 declared inactive), P(X2>r4 and X2>X1)=P(RX2 declared superior), P(X1>r2 and X1>X2)=P(RX1 declared superior), and P(X1=X2 or (X1<=r2 and X2<=r4))=P(neither declared superior). Note that P(neither declared superior) includes both the probability that both are declared inactive and the probability that both are declared active and are equal.

Value

Returns a vector of length 9, whose components are (in order) P(X1<=r1, X2>r4), P(r1<X1<=r2, X2>r4), P(X1>r2, X2>X1, X2>r4), P(X2<=r3, X1>r2), P(r3<X2<=r4, X1>r2), P(X1>r2, X2>r4, X1>X2), P(X1<=r2, X2<=r4), P(X1>r2, X2>r4, X1=X2), and the sum of the first 8 terms.