mHP.b

Compute the efficacy boundary (modified Haybittle-Peto) for the
first two stages

Clinical trial design for subgroup selection in three-stage group
sequential trial as described in Lai, Lavori and Liao (2014,
<doi:10.1016/j.cct.2014.09.001>). Includes facilities for design,
exploration and analysis of such trials. An implementation of
the initial DEFUSE-3 trial is also provided as a vignette.

Balasubramanian Narasimhan

ASSISTant

Adaptive Subgroup Selection in Group Sequential Trials

Tze Leung Lai

Philip Lavori

Olivia Liao

Ka Wai Tsang

mHP.b function

<dl><dt>prevalence</dt>
<dd>the vector of prevalences between 0 and 1 summing
to 1. \(J\), the number of groups, is implicitly the length
of this vector and should be at least 2.</dd>
<dt>N</dt>
<dd>a three-vector of total sample size at each stage</dd>
<dt>cov.J</dt>
<dd>the 3 x 3 covariance matrix for Z_J at each of the
three stages</dd>
<dt>mu.prime</dt>
<dd>a list of \(J\) mean vectors, each of length
\(J-1\) representing the conditional means of all the other
\(Z_j\) given \(Z_i\). This mean does not account for the
conditioned value of \(Z_i\) and so has to be multiplied by
that during use!</dd>
<dt>Sigma.prime</dt>
<dd>a list of \(J\) covariance matrices, each
\(J-1\) by \(J-1\) representing the conditional covariances
all the other \(Z_j\) given \(Z_i\)</dd>
<dt>alpha</dt>
<dd>the amount of type I error to spend</dd>
<dt>btilde</dt>
<dd>the futility boundary</dd>
<dt>theta</dt>
<dd>the effect size on the probability scale</dd></dl>

Arguments

Compute the efficacy boundary (modified Haybittle-Peto) for the
first two stages — mHP.b

<dl>

<dt>prevalence</dt>
<dd>the vector of prevalences between 0 and 1 summing
to 1. \(J\), the number of groups, is implicitly the length
of this vector and should be at least 2.</dd>


<dt>N</dt>
<dd>a three-vector of total sample size at each stage</dd>


<dt>cov.J</dt>
<dd>the 3 x 3 covariance matrix for Z_J at each of the
three stages</dd>


<dt>mu.prime</dt>
<dd>a list of \(J\) mean vectors, each of length
\(J-1\) representing the conditional means of all the other
\(Z_j\) given \(Z_i\). This mean does not account for the
conditioned value of \(Z_i\) and so has to be multiplied by
that during use!</dd>


<dt>Sigma.prime</dt>
<dd>a list of \(J\) covariance matrices, each
\(J-1\) by \(J-1\) representing the conditional covariances
all the other \(Z_j\) given \(Z_i\)</dd>


<dt>alpha</dt>
<dd>the amount of type I error to spend</dd>


<dt>btilde</dt>
<dd>the futility boundary</dd>


<dt>theta</dt>
<dd>the effect size on the probability scale</dd>

</dl>

mHP.b: Compute the efficacy boundary (modified Haybittle-Peto) for the first two stages

Description

Usage

Arguments