Sequential parallel comparison design with two coprimary endpoints

Abstract

A placebo-controlled randomized clinical trial is required to demonstrate that an experimental treatment is superior to its corresponding placebo on multiple coprimary endpoints. This is particularly true in the field of neurology. In fact, clinical trials for neurological disorders need to show the superiority of an experimental treatment over a placebo in two coprimary endpoints. Unfortunately, these trials often fail to detect a true treatment effect for the experimental treatment versus the placebo owing to an unexpectedly high placebo response rate. Sequential parallel comparison design (SPCD) can be used to address this problem. However, the SPCD has not yet been discussed in relation to clinical trials with coprimary endpoints. In this article, our aim was to develop a hypothesis-testing method and a method for calculating the corresponding sample size for the SPCD with two coprimary endpoints. In a simulation, we show that the proposed hypothesis-testing method achieves the nominal type I error rate and power and that the proposed sample size calculation method has adequate power accuracy. In addition, the usefulness of our methods is confirmed by returning to an SPCD trial with a single primary endpoint of Alzheimer disease-related agitation. KEYWORDSbivariate normally distributed endpoint, enrichment design, intersection-union test, placebo nonresponder, placebo responder Pharmaceutical Statistics. 2019;1-12. wileyonlinelibrary.com/journal/pst

A generalized weighted combination test of treatment effect for clinical trials with a sequential parallel comparison design and binary endpoint

“…• For patients with observed data for X i = 1 and missing data for Y i , we draw p 11 ∕p 1. from (6), then draw…”

“…In general, the SPCD consists of two consecutive stages and three treatment groups: active treatment for both stages, placebo for stage 1 and active treatment for stage 2, and placebo for both stages. There are different versions of the SPCD for different considerations 4‐13 . For example, there are two different SPCD randomization schemes: one is a rerandomization, and another is a prerandomization scheme.…”

“…There are different versions of the SPCD for different considerations. [4][5][6][7][8][9][10][11][12][13] For example, there are two different SPCD randomization schemes: one is a rerandomization, and another is a prerandomization scheme. Chen et al, 10 Rybin et al, 11 and Silverman et al 12 propose to unblind trial data after stage 1.…”

Sample size calculation in clinical trials with two co‐primary endpoints including overdispersed count and continuous outcomes

“…Some researchers have proposed sample size calculation methods in the context of co‐primary endpoints for various outcomes, including continuous outcomes, 21 binomial outcomes, 22 mixed continuous and binomial outcomes, 23 and survival outcomes 24 . These methodologies have also been extended to more complex trial designs, such as a sequential parallel comparison design with two co‐primary continuous endpoints 25 and group‐sequential trials with two co‐primary continuous endpoints 26 . Sample size calculation for co‐primary endpoints has been reviewed in further detail elsewhere 27 .…”