A graphical approach to sequentially rejective multiple test procedures

Abstract

For clinical trials with multiple treatment arms or endpoints a variety of sequentially rejective, weighted Bonferroni-type tests have been proposed, such as gatekeeping procedures, fixed sequence tests, and fallback procedures. They allow to map the difference in importance as well as the relationship between the various research questions onto an adequate multiple test procedure. Since these procedures rely on the closed test principle, they usually require the explicit specification of a large number of intersection hypotheses tests. The underlying test strategy may therefore be difficult to communicate. We propose a simple iterative graphical approach to construct and perform such Bonferroni-type tests. The resulting multiple test procedures are represented by directed, weighted graphs, where each node corresponds to an elementary hypothesis, together with a simple algorithm to generate such graphs while sequentially testing the individual hypotheses. The approach is illustrated with the visualization of several common gatekeeping strategies. A case study is used to illustrate how the methods from this article can be used to tailor a multiple test procedure to given study objectives.

Serelaxin in addition to standard therapy in acute heart failure: rationale and design of the RELAX‐AHF‐2 study

“…Primary and secondary efficacy outcomes, except all‐cause mortality, will be compared between treatment groups on an intent‐to‐treat basis using a sequentially rejective multiple testing procedure22 controlling the overall two‐sided α at 5%. Treatment groups will be compared regarding the time to CV death through day 180 with a log‐rank test at an initial significance level of (4/5)α, and regarding time to WHF through day 5 at an initial significance level of (1/5)α using a Gehan's generalized Wilcoxon test.…”

Assessing the pattern of covariance matrices via an augmentation multiple testing procedure

“…A similar comment applies for H 0 and H 0 . Hence, a source of possible improvement arises by looking for recent inferential methods that, taking advantage of the hierarchical structure of the hypotheses, allows for a better distribution of α on the various nodes of the tree (see, among others, Bretz et al 2009;Burman et al 2009;Goeman and Finos 2010). Many of these inferential methods can be seen as special cases of the sequential rejection principle of Goeman and Solari (2010), spurring us towards further developments in this direction.…”

Power Calculations in Clinical Trials With Complex Clinical Objectives

“…Unlike the fixed-sequence procedure which uses a restrictive α propagation rule, a general chain procedure offers more flexibility by defining an arbitrary set of transition parameters that govern the process of redistributing the error rate among the non-rejected null hypotheses. Specifically, after a null hypothesis is rejected, the hypothesis weights and transition parameters are updated using the algorithm described in Bretz et al (2009). Popular MTPs such as the Bonferroni, Holm, fixed-sequence and fallback procedures (Wiens, 2003;Wiens and Dmitrienko, 2005) can be constructed using the general chain approach.…”