6.3 Combination des deux approches

     

Les deux approches de la hiérarchisation et de la répartition sont assez fréquemment combinées. Par exemple, lorsque deux doses du nouveau traitement sont évaluées dans un essai (essai à 3 bras) le risque alpha global est d’abord réparti entre les 2 doses (2.5% pour chaque dose par exemple) et ensuite, pour chaque dose, une série de critères de jugement est testée de manière séquentielle.

A Bonferroni approach (splitting the overall α between the two dose levels of verubecestat) in conjunction with a hierarchical sequential testing approach of outcomes was used to control for the type 1 error rate, with testing of outcomes as described in the order listed in the Outcomes section.

Separately for each dose level, if significant superiority was not shown, all subsequent outcomes were assumed not to have differed significantly between the groups [ 10.1056/NEJMoa1812840 ]

Dans d’autres cas, les 2 doses peuvent être testées de manière complètement hiérarchique

Hypotheses were tested in the following order: the 20-mg cannabidiol group, followed by the 10-mg cannabidiol group, was compared with the placebo group with respect to the primary outcome; the 20-mg cannabidiol group was then compared with the placebo group with respect to each key secondary outcome in the order listed above, and then the 10-mg cannabidiol group was compared with the placebo group with respect to each key secondary outcome in the same order [ 10.1056/NEJMoa1714631 ]