|
Abstract:
|
Potential sources of multiplicity in adaptive seamless designs and their implications on the control of type I errors and decision making will be discussed. Repeated testing of the same hypotheses in a group sequential setting with early rejection or acceptance of hypotheses at interim analyses, testing of multiple initial hypotheses (e.g., multiple comparisons with a control), and change of design features (e.g., sample size) may need to be taken into account simultaneously. For each of these cases methods exist to account for multiplicity. Dealing with a combination of them within an adaptive trial often can be done in different ways. Options will be discussed together with their performance characteristics and possible operational restrictions.
|
