|
Abstract:
|
The statistical methodology for two arm group sequential clinical trials has been available for at least 35 years. The generalization to adaptive two-arm group sequential designs became available only in the last decade thanks to seminal papers by Lehmacher and Wassmer (1999), Cui, Hung and Wang (1999) and Müller and Schäfer (2000). The next stage of development is the generalization of these methods to multi-arm multi-stage (MAMS) group sequential trials. The statistical methodology already developed for the two-arm case can, in principle, be extended to MAMS designs. In practice, however, the formidable computational problems that must be overcome have inhibited making these methods accessible for realistic designs. We will discuss our recent work on overcoming these computational hurdles and will demonstrate the use of these methods for adaptive clinical trials that include early stopping, dropping of losers, and sample size re-estimation within the setting of adaptive dose selection and adaptive population enrichment. Comparisons will be made to alternative approaches that involve p-value combination and closed testing.
|
