Weighted Adaptive Closed Tests for Clinical Trials Testing Complex Multiple Hypotheses

• Activity Number: 278
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Florian Klinglmueller*+ and Martin Posch and Franz Koenig
• Companies: Medical University of Vienna and European Medicines Agency and
• Keywords: Adaptive Designs ; Interim Analysis ; Multiple Testing ; Hypothesis Testing ; Clinical Trials

Abstract

We propose weighted multiple test procedures that permit certain mid-trial design modifications based on internal and external information, e.g., changing the pre-planned sample size, inserting/dropping of treatment groups and endpoints. Both, the pre-planned weighting strategy, as well as potential mid-trial modifications thereof may be specified using intuitive graphical procedures. The approach is based on the closed test principle combined with the conditional error principle. It provides strong control of the family wise error rate. According simultaneous confidence intervals are available. It does not require knowledge of the multivariate distribution of the test statistics. Hence, it is suitable for, e.g. comparisons between combinations of treatment groups and/or multiple endpoints. The procedure may be applied sequentially resulting in an improvement of the pre-planned procedure by taking advantage of the (unknown) correlation between test statistics.