|
Abstract:
|
Statistical methods in clinical trial monitoring have been developed in many applications. With the recent surge of adaptive and complex innovative design programs, fixed design is rarely the case anymore. Pre-planned interim analyses and monitoring are routinely done in practice. The goal is to develop safe and effective products and serve patients as fast as possible by using all the knowledge gained in research and development. Futility due to safety concerns and suboptimal therapeutic effects should be carefully examined as well. We will look at some decision metrics for these interim analyses, in relation to Bayesian sample size determination and simulation-based design. Bayesian design is flexible and should be transparent. The proposed sampling approach and decision criterion attempt to connect with clinical trial design and implementation.
|
