ABSTRACT
The approach in Chapters 2–7 has used analytic methods with the occasional use of numerical methods when necessary. In practice, using a simulation approach of the type can avoid simplistic assumptions about the prior distribution and the model. In this chapter, we provide two simulation case studies. Both cases are related to interim analyses in clinical trials related to the approaches covered in Chapter 7. The first looks at the use of a Wilcoxon test in the context of a proportional odds model for an adaptive design. We show how to determine the operating characteristics using clinical trial simulation for the design in which the decisions at the interim are either to stop for futility or to increase the sample size based on predictive power. The second case study also uses a predictive approach this time to support decision-making in a trial for which patient recruitment is proving problematic. The sponsor needed to decide whether to stop the study, and consequently abandon the drug development, or to continue if the predictive probability of success was sufficiently high.
