ABSTRACT
This chapter deals with sample size determination for the purpose of testing hypotheses. Sample size calculations based on Bayesian concepts are discussed through three examples, one related to product lot release and the other two regarding futility and interim analysis in sequential trials. Sample size determination is an important aspect in study design, especially in the planning of clinical trials due to regulatory requirements. The “decision-theoretic Bayesian” approach determines a sample size that maximizes a utility function for the cost of the study and potential benefit of the new intervention or improved process. Futility represents an important extension of the Bayesian viewpoint of sample size/power calculations. The term “futility” refers to unlikeliness for the trial to achieve statistical significance for its primary endpoints. The statistical power of a test is the long-term probability to declare Ha using a statistical decision rule, conditioned on the value of population parameters.
