All Times EDT
Keywords: bayesian, probability of success, copula, seemingly unrelated regression, power prior
Given the cost and duration of phase III and phase IV clinical trials, the development of statistical methods for go/no-go decisions is vital. Most methods to compute probability of success are only for single endpoints e.g., Ibrahim et al. (2015). Methods dealing with multiple responses, which are common in clinical trials, are sparse, particularly for non-Gaussian endpoints. In this talk, we present novel techniques to compute probability of success (POS) for multiple endpoints under Gaussian and non-Gaussian assumptions. The tools developed will enable practitioners to compute the likelihood of a successful clinical trial for primary and key secondary endpoints.