Operational Characteristics of a Go/No-Go Decision Rule in an Early Clinical Trial

• Activity Number: 128
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Atalanta Ghosh*+ and Jose Carlos Pinheiro and Paul Rothenberg
• Companies: Johnson & Johnson and Johnson & Johnson and Johnson & Johnson
• Keywords: Posterior ; prior ; decision rule ; go/no-go ; OC ; Bayesian

Abstract

In an early phase clinical trial for Type 2 diabetes mellitus, 24-hour mean plasma glucose (PG) was used to set up a go/no-go decision rule to decide if the compound should undergo further development. Using published results for a comparator drug, a Bayesian decision rule was developed based on the posterior probability of the drug having a clinically relevant effect. A non-informative prior was used for the parameter in the decision rule, and different go and no-go thresholds for the posterior probability of treatment effect were investigated. The decision rule's operational characteristics (OCs), including frequentist properties, were evaluated via simulation, being presented and discussed in the talk. This technique was applied in a recent early clinical trial, providing the team with greater understanding of the rationale for the go/no-go rule, leading to better decision making.