|
Abstract:
|
In Early Clinical Development at AstraZeneca we apply a consistent statistical approach to go/no-go decisions. We choose a key endpoint or two, and the sample space for the corresponding point estimate is split into three zones: red, amber and green. The amber zone represents inconclusive evidence, where our senior leaders must use their impression of the data, as well as external factors, to come to a final decision. Decision criteria are set by the team at the design stage. To simplify this task, we have worked with Cytel to develop a software tool called dECiDe, which performs the necessary calculations and produces relevant tables and figures. Outputs have a consistent format that can be presented to senior leaders at internal governance bodies. It is sometimes helpful to quantify uncertainty in frequentist terms, and we will often need to assess the impact of interim analyses on the trial's operating characteristics. At other times we naturally make Bayesian probability statements, and may wish to formally combine evidence from different sources. dECiDe gives us flexibility to do all of this.
|
