Abstract:
|
Lalonde et al. (2007) proposed a model-based approach to drug development, which emphasized quantitative criteria and trial performance metrics for Go/No-Go decisions. This talk will outline some of the major advantages of this approach, including the flexibility to use multiple decision criteria within a frequentist or Bayesian framework, and generalization to multiple interim looks. Issues of criteria-setting and optimization of metrics will be discussed with examples using statistical software.
|