|
Abstract:
|
If a treatment effect in the overall population is marginally significant or not statistically significant, can a promising treatment effect in a subgroup pave a regulatory path for approval? What kind of scientific and statistical evidence is important for a successful regulatory negotiation? What are the main concerns about promising subgroup results? This presentation will use real examples to illustrate the key elements for a convincing case and effective ways of presenting statistical arguments. Probability of random chance is the top concern about any subgroup results and most statistical arguments are intended to address this concern. Credibility score, pre-specification, and multiple comparison issues are important concepts to be discussed. Any strong statistical arguments without good biological support are destined for failure.
|
