Abstract:
|
Quantitative approaches are useful for informing Go/No-Go (GNG) decision in proof-of-concept (PoC) trials. Because of the inherent small sizes of these trials, emanating treatment effect estimate are at risk of high random variation. Mitigating this possible high random variation calls for efficient study design and method of analysis. Using a clinical trial in kidney disease for example, we demonstrate that study design and analysis methods impact GNG decision. We evaluate the impact of design and analysis method on decision under three-outcome GNG decision framework. Results suggest that efficient study design and analysis methods impact GNG decision outcome. The prime contribution of this work is to remind project team to pay attention to study design and analysis method when evaluating operating characteristics for PoC studies based on GNG criteria; this can translate to trial cost savings.
|