Abstract:
|
Traditional efficacy interim design is based on alpha spending which doesn't have intuitive interpretation and hence is difficult to communicate with non-statistician colleagues. The alpha spending approach is based on efficacy alone and hence doesn't have the flexibility to incorporate newly emerged safety signal. Newly emerged safety signal may nullify the originally set efficacy boundary. In contrast, the probability of success (POS) concept has intuitive interpretation and hence can facilitate our communication with non-statistician colleagues and help to obtain health authorities (HA) buying. The success criteria of POS is not restricted to statistical significance. Hence POS has the capability to incorporate both efficacy and safety information. We propose to use POS and its credible interval to design efficacy interim. In the proposed method, the efficacy boundary is adjustable to offset newly emerged safety signal. (reference: http://www.tandfonline.com/doi/full/10.1080/10543406.2014.983646)
|