Bayesian prediction modelling for two-stage experimental trials for Poisson or Gamma distributed data

We consider Bayesian prediction modelling to evaluate a satisfaction index after a first phase of experiment in order to decide to stop or continue at the second stage. We apply this method to Poisson and Gamma distributed outcomes in many fields such as reliability or survival analysis for early termination due to either futility or efficacy. We look at two kinds of decisions making: an hybrid Bayesian-frequentist or a full Bayesian approach.

为决定是否进入第二阶段实验，我们采用贝叶斯预测模型（Bayesian prediction modelling）对第一阶段实验后的满意度指数进行评估。我们将该方法应用于多个领域中服从泊松分布（Poisson）和伽马分布（Gamma）的结果变量，例如可靠性分析或生存分析中因无效性或有效性导致的实验提前终止场景。我们考察了两种决策方式：贝叶斯-频率论混合方法以及纯贝叶斯方法。