论文部分内容阅读
在临床药物有效性和安全性评价中，有时出现小概率的反应结果事件，甚至有观察到０的情况。这类资料不能用Ｘ２检验法作一致性检验。本文建议根据Ｐｏｉｓｓｏｎ原理用１００（１－α）％可信限区间估计法作一致性检验并用实例作了说明。此法也可应用于差异显著性检验。
In the evaluation of clinical drug efficacy and safety, there are occasional events with low probability of reaction and even observed zero. This type of data cannot be checked for consistency using the X2 test. This paper proposes to use the 100(1-α)% confidence interval estimation method based on Poisson’s principle for consistency checking and illustrate it with examples. This method can also be applied to differential significance tests.