O/E法的检验效能和样本含量估计

Study on the Power of O/E Method

为探讨计算Ｏ／Ｅ法检验效能的方法和原理，提出Ｏ／Ｅ法样本含量的估计方法，利用Ｐｏｉｓｓｏｎ分布及假设检验的原理，求出不同基线率π０、增长因子γ及样本含量下的检验效能。结果显示，检验效能１－β随基线率、样本含量、增长因子的增加而增加，且随着样本含量的增加，Ｏ／Ｅ法的检验效能的增加是不连续的，跳跃的，有时会有回复。计算Ｏ／Ｅ法检验效能的方法和原理，同样适用于Ｐｏｉｓｏｎ分布资料的假设检验。Ｏ／Ｅ法样本含量的估计方法，亦可用于疾病干预性研究中的样本大小的估计。

This paper addresses the methodology of calculating the power of O/E method and introduces the relevant algorithm for estimating the sample size.Based on the principles of hyporthesis testing and Poisson distribution,the power can be worked out for given sample size,baseline rate(π0) and the factor of increasing (γ).The power increases with the increase in sample size,π0 and γ.In shape the power is not a continuous smooth line but a zigzag line.The method illustrated in this paper can be also used for estimaging the sample size of studies on the intervention of diseases.