相关差别指标的区间估计方法比较

在公共疾病控制领域,重大稀有疾病的发病率非常低,符合逆抽样特征,量化分析重大稀有疾病的发病率并对其特点进行分析,为了研究在带有群内相关条件下的整群抽样问题,通过β-二项分布抽样对比流行病学中相关差别指标的六种渐近置信区间的构造方法,综合考虑实际覆盖率与区间长度对各种方法的优劣及适用情况并对比分析。研究表明,Wald型置信区间与对数变换的置信区间对发病率的估计表现因参数而定,而Bootstrap类方法不稳定。本研究找出了不同区间估计方法的适用场合,认为应合理看待置信区间这种评估方法在流行病学中的实际应用。

A Comparative Study on the Interval Estimation for Relative Difference

In the field of public disease control, the incidence of major rare diseases is very low, in accordance with the inverse sampling characteristics, in order to quantify the incidence of major rare diseases and analysis of its characteristics.This paper mainly studies on the related conditions within the group with cluster sampling,by β-binomial distribution, under inner-correlated cluster sampling conditions by comparing the differences related to the epidemiology of this indicator in the six asymptotic confidence intervals constructed.The advantages and disadvantages of various methods and their application conditions are analyzed by considering the actual coverage ratio and the interval length.The results show that the estimation performance of the confidence interval of the Wald type confidence interval and the logarithmic transformation is dependent on the parameters, and the Bootstrap method is not stable.In this paper, the application of different interval estimation methods is found, and the practical application of confidence interval estimation method in epidemiology is reasonable.