Mary-Krause M, Mary J Y, Valleron A J
B3E-INSERM U.263 Faculté de médecine Saint-Antoine, Paris, France.
Rev Epidemiol Sante Publique. 1993;41(2):155-60.
If PV is the proportion of the target population vaccinated and PCV the proportion of cases in the same population who have been previously vaccinated, vaccine efficacy can be computed as (PV-PCV)/[PV(1-PCV)]. In the screening method, this formula is currently applied to sample values. The objective of this paper is to provide the corresponding statistical tools. Taylor series expansion has been used to derive the bias, variance and confidence interval of the estimate. The bias may be very high when vaccine efficacies are low. It is below 1% for vaccine efficacies higher than 80%. Monte Carlo simulations were used to assess the departure from normality of the estimate. The number of subjects required by the method for given precisions are presented.
如果PV是接种疫苗的目标人群比例，PCV是同一人群中先前已接种疫苗的病例比例，那么疫苗效力可计算为(PV - PCV)/[PV(1 - PCV)]。在筛查方法中，此公式目前应用于样本值。本文的目的是提供相应的统计工具。已使用泰勒级数展开来推导估计值的偏差、方差和置信区间。当疫苗效力较低时，偏差可能会非常高。对于高于80%的疫苗效力，偏差低于1%。使用蒙特卡罗模拟来评估估计值与正态性的偏离情况。给出了该方法在给定精度下所需的受试者数量。
