用非参数统计方法评估检测体系的最低检出限

目的建立一套基于非参数统计原理的评估方案,用于空白标本检测结果为非正态分布、标准曲线为非直线线性的检测体系最低检出限(limit of detection,LoD)的确定。方法根据数据分布模型,以非参数统计方法确定空白标本检测结果的第95百分位数作为空白检测限(limit of blank,LoB),通过LoD=LoB+tβ×s低值的数学模型来评估LoD。结果以非参数统计方法建立的评估程序可通过验证程序证明该LoD值有效。结论对于空白标本检测结果不符合正态分布,标准曲线为非直线线性的资料可用非参数统计的评估方法来确定LoD。

Evaluate the limit of detection of quantitative system by non-parametric statistics

Objective To set up a new procedure based on non-parametric statistics principle for determining the limit of detection （LoD）, and estimate the LoD of an detection system which blank measurements is non-normal distribution or calibration curve is nonlinear. Methods By using theoretical distribution models and non-parametric approach to determine the 95th percentile of the blank measurements defines the limit of blank （LoB）, then LoD is estimated as： LoD=LOB＋tβx SDs. Results LoD detection system that was partly based on non-parametric statistics can be proved effective by proving program. Conclusion The proposed approach which is based on the theory of non-parametric theory appears useful for application in the field that blank distributions are nongaussian distribution and calibration curve is nonlinear.