Z-score分数和泊松分布在血站血液核酸筛查室内质控的应用

Establishment and application of quality control charts using Z-score and Poisson distribution in blood nucleic acid screening laboratory

目的制作1种采用Z分数(Z-score)和泊松(Poisson)分布方法对数值自定义的核酸定性室内质控图以反映血站血液(核酸)筛查实验室在控、假阳性、假阴性质控点。方法采集2018年12月份本站核酸实验室检测数据作出Excel表,自定义室内质控品检测阴性和阳性结果,以时间为横坐标,自定义值为纵坐标,绘制自定义室内质控核酸室内质控图;以时间为横坐标,以±3为告警限,±4为失控限,采用Z分数法建立核酸实验室内部质控品(IQ)、外部质控品(EQ)及标本内对照(IC)的核酸检测系统产生的扩增循环数(Ct值)Z-score质控图;采用Poisson分布概率法计算每日核酸检测阳性概率,在Excel表中以时间为横坐标,以概率为纵坐标,以P=0.05为失控限,绘制Poisson分布阳性概率质控图。验证将这3种统计方法质控图在血液筛查实验室的实用性、可行性。结果自定义核酸定性室内质控图简单明了可适用于不产生数值的室内质控;Z-score质控图能够同时呈现3个联检项目和内标的Z值,能够避免数据非正态分布引起的假失控;Poisson分布阳性概率图能够反应出实验室污染或检测过程有效性。结论 Z-score和Poisson统计方法建立的室内质控图满足了血站血液筛查实验室NAT定性检测室内质控的要求,具有预警失控和试验有效性的功能,能有效提升献血者血液核酸检测质量控制和实验室管理能力。

Objective To propose a statistical internal quality control(IQC) chart of nucleic acid test(NAT) on blood donors′ samples using Z-score and Poisson distribution,Methods Based on Excel sheets of NAT data collected in our lab in December 2018 and the customized negative and positive results of IQC products, the IQC chart of self-defined qualitative NAT was plotted, with date as the abscissa and definition value as the ordinate. Then a Z-score control chart based on the amplification cycle number(Ct value), which came from IQC, external quality control(EQC), and sample internal control(IC) of the NAT in our lab, was plotted with date as the abscissa, ±3 as warning limit and ±4 as out-of-control limit.Also, quality control chart of positive probability based on the positive probability of daily NAT calculated by Poisson distribution was plotted with date as the abscissa, probability as the ordinate, and P=0.05 as the out-of-control limit.These three statistical control charts were verified for their practicability and feasibility in blood screening lab. Results The IQC chart of self-defined qualitative NAT was simple and clear, which could be applicable to IQC yielding with no numerical values. The Z-score quality control chart was possible to present the Z values of three joint detection items and ICs at the same time, which can avoid the false out of control caused by the nonnormal distribution of data. Poisson distribution positive probability chart can present laboratory contamination or the effectiveness of the detection process. Conclusion The statistical IQC charts of NAT established by Z-score and Poisson in blood screening laboratory of blood center guarantee a good usefulness. With such features as early warning out of control and effectiveness etc., they can further improve the quality control of NAT and laboratory management capabilities.