基于EWMA质控图和泊松分布的核酸检测系统稳定性预警机制的建立

Establishment of stability early warning mechanism of nucleic acid detection system based on EWMA quality control diagram and poisson distribution

目的建立Procleix TIGRIS联检的稳定性预警机制,用以监测日常检测工作的稳定性。方法收集2017年1月—12月Procleix TIGRIS核酸检测系统外部质控品的HIV/HCV/HBV联检S/CO值。对数据进行正态分布检验,判断是否可以使用基于数据正态分布的统计学质量控制方法。应用指数加权移动平均(EWMA)质控图对外部质控品检测S/CO值进行分析,验证该方法对核酸检测稳定性的预警作用。使用泊松分布公式对核酸检测中反应性样本的检出率进行计算,建立阳性率异常的早期预警。结果HIV、HCV和HBV外部质控品联检S/CO值均不符合正态分布。EWMA质控图对外部质控品核酸联检S/CO值的波动敏感,可显示出不同试剂批号间的微小系统性差别。当阳性检出率在4.9‰时,169例标本中反应性标本数目>3的概率为1.02%,500例标本中反应性标本数目>5的概率为3.88%,均属于小概率事件,可根据此计算结果建立核酸联检反应性标本数目异常界限。结论应用EWMA质控图和泊松分布可建立核酸联检的早期预警机制,利于加强核酸日常检测过程的稳定性监控。

Objective To monitor the stability of daily inspection work through the establishment of stability early warning system of Procleix TIGRIS HIV/HCV/HBV multiple detection.Methods HIV/HCV/HBV S/CO values for external quality control of Procleix TIGRIS nucleic acid detection system from January to December in 2017 were collected.The normal distribution of the data was tested to determine whether the statistical quality control method based on the normal distribution of data can be used.The S/CO value of external quality products was analyzed using exponential weighted moving average(EWMA)quality control chart to find out whether this method played an early warning role in the stability of nucleic acid detection.Poisson distribution formula was used to calculate the detection rate of reactive samples in nucleic acid detection,and early warning of abnormal positive rate was established.Results The external quality control S/CO value of HIV,HCV and HBV did not conform with normal distribution.The EWMA quality control chart was sensitive to the fluctuation of nucleic acid data of external quality control products,and could display the slight systematic difference between different batch numbers of reagents.When the positive rate was 4.9‰,the probability of the number of reactivity samples>3 in 169 samples was1.02%,and the probability of the number of reactivity samples>5 in 500 samples was 3.88%,all of which belonged to small probability events.According to the results,the number of abnormal reactivity samples of NAT can be established.Conclusion The use of EWMA quality control chart and Poisson distribution can establish the early warning mechanism of multiple NAT which is helpful to strengthen the monitoring of the stability of nucleic acid routine detection process.