粪大肠菌群(MPN法)测定结果的评价方法比较
2021-07-25王宠吴旸孙开争郭萌萌徐国栋张水燕
王宠 吴旸 孙开争 郭萌萌 徐国栋 张水燕



摘要:目的:探索8家实验室粪大肠菌群(MPN法)检测结果的一种评价方法。方法:采用Grubbs检验法、泊松分布置信区间法和改进的En值法分别评价了检测结果,结合“不合格”或“不满意”实验室问题追踪结果,比较了3种方法的评价结论。结果:较其他2种方法,改进的En值评价方法不但考虑了标准定量菌株和MPN方法的不确定度,而且还考虑了其分布特点,其评价结论更符合实际情況,更加科学合理。结论:改进的En值评价方法适用于MPN法的准确度评价,不仅可以评价能力验证或能力考核结果,而且还可以评价日常质控样品的准确度。
关键词:粪大肠菌群;MPN;En值
中图分类号:X820.3 文献标识码:A 文章编号:2095-672X(2021)01-0072-06
DOI.10.16647/j.cnki.cn15-1369/X.2021.01.011
Comparison of evaluation methods for the determination of fecal coliform (MPN)
Wang Chong ,Wu Yang,Sun Kaizheng,Guo Mengmeng,Xu Guodong,Zhang Shuiyan
(Ji'nan Eco-environmental Monitoring Center of Shandong Province, Ji'nan Shandong 250101,China)
Abstract:Objective:Explore an evaluation method of fecal coliform (MPN) detection results of 8 laboratories.Method:Grubbs test method, Poisson distribution confidence interval method and improved en value method were used to evaluate the test results. Combined with the "unqualified" or "Unsatisfactory" laboratory problem tracking results, the evaluation conclusions of the three methods were compared.Result :Compared with the other two methods, the improved en value evaluation method considered not only the uncertainty of standard quantitative strains and MPN method, but also their distribution characteristics. The evaluation conclusion is more in line with the actual situation and more scientific and reasonable.Conclusion:This method was suitable for the accuracy evaluation of MPN method, which can not only evaluate the ability verification or ability examination results, but also evaluate the accuracy of routine quality control samples.
Key words:Fecal coliform;most probable numbers; MPN;En value
最大可能数(MPN)法[1-5]很早就应用于总大肠菌群和粪大肠菌群等微生物指标分析。该方法是生态环境监测领域多管发酵法[6]和酶底物法[7]的理论基础,也在卫生检验、食品检验领域获得了广泛应用。该方法不同于平板计数法用菌落数量直观表示测量结果,而是通过概率理论估算细菌浓度的一种方法,同时给出检测结果的置信区间,可以是95%概率(包含因子k=2)和99%概率(包含因子k=3),因为MPN值本身的不确定度是MPN法检测结果不确定度的主要来源,因此可以忽略其他显著小的不确定度,该置信区间可以作为MPN值的扩展不确定度[8]使用。该不确定度的分布与常见的正态、三角型、梯形、矩形等对称分布不同,它属于不对称分布,多管发酵法MPN值的置信区间形状类似于色谱分析中的拖尾峰,表现为MPN值与置信区间下限的差值小,而与置信区间上限的差值大。……
