GUF法和蒙特卡洛法在TBil参考测量程序测量结果不确定度评定中的应用  

Applications of GUF method and MCM in evaluating measurement uncertainty of TBil concenteation by JCTLM reference procedure

作　　者：周卫平 陶赟[2] 戴绘芬 王惠民[2] 王建新[2] 沈蕾[2] 陈峰[4] ZHOU Weiping;TAO Yun;DAI Huifen;WANG Huimin;WANG Jianxin;SHEN Lei;CHEN Feng(Department of Clinical Laboratory,Tongzhou People's Hospital of Nantong,Jiangsu Province,Nantong 226300;Department of Clinical Laboratory,the Affiliated Hospital of Nantong University;Department of Clinical Laboratory,Traditional Chinese Medical Hospital of Haian County;Nantong Entry and Exit Inspection and Quarantine Bureau)

机构地区：[1]江苏省南通市通州区人民医院检验科,通州226300 [2]南通大学附属医院检验科 [3]江苏省海安市中医院检验科 [4]南通出入境检验检疫局

出　　处：《南通大学学报（医学版）》2019年第6期455-459,共5页Journal of Nantong University(Medical sciences)

基　　金：南通市科技计划指导项目(YYZ15009)

摘　　要：目的:用测量不确定度表示指南(gudie to the expressin of uncertainty in measurement, GUM)不确定度框架法(GUM uncertainty framework, GUF)与蒙特卡洛法(Monte Carlo method, MCM)评定本实验室条件下根据国际检验医学溯源联合委员会推荐的Doumas血清总胆红素(total bilirubin, TBil)浓度测量结果的不确定度,比较两种评定结果并寻找和探讨适用的不确定度模式。方法:应用Doumas参考测量程序测定国际临床化学和检验医学联合会2014年参考实验室外部质量评价计划(external quality assessment scheme for reference laboratories, RELA)样本A和B,分析测定过程中各中不确定度分量来源,建立实验室TBil参考测量程序的测量数学模型。用GUF法与MCM评定测量结果的不确定度,并以自适应(adaptive MCM)对GUF法进行验证。结果:本实验室条件下TBil浓度测定GUF法评定结果:RELA样本A为(39.37±1.20) mg/L(k=1.96),95%置信区间为[38.16 mg/L, 40.57 mg/L],样本B为(18.70±0.59) mg/L(k=1.96),95%置信区间为[18.11 mg/L, 19.29 mg/L]。MCM评定结果:RELA样本A为(39.37±1.18)mg/L(k=1.96),95%置信区间为[38.19 mg/L, 40.55 mg/L],样本B为(18.70±0.58) mg/L(k=1.96),95%置信区间为[18.12 mg/L, 19.29 mg/L]。用自适应MCM验证GUF法包含区间,两组绝对偏差dlow和dhigh值均大于数值容差0.05,GUF法的结果未通过MCM验证。结论:MCM是一种更为准确的不确定度评定方法,GUF法较为简易方便。MCM对GUF法的结果验证后,发现GUF法虽然简单方便,但与MCM存在一定差异。Objective:To evaluate uncertainy in measurement of TBil using gudie to the expressin of uncertainty in measurement uncertainty framework(GUF)and Monte Carlo method(MCM)according to Doumas serum TBil reference procedure of the Joint Committee for Traceability in Laboratory Medicine(JCTLM),compare the results of GUF with MCM and try to find an appropriate model for evaluating uncertainty.Methods:The International Federation of Clinical Chemistry and Laboratory Medicine external quality assessment scheme for reference laboratories in laboratory medicine(RELA)sample in 2014 was measured by Doumas serum TBil reference procedure,all the possible sources of uncertainty during the process of measurement were defined and we established the mathematical model of TBil reference procedure in the laboatory.Using GUF and MCM respectively,in turn,to calculate the expanded uncertainty and the set range of model.We used the adaptive MCM to validate the result of GUF.Results:When the uncertainty was evaluated by GUF,the result of sample A was(39.37±1.20)mg/L(k=1.96)95%confidence interval was[38.16 mg/L,40.57 mg/L];the result of sample B was(18.70±0.59)mg/L(k=1.96),95%confidence interval was[18.11 mg/L,19.29 mg/L];when the uncertainty was evaluated by MCM,the result of sample A in 2014 was(39.37±1.18)mg/L(k=1.96),95%confidence interval was[38.19 mg/L,40.55 mg/L],the results of sample B was(18.70±0.58)mg/L(k=1.96),95%confidence interval was[18.12 mg/L,19.29 mg/L].The coverage intervals of GUF could not verified by adaptive MCM,and both the dlow and dhigh values of 2 tests were higher than the numerical tolerance,so the results of GUF could not be validated by MCM.Conclusions:MCM is a more accurate uncertainty assessment method,but GUF method is more simple and convenient.After MCM verified the results of GUF method,it was found that although GUF method was more simple and convenient,it was different from MCM.

关 键 词：不确定度 总胆红素浓度 测量不确定度表示指南不确定框架法 蒙特卡洛法 

分 类 号：R446.11[医药卫生—诊断学]