检验医学 ›› 2022, Vol. 37 ›› Issue (7): 674-679.DOI: 10.3969/j.issn.1673-8640.2022.07.015

• 实验室管理·论著 • 上一篇    下一篇

蒙特卡洛法对校准曲线法测定总胆红素浓度的不确定度评定

孟舒婷1, 周卫平2, 戴绘芬3, 董德平1, 董业峰1, 葛仁美1, 王惠民4()   

  1. 1.南通大学附属海安医院检验科,江苏 南通 226600
    2.南通市通州区人民医院检验科,江苏 通州 226300
    3.海安市中医院检验科,江苏 南通 226600
    4.南通大学附属医院检验科,江苏 南通 226601
  • 收稿日期:2021-12-19 修回日期:2022-03-05 出版日期:2022-07-30 发布日期:2022-08-26
  • 通讯作者: 王惠民
  • 作者简介:王惠民,E-mail: whm_jyk@ntu.edu.en
    孟舒婷,女,1988年生,硕士,主管技师,主要从事免疫、分子生物学检测和实验室质量管理工作。
  • 基金资助:
    南通市科技计划指导项目(JCZ19075)

Evaluation of uncertainty of serum total bilirubin in calibration curve method by Monte Carlo method

MENG Shuting1, ZHOU Weiping2, DAI Huifen3, DONG Deping1, DONG Yefeng1, GE Renmei1, WANG Huimin4()   

  1. 1. Department of Clinical Laboratory,Hai'an Hospital of Nantong University,Nantong 226600,Jiangsu,China
    2. Department of Clinical Laboratory,Tongzhou People's Hospital,Nantong 226300,Jiangsu,China
    3. Department of Clinical Laboratory,Traditional Chinese Medical Hospital of Hai'an County,Nantong 226600,Jiangsu,China
    4. Department of Clinical Laboratory,the Affliated Hospital of Nantong University,Nantong 226601,Jiangsu,China
  • Received:2021-12-19 Revised:2022-03-05 Online:2022-07-30 Published:2022-08-26
  • Contact: WANG Huimin

摘要:

目的 优化蒙特卡洛法(MCM)对国际检验医学溯源联合委员会(JCTLM)推荐的总胆红素参考测量程序不确定度的评定过程,参考实验室条件实现对血清总胆红素不确定度评定的量值溯源。方法 应用JCTLM推荐的参考测量方法(Doumas法)测定国际临床化学和检验医学联合会(IFCC)2014年参考实验室外部质量评价计划国际比对样本A和样本B,基于MCM原理,借助可视化软件MATLAB对总胆红素浓度测量不确定度进行评定。结果 建立了实验室测量不确定度表达指南(GUM)的不确定度的评定模型,校准曲线相关参数为:斜率(b)=0.007 531 6,斜率的标准差为0.000 025 51;浓度均值为87.284 4 mg/L,浓度均值的标准差为0.196 5;吸光度均值($\bar{y}$)=0.657 917,吸光度均值的标准差为0.000 365 4;截距(a)=0.000 5251 6,截距的标准差为0.002 699。参考实验室条件下总胆红素测定结果:样本A浓度x=39.368 8 mg/L,合成标准不确定度(u)=0.603 2,95%可信区间为38.188 0~40.553 5;样本B浓度x=18.703 1 mg/L,合成标准不确定度(u)=0.296 4,95%可信区间为18.122 5~19.284 4。MCM输出量为偏正态分布,呈尖峰肥尾。用自适应MCM验证GUM框架的测量不确定度评定法(GUF)包含区间,2组绝对偏差dlow和dhigh均值>数值容差0.05,GUF的结果未通过 MCM验证。结论 MCM借助可视化软件MATLAB,可操作性强,能避免经典GUF评定不确定度的繁琐,是较GUF评定不确定度更方便和精确的方法。

关键词: 蒙特卡洛法, 不确定度, 总胆红素, 标准曲线

Abstract:

Objective To evaluate the uncertainty process for total bilirubin reference measurement procedure which was recommended by the Joint Committee for Traceability in Laboratory Medicine(JCTLM) used by the Monte Carlo method(MCM) through optimizing. To realize quantity traceability about the uncertainty evaluation of serum total bilirubin under reference laboratory conditions. Methods Applying Doumas method which was recommended by JCTLM,sample A and sample B from external quality assessment scheme for reference laboratories of the International Federation of Clinical Chemistry and Laboratory Medicine(IFCC) in 2014 were determined. Based on software MATLAB,the uncertainty of total bilirubin determination was evaluated by MCM. Results The uncertainty model which was evaluated by the Guide to the Expression of Uncertainty in Measurement(GUM)was established. Parameters related to the calibration curve were as follows:slope(b) was 0.007 531 6,the standard deviation of slope was 0.000 025 51,mean concentration was 87.284 4 mg/L,the standard deviation of mean concentration was 0.196 5,mean absorbance($\bar{y}$) was 0.657 917,the standard deviation of mean absorbance was 0.000 365 4,intercept(a) was 0.000 525 16,and the standard deviation of intercept was 0.002 699. The results of total bilirubin determination under laboratory conditions were as follows:the concentration of sample A x was 39.368 8 mg/L,the standard uncertainty of sample A synthesis(u) was 0.603 2,95% confidence interval(CI) was 38.188 0-40.553 5,the concentration of sample B x was 18.703 1 mg/L,the standard uncertainty of sample B synthesis(u)was 0.296 4,and 95%CI was 18.122 5-19.284 4. Output quantity of MCM presented skewed normal distribution,showing peak fat-tail. The GUM uncertainty framework(GUF) inclusion interval was verified with adaptive MCM. The absolute deviation dlow and dhigh values of the 2 groups were greater than the numerical tolerance 0.05. The results of GUF did not pass the MCM verification. Conclusions With the help of visual software MATLAB,MCM can optimize calculation process,improve the accuracy of inspection results highly compared with the classical GUF.

Key words: Monte Carlo method, Uncertainty, Total bilirubin, Standard curve

中图分类号: