临床化学自建检测系统的误差分析

Investigation of the measuring error of analytical systems in clinical chemistry

目的通过对上海市医院实验室使用的部分临床化学试剂的患者样本比对实验,分析临床化学检测系统误差。方法选择5家国产试剂厂家(分别为MD、YY、KH、FH、CZ),根据厂家的校准声明校准其试剂,在日立自动生化分析仪上与罗氏诊断产品有限公司试剂进行丙氨酸转氨酶(ALT)、尿素(Urea)、胆固醇(Chol)、白蛋白(A lb)4个项目的比对实验。用回归方法统计比对结果,并计算4个项目在医学决定水平处的偏倚。结果MD的ALT试剂与罗氏试剂比对结果的相关系数(r)为0.735 0。其余试剂比对结果的r均>0.975。MD试剂ALT在20、60和300 U/L,Urea在9.28和17.84 mmol/L,Chol在2.33 mmol/L,A lb在20 g/L时的偏倚高于允许误差。其余试剂在医学决定水平处的偏倚均低于允许误差。结论5家国产试剂与进口配套试剂进行4个常规生化项目的患者样本比对实验,除一家厂家外,其余试剂质量基本符合临床使用要求。

Objective To investigate the measuring error of clinical chemistry analytical systems in the clinical laboratory in Shanghai with a comparison experiment using fresh patient samples. Methods To evaluate the comparison experiment between five domeslie reagent manufacturers （ MD, YY, KH, FH, CZ） and Roche Diagnostics of four items [ alanine aminotransferase （ ALT）, Urea, choleaterolesters （Chol） , albumin （Alb） ]. The coefficient of correlation （r） and the bias at clinical decision level were determined. Results The coefficient of correlation was greater than 0.97.5 except ALT of MD. The biases at clinical decision level were less than CLIA Requirements for Analytical Quality, except ALT 20,60,300 U/L;Urea 9.28,17.84 mmol/L; Chol 2.33 mmol/L; Alb 20 g/L. of MD. Conclusions The results of the comparison experiment in clinical chemistry between five domestic reagent manufacturers and Roche Diagnostics of tour items in clinical chemistry are satisfactory except one domestic reagent manufaeturer.