红细胞、白细胞、血红蛋白测定的不确定度评定

Evaluation of the uncertainty of blood cell count and hemoglobin determination

目的探讨全自动血液分析仪测定红细胞(RBC)数、白细胞(WBC)数和血红蛋白(Hb)的不确定度评估的方法。方法对全自动血液分析仪测定RBC、WBC和Hb的过程进行描述,分析测量不确定度的来源,对各个不确定度分量进行评估,然后计算合成不确定度和扩展不确定度。不确定度来源包括不准确度、不精密度、携带污染以及分析前不确定度;不准确度的不确定度采用计量部门校准证书上给出的不确定度;不精密度分为批内不精密度、批间不精密度和天间不精密度,批内和批间不精密度以及携带污染率采用国际实验血液学学会(ICSH)推荐的评价方法进行评价,天间不精密度由室内质控的在控结果计算得到。结果取k=2,3个项目的高、中、低水平的扩展不确定度分别为U95(RBC)高=5.0%,U95(RBC)中=4.8%,U95(RBC)低=4.6%;U95(WBC)高=5.0%,U95(WBC)中=4.6%,U95(WBC)低=5.6%;U95(Hb)高=6.4g/L,U95(Hb)中=5.6g/L,U95(Hb)低=4.8g/L。结论我们通过建立数学模型、分析寻找各项影响因素及运用数理统计方法,对全自动血液分析仪测定RBC、WBC和Hb的测量不确定度进行了较全面和合理的评定。

Objective To study the method of uncertainty evaluation of red blood cell （ RBC ） count, white blood cell （WBC） count and hemoglobin （Hb） determination by hematology analyzer. Methods To describe the process of RBC, WBC and Hb determination, analyze the resources of uncertainty, and evaluate the branches of the uncertainty, then calculate the combined uncertainty and the expanded uncertainty, the resources of uncertainty of these three items included bias, imprecision, sample carry-over and pre-analytical uncertainty. The uncertainty of bias adopted the uncer- tainty given in the calibration certificate of Zhongshan Supervision Testing Institute of Quality ＆ Metrology. The impre- cision was divided into within-run imprecision, between-run imprecision and between-day imprecision. Within-run and between-run imprecision were evaluated according to the International Committee for Standardization in Haemtology （ICSH）. Between-day imprecision was computed by the quality control data. Results When k = 2, the expanded uncertainties of RBC, WBC and Hb in high, median and low concentrations were U95（RBC）high = 5. 0%, U95（RBC）madian = 4. 8% ,U95（RBC）low = 4. 6% ; U95（WBC）high = 5. 0% , U95（WBC）median = 4. 6%, U95（WBC）low = 5. 6% ; U95（Hb）high = 6. 4 g/L, U95（Hb）median =5. 6 g/L,U95（Hb）low =4.8 g/L respectively. Conclusions This study evaluates the uncertainties of RBC, WBC and Hb comprehensively and reasonably through building a mathematical model, analyzing the resources of influence and using the methods of statistics.