使用室内质控与能力验证数据评估测量不确定度

目的使用室内质控数据与"能力验证(PT)"数据评估临床实验室酶学项目肌酸激酶(CK)、乳酸脱氢酶(LDH)、γ-谷氨酰转移酶(GGT)检测结果的测量不确定度,以找出引入不确定度的主因。方法北京市临床检验中心(BCCL)收集北京地区三级医院医学检验科CK、LDH、GGT项目同一时间段2个浓度连续3个月的室内质控数据,以各实验室室内质控数据评定实验室内测量复现性引入的相对测量不确定度[U_(rel)(R_W)];同时收集各实验室2011~2013年连续6次的PT数据,用以评定偏移引入的相对测量不确定度[Ucrel(bias)]。再通过U_(rel)(R_W)与Ucrel(bias)两个分量计算相对合成不确定度(Ucrel)和相对扩展不确定度(U_(rel))。结果测量不确定度评估结果显示,47家三级医院检验科U_(rel)(R_W)的中位数与四分位间距分别为CK:2.33、1.70,LDH:2.69、1.58,GGT:2.30、1.43;Ucrel(bias)的中位数与四分位间距分别为CK:3.92、2.40,LDH:4.84、3.17,GGT:4.33、2.70。整体而言,Ucrel(bias)明显大于U_(rel)(R_W)。CK、LDH、GGT在浓度分别为(235.3±28.5)、(234.7±26.6)、(55.0±3.0)U/L时,3个项目的U_(rel)分别为CK:10.29、5.56,LDH:12.00、6.36,GGT:10.77、4.96。结论基于室内质控与PT数据评估测量不确定度是适合目前临床实验室现状的一种简单可行的方式,减少和控制RMSrel(bias)是改善酶学结果可比性的关键。

Using internal quality control and Proficiency testing data in clinical laboratories for estimating measurement uncertainty

Objective Based on the data of internal quality control and Proficiency testing result in medical laboratory from the tertiary hospitals in Beijing to estimate the measurement uncertainty of the test results of creatinekinase(CK) ,lactate dehydrogenase (LDH ,γ‐glutamyltransferase(GGT) ,in order to find the main resource of measurement variation .Methods Internal‐quality control and external‐quality control data were collected by Beijing Center for Clinical Laboratories(BCCL) .The component induced from in‐ternal laboratory measurement reproducibility [Urel (RW )] was estimated with the statistics of internal quality control results of three months .The relative standard measurement uncertainty from bias [Ucrel (bias)]was estimated with six surveys of PT .Then , we calculated the relative combined standard uncertainty(Ucrel )and the relative expanded uncertainty(Urel )with combination of Ucrel (bias) and Urel (RW ) .Results The estimation of Urel (RW ) from all laboratories showed that the median and interquartile range of CK ,LDH ,GGT were 2 .33 and 1 .70 ,2 .69 and 1 .58 ,2 .30 and 1 .43 relatively .And the median and interquartile of Ucrel (bias) of CK ,LDH ,GGT were 3 .92 and 2 .40 ,4 .84 and 3 .17 ,4 .33 and 2 .70 relatively .The uncertainty introduced by Ucrel (bias) was larger than Urel (RW )′s and the Ucrel (bias) contributed more weight to Ucrel .The median and interquartile of Urel for CK ,LDH ,GGT were 10 .29 and 5 .56 ,12 .00 and 6 .36 ,10 .77 and 4 .96 at the concentrations of(235 .3 ± 28 .5) ,(234 .7 ± 26 .6) ,(55 .0 ± 3 .0)U/L .Con‐clusion The approach that based on the internal quality control data and PT results is a kind of simple and feasible way to evaluate the measurement uncertainty of results for clinical laboratory .To reduce and control RMSrel (bias) is the key to improve the enzy‐mology results comparable .