患者数据质控在定义室内质控分析批长度中的应用

Optimation of analytical run length for clinical laboratory internal quality control with the combination of patient-based and control-based quality control

目的将传统的控制物质量控制与患者数据均值(average of norm als,AON)法质量控制结合,建立一种客观定义临床实验室室内质量控制分析批长度的方法。方法从我室常规生化检验项目中,选择不同性能水平的8个项目。(1)根据方法分析性能的西格玛(σ)值设计个体化的控制物质量控制方案,利用通过计算机模拟建立的AON法患者数据质控规则系统设计AON法质量控制方案。(2)选择工作量较大且较稳定的一周时间,每日8、10、12、14时各做一次控制物质控,并收集同期的患者检验结果,用AON法质控判断分析系统状态。(3)结合控制物质控与AON法质控的结果,确定分析批长度。结果对于分析性能好(>3.5σ)、控制物质控误差检出率高的项目,控制物质控与AON法质控的结果较一致,即二者误差检出效能相当。此时,主要根据控制物质控结果确定分析批长度;对于分析性能差(<3.5σ)、控制物质控误差检出率低的项目,AON法质控的误差检出效能高于控制物质控。此时,主要根据AON法的结果确定分析批长度。本文所选的8个项目的分析批长度设定如下:三酰甘油、钾、总蛋白8 h,氯6 h,镁4 h,钙、二氧化碳结合力、钠2 h。结论分析性能高于3.5σ的项目,可主要根据控制物质控结果确定分析批长度,分析性能低于3.5σ的项目,可主要根据AON法质控确定分析批长度。

Objective The main purpose is to establish a simple method of analytical run length definition through combination of controlbased quality control （QC） and average of normals method （AON）. Methods Eight test items with different analytical performance were chosen. First, the individualized control-based quality control strategy was designed in the direction of sigma metrics, and the suitable AON rules were selected for each analyte. Then, the selected AON rules were applied to the patient data of successive five workdays. Meanwhile, new individualized control-based QC procedures were also used at 8：00, 10：00, 12：00 and 14：00 in those days. At last, AON Qc result were compared with control-based QC result to define the analytical run for 8 items and the strategy through which laboratory can optimize analytical run length. Results The error detection power of AON algorithms was as good as control-based QC whose performance was excellent. Analytical run length for 8 items involved in this study were defined as follows：triglyceride, potassium, total protein： 8 hours, Chlorine： 6 hours,magnesinm：4 hours, calcium, carbon dioxide combining power,sodium： 2 hours. Conclusions For the items with performance above 3.5 sigma, the analytical run was defined mainly depending upon control-based QC, and AON QC was just used to validate control-based result. For the items with performance below 3.5 sigma, the analytical run was defined mainly depending upon AON QC.