不同检测条件下血糖检测结果的偏倚分析

目的分析不同检测方法、检测仪器、试剂和校准物的检测条件下同一血清样本血糖含量的测定结果，研究影响血糖检测结果的因素，为临床准确检测血糖提供实验依据。方法由卫生部在全国范围内选取816家医院，将卫生部专家实验室确定的公议值作为血糖含量的靶值，以靶值为标准计算出各个检测部门血糖检测结果的偏倚，剔除样本量较小（小于25）的检测部门，用SPSSI6．0软伴对不同的检测条件的偏倚进行方差分析，确定影响血糖检测结果的因素。结果不同方法、仪器、试剂、校准物测定血糖的实验室间精密度变异系数均小于5．30％。经方差分析，不同检测方法、不同检测仪器检测同一血清样本血糖值偏倚的总体差异均无统计学意义（F=0．067，P〉0．050；F=0．831，P〉0．050），不同检测试剂和不同校准物检测血糖的偏倚差异有统计学意义（F=1．893，P〈0．050；F=13．687，P〈0．050）。结论血糖测定结果的差异主要来自不同的校准物。

Bias analysis of determination results of blood glucose under different conditions

Objective To analyze these values of blood glucose of the same sample under different methods, instruments, reagents and standards, study influencing factors of determination results of blood glucose, and provide experiment evidences for accurate detection of blood glucose in clinic. Methods 816 hospitals around the country were selected by the Ministry of Public Health , and the target value of blood glueose was decided by the professional laboratory of the Ministry of Public Health. The bias of different detection departments were calculated out by the target value and those values of sample size which were less than 25 were rejected . After that, SPSSI6.0 was used to conduct analysis of variances on the bias of different conditions, and determine influencing factors of determination results of blood glucose. Results CV values of precision were less than 5.30% among many kinds of laboratories when blood glucose were detected by different methods, instruments, reagents and standards. And tota~ difference of these bias values of blood glucose had no statistical significance under different detection conditions such as method and instrument by anlysis of variance（ F= 0. 067,P〉0. 050; F= 0. 831, P〉0.050）, whereas a significant difference showed among these values of blood glucose detected by reagent and different standards （ F= 1. 893,P〈0. 050 ; F= 13. 687, P〈0. 050）. Conclusion The main influencing factor was different standards when values of blood glucose were detected.