不同实验条件对高密度脂蛋白测定的偏倚分析

Bias Analysis on Determination Results of High Density Lipoprotein Cholesterol Under Different Kinds of Experimental Conditions

目的通过不同检测条件对同一样本高密度脂蛋白(HDLC)含量的测定,研究影响检测结果的因素,为临床准确检测高密度脂蛋白提供实验依据。方法在全国范围内选取815家医院,按方法原理进行分组后采用剔除离群值的方法得出的均值为高密度脂蛋白的含量为靶值,计算出各个检测部门检测高密度脂蛋白结果的偏倚,剔除样本量较小(小于25)的检测部门,然后用SPSS13.0软件对不同的检测条件的偏倚(检测方法、检测仪器、试剂、校准物)进行方差分析,确定对高密度脂蛋白检测结果的影响因素。结果不同的检测方法、不同的检测仪器检测同一样本血清高密度脂蛋白的含量其差异无统计学意义(F=0.040,P>0.05;F=1.336,P>0.05)。不同的试剂、不同的校准物检测同一标本血清高密度脂蛋白的含量其差异有统计学意义(F=9.762,P<0.01;F=6.479,P<0.01)。结论血清高密度脂蛋白测定主要与试剂和校准物有关。

Objective The content of high density lipoprotein cholesterol （HDLC） selected from the same sample was detected in different experimental conditions to explore the influential factors leading to the differences, which could provide an experiment basis for clinical use. Methods The detection results of the same sample, concluded from 815 hospitals around the country, were divided into several groups by method and principle, and the average was regarded as the content of HDLC by rejecting the Studentized Residual Error, and then calculated the bias of the results of HDLC examined by different hospitals after rejecting the values, the sample size of which was less than 25. The software of SPSS13,0 was employed to conduct analysis of variance （ANOVA） of the bias （method, instrument, reagents, calibrated object）, and the influential factors for the results were determined. Results The discrepancy which summarized from different ways of detection and various detecting instruments showed no statistical significance （F = 0. 040, P 〉 0. 05; F = 1. 336, P 〉 0.05）, The contents of HDLC from the same sample which were detected by different reagents and different calibrated objects showed statistically significant differences （F=9.762, P〈0.01; F=6.479, P〈0.01）, Conclusion Reagent and calibrated object have a great influence on the results of HDLC detection.