目的 在体外诊断试剂（IVDR）临床试验中，准确快速地完成EP09c的统计分析工作。方法 利用MedCalc软件对差异图（Difference Plot）执行偏倚量化；利用SPSS或MedCalc软件的回归模型对散点图（Scatter Plot）执行偏倚量化；差异图和散点图被用于确认偏倚量化结果的正确性。结果 MedCalc的差异图确定工作样本的低浓度数据符合恒定标准差分布，偏倚为0.10（0.09，0.28）；高浓度数据符合恒定变异系数分布，偏倚为-5.0%（中位数：-7.2%，1.5%）或-3.2%（Hodges-Lehmann中位数：-5.5%，-0.8%）。散点图确认Passing&Bablok回归和加权线性回归（权重1/x_i ^1.4）适用于工作样本的数据，回归方程分别为y=0.122+0.952x（R²=0.983）和y=0.269+0.977^x（R²=0.981）。结论 偏倚量化结果必须得到差异图和散点图的确认，否则结果的正确性无法保证。

      Objective To perform accurate and efficient statistical analysis of EP09c in clinical studies of in vitro diagnostic reagents（IVDR）．Methods Bias quantification was performed on the difference plot using MedCalc．Bias quantification was conducted on the scatter plot using regression modeling in SPSS and MedCalc．The difference plot and scatter plot were employed to corroborate the accuracy of the bias quantification results．Results Difference plots demonstrated that the low-concentration data of the working sample exhibited a constant standard deviation distribution with a bias of 0.10（0.09，0.28），while the high-concentration data conformed to a constant coefficient of variation distribution with a bias of -5.0%（median：-7.2%，1.5%） or -3.2%（Hodges-Lehmann median：-5.5%，-0.8%）．Scatter plots confirmed that Passing & Bablok（P-B） regression and weighted least square（WLS） regression（weight 1/x_i ^1.4） were applicable to the data from the working sample，with the resulting regression equations being y=0.122+0.952x（R²=0.983） and y=0.269+0.977x（R²=0.981），respectively．Conclusions The results of bias quantification must be confirmed by difference plot and scatter plot，otherwise the correctness of the results cannot be assured．