两种总胆红素试剂测定结果的偏倚分析

目的 探讨两种总胆红素试剂测定结果间是否具有可比性.方法 依据EP9-A2文件要求,每天随机从临床样本中抽取8份不同浓度水平的血清样本,分别用两种试剂进行总胆红素测定,共测定5天,记录检测结果,将结果进行相关回归分析.计算在其医学决定水平处的预期偏倚（c）和预期偏倚的95%可信区间,并判断偏倚是否可以接受.结果 相关系数（r）=0.999, r2=0.998,截距（a）=-0.079,斜率（b）=0.998,回归方程：Y=0.998X-0.079.无方法内、间的离群点,测定结果的偏差在允许误差范围内.结论 在实验室用一种试剂替代另一种试剂前,可参照EP9-A2文件进行方法学比对和偏倚分析,判断其临床可接受性能,以保证实验室检测结果的准确性和可比性.

The bias analysis of two kinds of total bilirubin reagents detection results

Objective To investigate the total Bilirubin（TBil）comparability between two kinds of reagents by comparative analyzing Tbil test resets. Methods Based on the guideline of EP9 - A2, eight different concentrations of samples involved randomly were measured by two .types of TBil reagents for five days consecutively, and collect the results of five days together. All results were subjected to accumulate the linearity equation and the relativity eoeffieient. The 95% confidence interval of Be （bias at medical deeision level X） were calculated to estimate acceptability of the bias. Results The correlation coeffieient of using two kinds of TBil reagents was 0.999 （2 = 0.998）, intercept（a） ： - 0. 079, slope（b）：0.998, regression equation： Y = 0.998X- 0.079, without outlier of within- method and between- method, the bias of results were accepted in the range of the allowable error. Conclusion When the same test item is analyzed by a new reagent, it is necessary to evaluate the clinical acceptability to insure the accuracy and comparability by method e.omparison and bias estimation based on EP9 - A2 document.