全自动生化分析仪检测血糖的校准验证

Calibration verification for blood glucose test in automatic biochemistry analyzers

目的对日立和奥林巴斯全自动生化分析仪检测血糖作校准验证，旨在介绍一种作校准验证的方法并对校准验证的结果作评价。方法按照不同分析仪器将血糖的检测数据分为日立组和奥林巴斯组。利用美国病理学家协会日前使用的校准验证方法对日立和奥林巴斯全自动生化分析仪检测血糖作校准验证，即对校准线的斜率和截距作单样本t检验并计算百分差异值，并将百分差异值与允许误差作比对。如果校准线的斜率与1和截距与0比没有统计学意义且百分差异值没有超过允许误差，判定为验证1；如果校准线的斜率与1和截距与0比没有统计学意义且至少一个百分差异值超过允许误差，判定为差异1；如果校准线的斜率与1或截距与0比有统计学意义且百分差异值没有超过允许误差，判定为验证2；如果校准线的斜率与1或截距与0比有统计学意义且至少一个百分差异值超过允许误差，判定为差异2。结果在日立组，验证1、差异1、验证2和差异2分别占5．4％（4／74）、0．0％（0／74）、68．9％（51／74）和25．7％（19／74）。在奥林巴斯组，验证1、差异1、验证2和差异2分别占16．4％（12／73）、0．0％（0／73）、58．9％（43／73）和24．7％（18／73）。结论美国病理学家协会使用的校准验证方法是合理的和可行的。大部分实验室（74．8％，110／147）可以通过这次校准验、证评。

Objective To perform calibration verification for the blood glucose test in Hitachi and Olympus automatic biochemistry analyzers and present a method of calibration verification and evaluate the results. Methods The glucose test results were divided into two groups： Hitachi and Olympus groups according to the different analytical instruments. The calibration verification was performed following the current College of American Pathologists strategies. The slope and intercept of the calibration line were tested by one-sample t test and the percentage differences were calculated. The percentage differences were compared with allowable errors. If the slope and intercept had no statistically significant difference from 1 and 0 respectively and the percentage differences were within the limits of allowable errors, the calibration line was evaluated as verified 1 ; if the slope and intercept had no statistically significant difference from 1 and 0 respectively and at least one of the percentage differences exceeded the limits of allowable errors, the calibration line was evaluated as different 1 ; if the slope or intercept had statistically significant difference from 1 and 0 respectively and the percentage differences were within the limits of allowable errors, the calibration line was evaluated as verified 2; if the slope or intercept had statistically significant difference from 1 and 0 respectively and at least one of the percentage differences exceeded the limits of allowable errors, the calibration line was evaluated as different 2. Results In Hitachi group, verified 1, different 1, verified 2 and different 2 were 5.4% （4/74）, 0.0% （0/74）, 68.9% （51/74） and 25.7% （19/74）, respectively. In Olympus group, verified 1, different 1, verified 2 and different 2 were 16.4% （12/73）, 0. 0% （0/73）, 58.9% （43/73） and 24. 7% （ 18/73 ）, respectively. Conclusions The current College of American Pathologist strategies could be performed as a reasonable and feasible method for calibration verification. Most of participating laboratories （ 74. 8%, 110/147 ） passed the assessment of calibration verification.