方差随机效应模型分析凯氏定氮法蛋白质含量测定的中间精密度

Application of variance analysis to random-effect model in assessment of intermediate precision of Kjeldahl method

目的利用方差随机效应模型分析本实验室凯氏定氮法蛋白质含量检测过程中变异来源及各种来源变异在总变异中所占比例。方法将凯氏定氮法蛋白质含量检测过程中,检测人员、检测仪器、检测时间3个主要影响因素纳入方差随机效应模型进行方差成分分析,并计算中间精密度(intermediate precision,IP)和组内相关系数(intra-class correlation coefficient,ICC)。结果本实验室凯氏定氮法蛋白质含量检测过程中,49.20%的变异来源于检测人员因素,其次为检测仪器和检测时间;本实验室凯氏定氮法蛋白质含量检测IP和ICC分别为2.58%和0.59。结论在本实验室中,检测人员是凯氏定氮法蛋白质含量检测主要的变异来源;本实验室凯氏定氮法蛋白质含量检测的变异度较小,检测结果准确度较高;应用方差随机效应模型能有效分析检测过程中的随机影响因素。

Objective To assess the source and proportion of variability in Kjeldahl method by using random-effect model in analysis of variance（ANOVN）. Methods Three variables during determination of protein content by Kjeldahl method,date, measurement instrument and operators, were included in ANOVN, based on which the intermediate precision（IP）and intra-class correlation coefficient（ICC）of the method were calculated. Results During determination of protein content by Kjeldahl method, 49. 20% of total variability was due to operators, followed by measurement instrument and date.The IP and ICC of the method were 2. 58% and 0. 59 respectively. Conclusion In the process of determination of protein by Kjeldahl method, a majority of the variability might be due to the operators. The determination result of protein content was reliable with high accuracy and good precision in this laboratory. The random-effect of variance analysis may be applied to the analysis of randomly influencing factors during determination.