臨床實驗室中,更換試劑或儀器,須評估其精密度及準確度,以至於新的檢驗方法可直接取代舊的檢驗方法或者兩種方法可交替的使用,但評估時常誤用了相關係數來進行分析,本研究為以醫療機構最常使用的血糖機及生化自動分析儀,兩種測定血糖的方法,進行方法比較的研究,以線性迥歸;paired-t test;Error Index Plot;MEDx chart等統計方法評估其誤差性質,再現性及performance等,以實例分析的方式提供給實驗室簡單且明確的方式,作為評估方法比較研究的參考。
When we want to change reagents or instruments in clinical laboratory, we must determine their precision and accuracy in order to judge whether the old method can be replaced by the new or use the two interchangeably. Such investigations are often analyzed inappropriately by using correlation coefficients. This study compared two methods of measuring blood glucose by glucometer and automatic chemistry analyzer that are often used to measure blood glucose in medical care institutions. Use statistical techniques such as the linear regression, paired-t test, Error Index Plot, and MEDX chart to estimate the nature of the error, repeatability, and performance. This study by case analysis can provide a simple and definite approach in method-comparsion studies for clinical laboratory reference.