Abstract:

Objective To investigate the influence of different sigma（σ） levels by using allowable total errors（TEa） and biases（Bias） from different sources on the evaluation of clinical biochemical indicators. Methods Imprecisions [coefficients of variation（CV）] were calculated from the internal quality control（IQC） statistics collected by the Department of Clinical Laboratory of the Third People's Hospital of Mianyang in 2017. The Bias calculation was based on the 2017 external quality assessment（EQA） of the National Center for Clinical Laboratories and trueness verification scheme data. The TEa came from the National Standard GB/T 20470—2006 and the Health Industry Standard WS/T 403—2012. Totally,12 biochemical laboratory items that participated the trueness verification scheme were chosen to calculate the σ value,and the σ values of 4 groups were compared in pairs by SPSS 19.0 software. Results According to the health industry standard,alanine aminotransferase,total protein,urea and glucose had poor σ levels. With the TEa from the same source and Bias from different sources there was no statistical significance in the calculated σ value（P>0.05）. The calculated σ levels of different sources of TEa were different（P<0.05）. Conclusions Bias from different sources has little effect on σ value. Different sources of TEa will generate different σ values. It is recommended that laboratories should choose TEa from health industry standard and adopt trueness verification scheme result as Bias to calculate σ value. If the conditions do not meet,Bias can be calculated from the result of EQA.

Key words: Allowable total error, Bias, Sigma, Performance evaluation