In this study, the data were collected from the second internal quality control of routine chemistry and the first external quality assessment of routine chemistry in 2014 organized by the National Center for Clinical Laboratory. One of the laboratories was selected for its coefficient of variation (CV) and the bias of 19 clinical chemistry items from the data. σ of 2 runs were calculated by 5 different TEa. The σ metrics' performance for assessing the analytical quality of clinical chemistry determination was analyzed comparatively.

σ metrics varied with the changes of TEa and imprecision. Under the National Health Industry Standard, the major σ values(68.4%)for control 1 ranged from 2 to 4 and from 3 to 6 for control 2(58%). Under RiliBÄK, except triglyceride (negative) and alanine aminotransferase (ALT)(<3), others had a σ value from 3 to 6, even up to 14.78. Under the Clinical Laboratory Improvement Amendment of 1988 (CLIA'88), 89.47% of control 1 showed a σ value>3, up to 7.69, and 84.2% of control 2 showed a σ value > 3, up to 10.43. Under biological variability, the σ value of control 1 ranged from 1 to 5, and the most (63%) was < 3, and that of control 2 ranged from 1 to 6, but those of 9 from 19 were < 3. Under the TEa of Australian, the σ value of control 1 was <3, and that of 79% control 2 was <3 .The σ value of control 2 was generally higher than that of control 1.

The 6σ is an efficient way to control quality, but the lack of TEa for many analytes and inconsistent TEa from different sources are important variables for the interpretation of σ metrics in a routine clinical laboratory.