Biochemia Medica, Vol. 28 No. 2, 2018.
Original scientific paper
https://doi.org/10.11613/BM.2018.020709
Application of Sigma metrics in assessing the clinical performance of verified versus non-verified reagents for routine biochemical analytes
Shuang Cao
; Department of Medical Laboratory, Shengjing Hospital of China Medical University, Shenyang, China
Xiaosong Qin
; Department of Medical Laboratory, Shengjing Hospital of China Medical University, Shenyang, China
Abstract
Introduction: Sigma metrics analysis is considered an objective method to evaluate the performance of a new measurement system. This study
was designed to assess the analytical performance of verified versus non-verified reagents for routine biochemical analytes in terms of Sigma metrics.
Materials and methods: The coefficient of variation (CV) was calculated according to the mean and standard deviation (SD) derived from the
internal quality control for 20 consecutive days. The data were measured on an Architect c16000 analyser with reagents from four manufacturers.
Commercial reference materials were used to estimate the bias. Total allowable error (TEa) was based on the CLIA 1988 guidelines. Sigma metrics
were calculated in terms of CV, percent bias and TEa. Normalized method decisions charts were built by plotting the normalized bias (biasa: bias%/
TEa) on the Y-axis and the normalized imprecision (CVa: mean CV%/TEa) on the X-axis.
Results: The reagents were compared between different manufacturers in terms of the Sigma metrics for relevant analytes. Abbott and Leadman’s
verified reagents provided better Sigma metrics for the alanine aminotransferase assay than non-verified reagents (Mindray and Zybio). All reagents performed well for the aspartate aminotransferase and uric acid assays with a sigma of 5 or higher. Abbott achieved the best performance for the urea assay, evidenced by the sigma of 2.83 higher than all reagents, which were below 1-sigma.
Conclusion: Sigma metrics analysis system is helpful for clarifying the performance of candidate non-verified reagents in clinical laboratory. Our study suggests that the quality of non-verified reagents should be assessed strictly.
Keywords
quality assessment; Sigma metrics; method decision chart; total allowable error
Hrčak ID:
201529
URI
Publication date:
15.6.2018.
Visits: 1.527 *