Two approaches for measurement uncertainty estimation: which role for bias? Complete blood count experience

Abstract

Objectives: Measurement uncertainty is described as a magnitude indicates the distribution of the measurement results. AACB Guideline suggests that bias should not be included in the uncertainty calculation contrary to Nordtest Guideline. The aim of the study is to calculate the uncertainty values of certain complete blood count (CBC) parameters and evaluate the contribution of bias. Methods: This retrospective study was performed with the quality control data of January–December 2017 of two different CBC autoanalyser models (Beckman Coulter LH780 and DXH800). Measurement uncertainties were calculated according to AACB and Nordtest Guidelines. Imprecision, i.e. measurement uncertainty, varies with concentration. Imprecision values of user manuals, as performance characteristics of autoanalyzers, were used for assessment. Results: User manuals imprecision values of different levels of platelets are between 3.3 and 14%. As the concentrations of platelets decrease, imprecision is observed to increase. This is expected to be parallel with measurement uncertainty. Contrary to user manuals, uncertainty values of AACB found to be so close to each other (between 3.41% and 4.80%), regardless of concentration level. However Nordtest guideline is more compatible with user manuals (between 6.97% and 15.35%). Conclusions: When evaluated with the performance expectations, bias should be used in measurement uncertainty. Calculation of uncertainty for different concentration level is also important.