Japanese Journal of Clinical Chemistry, 34(1): 40~46, 2005 Estimation Methods for Various Cases of Uncertainty of Measurement in Clinical Chemistry

Regarding the expression of the reliability of measurement results, an international consensus is emerging that analytical values are expressed in combination with the uncertainty of their measurement. This paper proposes basic procedures for the estimation of uncertainty in analytical methods used for laboratory testing. There are two situations where uncertainty estimation is necessary: (1) estimation of the uncertainty of routine test values and (2) estimation of the uncertainty of assigned values of calibrators and quality assurance (QA) samples. For each of these situations, methods of uncertainty estimation are discussed in four cases: the cases where the analytical procedures are calibrated at every batch of assay, where the calibration takes place only at the first batch of assay, where multipoint linear calibration is performed, and where calibration is performed using a calibration curve. Key words: uncertainty of measurement, estimation method, routine test value, assigned value of calibrator and quality assurance sample, traceability, metrology, reliability, measurand, component of uncertatinty, reference material, linear calibration, nonlinear calibration, repeatability