Uncertainty of measurement for 14 immunoassay analytes: application to laboratory result interpretation

Abstract

Laboratory tests are an integral part of clinical decision making. Therefore, measurement uncertainty comes into prominence in the context of the accuracy of the laboratory result. This study aims to investigate measurement uncertainty of 14 immunoassay analytes, to compare them with different quality goals and to utilize them in the result interpretation. Measurement uncertainties of 14 immunoassay analytes were estimated by using internal and external quality control data by using Nordtest approach. Expanded uncertainties (U) were compared with allowable total error (TEa%), permissible relative deviation in the external quality assessment (PRD_EQA%) and permissible expanded uncertainty for external quality assessment (pU_EQAS%). Uncertainties were incorporated into the calculation of reference change values (RCV) and uncertainty adjusted reference intervals. RCVs of 14 analytes were calculated by three different methods reported by Harris, Clinical Laboratory Standards Institute (CLSI), and National Pathology Accreditation Advisory Council (NPAAC). Measurement uncertainties of TSH, estradiol, LH, progesterone, prolactin, and vitamin B12 were within defined allowable limits. U_{one-sided FT3} and U_{one-sided ferritin} exceeded defined TEa% but U_FT3 and U_ferritin were found below the limits of pU_EQAS%. Measurement uncertainties of FT4, cortisol, DHEAS, FSH, testosterone, and folate did not meet the specification limits. Recently defined permissible expanded uncertainty promises new targets to compare estimated measurement uncertainty. Measurement uncertainty should be applied to the laboratory result interpretation within the scope of RCV and reference interval to obviate misdiagnosis. Furthermore, we suggest that laboratories should inform clinicians about the tests with high uncertainties to assist them making the right clinical diagnosis.

Abbreviations

CLSI: Clinical Laboratory Standards Institute; CV: coefficient of variation; CV_A: analytic coefficient of variation; CV_G: inter-individual coefficient of variation; CV_I: intra-individual coefficient of variation; DHEAS: dehydroepiandrosterone sulfate; FSH: follicle-stimulating hormone; FT3: free triiodothyronine; FT4: free thyroxine; k: coverage factor; LH: luteinizing hormone; LRL: lower reference limit; MD: minimal difference; NPAAC: National Pathology Accreditation Advisory Council; PRD_EQA%: permissible relative deviation in the external quality assessment; pU_EQAS%: permissible expanded uncertainty for external quality assessment; RCV: reference change value; RCV': uncertainty-adjusted reference change value; TSH: thyroid-stimulating hormone; Rw: within-laboratory reproducibility; RMSbias: root mean square of biases; u(Cref): the uncertainty of nominal values; u(bias): uncertainty component for bias; u_c: combined standard uncertainty; TEa%: allowable total error; U: expanded uncertainty; U_one-sided%: one sided estimation of expanded measurement uncertainty using coverage factor “1.65”; URL: upper reference limit

Keywords:

• Uncertainty
• immunoassay
• reference values
• quality control
• laboratories

Disclosure statement

No potential conflict of interest was reported by the authors.