Measurement Uncertainty (MU) in Diagnostic Assays - Estimation, Repeatability, and Cut-off Applications

Aug 27, 2026

Measurement uncertainty (MU), also called measurement imprecision, is required for diagnostic laboratories to express results as an estimate with an associated level of imprecision rather than a single exact value. MU is limited to the measurement process and complements assay validation. Compliance with ISO/IEC 17025-2005 requires MU estimates for quantitative test outputs (e.g., OD, positivity/inhibition percentages, titres, and CT values). Guidance covers statistical approaches for serology and RT PCR, the role of repeatability studies, internal process controls, the “top-down” uncertainty approach, and why uncertainty becomes most critical near diagnostic cut-off thresholds.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages4

Reading Time10 min

AccessFree to download

Measurement uncertainty (MU) in diagnostic testing

Definition and scope of MU

MU requirements under ISO/IEC 17025-2005

Statistical estimation for serology and RT PCR

Repeatability and expected assay precision

Internal quality/process controls and robustness

Top-down uncertainty approach

MU relevance near diagnostic cut-off values

The OIE Validation Recommendations provide detailed information and examples in support of the OIE Validation Standard that is published as Chapter 1.1.6 Principles and methods of validation of diagnostic assays for infectious diseases of this Terrestrial Manual, or Chapter 1.1.2 of the Aquatic Manual. The Term “OIE Validation Standard” in this chapter should be taken as referring to those chapters. Estimation of measurement uncertainty (MU), sometimes termed measurement imprecision, is a requirement for testing laboratories based on international quality standards such as ISO/IEC 17025-2005, General requirements for the competence of testing and calibration laboratories (ISO/IEC 17025) . The measurement process for detection of an analyte in a diagnostic sample isnot entirely reproducible and hence, there is no exact value that can be associated with the measured analyte. Therefore the result is most accurately expressed as an estimate together with an associated level of imprecision. This imprecision is the measurement uncertainty (MU) . MU is limited to the measurement process. It is not a question of whether the measurement is appropriate and fit for whatever use to which it may be applied. It is not an alternative to test validation, but is rightly considered a component of that process (see the OIE Validation Standard, Section B.1. 1) . To assure compliance with ISO/IEC 17025-2005 requirements, national accreditation bodies for diagnostic laboratories require MU estimates for test methods that produce quantitative results, e.g. optical densities (OD), percentage of positivity or inhibition (PP, PI), titres , cycle threshold (CT) values, etc. This includes tests, where numeric results are calculated and then are expressed as a positive or negative result at a cut-off value. For the purpose of estimating MU in serology and RT PCR, suitable statistical measures are mean target values ± 2 standard deviations (SD), which is approximately equal to a 95% confidence interval (CI), relative standard deviation (RSD = SD / mean of replicates) and coefficient of variation (CV = RSD × 100%) . The concept of MU does not apply to strictly binary results (positive or negative) . Repeatability is the level of agreement between results of replicates of a sample both within and between runs of the same test method in a given laboratory. During assay development, repeatability is estimated by evaluating variation in results of independent replicates from a minimum of three (preferably five) samples representing analyte activity within the operating range of the assay (see the OIE Validation Standard , Sections A.2.5 and B.1.1 , and Chapter 2.2.6 Selection and use of reference samples and panels , Section 3.1) . Typically, the variation in replicate results is expressed as RSD or CV. The significant feature is that repeatability studies can be used to define the expected precision of the assay in the detection of a range of analyte concentrations. The use of internal quality or process controls over a range of expected results has become part of daily quality control and quality assurance operations of accredited facilities (see the OIE Validation Standard , Sections A.2.6 and B.5.1 , and Chapter 2.2.6 , Section 1.4) . These results provide a continuous monitor relative to different aspects of repeatability, e.g. intra-and inter-assay variation, intra-and inter-operator variation and intra-and interbatch variation, which, when subjected to statistical analysis, provide an expression of the level of robustness (precision) of a test procedure. The monitoring of assay quality control parameters for repeatability provides evidence that the assay is , or is not performing as expected. In order for control samples to provide valid inferences about assay precision, they should be treated in exactly the same way as test samples in each run of the assay, e.g. including sample preparation such as extraction steps or dilution of serum samples f

FAQs

Why must diagnostic laboratories estimate measurement uncertainty (MU) instead of reporting a single exact result?

The measurement process for detecting an analyte in a diagnostic sample is not fully reproducible, so no exact value can be assigned. Results should be expressed as an estimate together with an associated level of imprecision (MU).

How is MU estimated for serology and RT PCR in this guidance?

Use suitable statistical measures such as mean target values ± 2 standard deviations, which corresponds approximately to a 95% confidence interval. Related measures include relative standard deviation (RSD) and coefficient of variation (CV).

When does MU become most important for interpreting diagnostic outcomes?

MU becomes increasingly important as test values approach the diagnostic cut-off value, because interpretation is made relative to the assay threshold (positive, negative, or inconclusive). Low positive samples are especially appropriate for estimating MU.

Aug 27, 2026

Upload by Theodore