MEASUREMENT UNCERTAINTY / 04
Calibration, traceability and bias
Separate the uncertainty of assigned values from the variability of using them.
Follow the value, not just the label
Traceability links a measurement result to a stated reference through a documented calibration chain, with each step contributing uncertainty. A traceability claim alone does not demonstrate that an assay is suitable for every clinical use.
| Contribution | Question to ask |
|---|---|
| Assigned calibrator value | What value, units, uncertainty and coverage factor are supplied? |
| Local calibration process | Are fitting, repetition and recalibration effects already captured? |
| Material suitability | Does the material behave appropriately in this procedure? |
| Bias or correction | What comparison supports it, and what uncertainty remains? |
Convert certificates carefully
If a certificate gives expanded uncertainty U with a stated coverage factor k, its corresponding standard uncertainty is U/k. Do not assume that an unspecified manufacturer range is an expanded uncertainty, or that a control acceptance range describes calibrator uncertainty.
Worked example: an independent missing term
Suppose the result-specific long-term precision contribution is 2.4%, and a genuinely additional, independent assigned-value contribution is U = 2.0% with k = 2. The latter gives u = 1.0%. If the model justifies unit relative sensitivity, combined relative standard uncertainty is √(2.4² + 1.0²) = 2.60%. At k = 2, expanded relative uncertainty is about 5.2%. Independence and absence of overlap are assumptions to demonstrate—not defaults.
Handle bias explicitly
A comparison can reveal a discrepancy without establishing which procedure is correct. Examine reference quality, sample comparability and concentration dependence. Where a justified correction is applied, include its uncertainty and the relevant model terms. Do not automatically treat the magnitude of an uncorrected bias as an independent random SD.
Ask for missing information
Request the relevant certificate, definition of uncertainty, coverage factor, traceability statement and lot information. If evidence is unavailable, record the gap and its consequences rather than substituting an unrelated QC target range.
Sources & further reading
- JCGM 200:2012 — International Vocabulary of Metrology
- JCGM 100:2008 — Guide to the Expression of Uncertainty in Measurement
- ISO/TS 20914:2019 — public scope and status
Educational guidance, not a substitute for a locally approved procedure. Worked examples are synthetic. Reviewed against the sources indicated; licensed standards must be consulted for their full requirements.