Assessment of uncertainties in analytical results

ISO/CEI 17025 "General requirements for the competence of testing and calibration laboratories" calls for the validation of methods when laboratories use non-standardized methods or methods outside the scope of the standard. The standard also requires that uncertainty be associated with the results provided. Chemical analysis laboratories have a long experience and tradition of validating their analytical methods, but they sometimes find it difficult to assess the uncertainty of their results.

In this article, the authors take stock of the different approaches to assessing the uncertainty of analytical results, and show that the traditional method presented in the "Guide to the Expression of Uncertainty in Measurement" (GUM) (NF ENV 13005), can be supplemented and enriched by using method validation data.

A method's performance can be expressed in terms of characteristics such as: selectivity, specificity, accuracy, linearity, repeatability, reproducibility, robustness, detection limits, etc. These characteristics are assessed through experimental work carried out in-house (intralaboratory approach). The performance of a method can also be assessed collectively, involving several laboratories (interlaboratory approach).