From validated method to validated results

This article provides important additions to [P 224] "Validation of quantitative analysis methods using the accuracy profile". For this reason, it does not repeat the elements then detailed, such as the basic vocabulary, the computational aspects of the accuracy profile or software applications, as well as the bibliographical references which remain current.

This article presents a recent development, based on the widespread use of measurement uncertainty and uncertainty functions. Measurement uncertainty is a natural methodological development in the field of method validation in analytical chemistry.

Validation is mainly a preoccupation of analysts, yet their role has changed considerably over the last 20 years. In the 1970s, chemometricians proposed defining an analyst as a problem-solver. This highly ambitious definition is now outdated, as many analysts now work in subcontracting laboratories for customers or principals who outsource their testing. Analysts have gradually gone from being genuine problem-solvers to data producers. Of course, not all laboratories are like this. Many university and industrial research facilities still employ analysts to solve problems. However, the setting up of specialized instrumental platforms, to reduce the cost of measurement and federate the use of delicate instruments, tends to specialize them and distance them from the general understanding of the problems to be solved. Yet the demand for analysis results is growing, as they are increasingly taken into account as a basis for decision-making. Paradoxically, in the future, the concept of "decision-making science" could be seen as an implicit objective of the analytical sciences.

Based on a critical presentation of developments in method validation and measurement uncertainty estimation, it is therefore possible to highlight the trends expected over the coming decades.

A glossary at the end of the article (§ 6 ) lists definitions that are important...