Methods to Express Dispersion of Results in Food Composition Data

In Europe one feature of National Food Composition Databanks (nFCDBs) is to provide data soundly supported in standardized quality assurance procedures. It is now widely recognized that the evaluation of the degree of dispersion associated with a result is an essential part of any quantitative analysis. According to recent requirements the concept of data quality incorporates the evaluation of the measurement uncertainty (MU) as an indicator of the reliability of the result. The aim of this work is to study the typification of approaches used to estimate measurement uncertainty in food composition analysis in compliance with the criteria established in the “Guide to the expression of uncertainty in measurement (GUM)”. The work addressed the approaches founded on the modelling of the measurement process as described in the GUM (chapter 8), and on the experimental approaches, typically precision and bias data, obtained from within-laboratory validation studies, quality controls, inter-laboratory method validation studies or proficiency testing schemes.This work was completed on behalf of the EuroFIR Nexus consortium and funded under the EU 7th Framework Food Quality and Safety Programme