La classification NOVA répartit les aliments en quatre groupes (de 1, peu transformé à 4, le plus transformé). Sa fiabilité et sa reproductibilité ont été très peu explorées alors qu’elle est très utilisée. Nous avons étudié la cohérence entre utilisateurs (environ 170 professionnels) de l’affectation de 120 aliments du commerce et 111 aliments génériques dans les groupes NOVA. La cohérence des affectations (κ de Fleiss) était de 0,32 (aliments du commerce) et 0,34 (aliments génériques). Des clusters d’aliments présentant des distributions d’affectations similaires ont été définis par classification hiérarchique. Parmi les aliments du commerce, un cluster réunissait 90 aliments très largement affectés à NOVA4 (91 % des affectations) et, parmi les aliments génériques, trois clusters contenaient des aliments principalement affectés à NOVA 1 (79 %), NOVA2 (75 %) et NOVA4 (70 %), respectivement. Les affectations étaient particulièrement hétérogènes pour 30 aliments du commerce et 28 aliments génériques (25 % du total dans les deux cas). Les critères actuels de la classification NOVA ne permettent pas d’affecter de manière non ambiguë un aliment à un groupe. Ce système doit être amélioré, en raison de son utilisation en recherche et dans les politiques publiques.

In the NOVA classification system, descriptive criteria are used to assign foods to one of four groups based on processing-related criteria. Although NOVA is widely used, its robustness and functionality remain largely unexplored. We determined whether this system leads to consistent food assignments by users. French food and nutrition specialists completed an online survey in which they assigned foods to NOVA groups. The survey comprised two lists: one with 120 marketed food products with ingredient information and one with 111 generic food items without ingredient information. We quantified assignment consistency among evaluators using Fleiss’ κ (range: 0–1, where 1=100% agreement). Fleiss’ κ was 0.32 and 0.34 for the marketed foods (n=159 evaluators) and generic foods (n=177 evaluators), respectively. The consistency of the assignments, as assessed by Fleiss’ κ, was 0.32 (commercial foods) and 0.34 (generic foods). Clusters of foods with similar assignment distributions were defined by hierarchical ascending classification. Among the 120 commercial foods, one cluster contained 90 foods that were predominantly assigned to NOVA4 (91% of the assignments) and, among the 111 generic foods, three clusters contained foods that were predominantly assigned to NOVA 1 (79% of the assignments), NOVA2 (75%) and NOVA4 (70%), respectively. Assignments were particularly heterogeneous for 30 commercial and 28 generic foods (25% of the total in both cases). Although assignments were more consistent for some foods than others, overall consistency among evaluators was low, even when ingredient information was available. These results suggest current NOVA criteria do not allow for robust and functional food assignments.