Updated Nutri-Score Algorithm Improves Alignment with NOVA Processing Classifications for Foods

An updated version of the Nutri-Score front-of-pack nutrition labelling algorithm demonstrated improved alignment with the NOVA food processing classification system, according to an analysis of 450 food products evaluated in Polish retail stores. Nutri-Score serves as one of the most widely implemented front-of-pack labelling frameworks in Europe, designed primarily to evaluate the overall nutritional quality of packaged foods. In contrast, the NOVA classification categorises food items based entirely on the extent and technical purpose of industrial processing. The modification to the Nutri-Score calculation formula was developed specifically to reflect updated dietary guidelines and reduce the incidence of favourable grades being assigned to highly processed and ultra-processed food items.

To investigate the real-world operational impact of the formula change, investigators evaluated whether the updated Nutri-Score system produced better agreement with NOVA categories when applied across diverse grocery items.

The study sample comprised exactly 450 distinct food products gathered from Polish supermarkets, representing a broad selection of retail food categories. For each item in the dataset, researchers calculated nutritional profiles using both the original Nutri-Score algorithm and the revised 2023 updated scoring formula. Concurrently, all items were independently categorised using the NOVA framework to identify their level of industrial alteration. To further examine shifts across different nutritional qualities, the researchers incorporated an independent healthfulness classification based on Part B of the KomPAN questionnaire along with supplementary author-defined criteria, allowing comparison of rating shifts across products possessing varied baseline nutritional characteristics.

Statistical evaluation revealed a measurable increase in overall concordance between Nutri-Score grades and NOVA categories following the algorithm update. The rank correlation between the two classification methodologies rose from Spearman’s rho of 0.518 under the original calculation formula to rho of 0.571 under the updated system.

In addition to the increase in overall agreement, the revision significantly reduced outright conflicts between the two assessment systems across the examined supermarket sample. Following the implementation of the updated calculation criteria, the proportion of these Nutri-Score and NOVA classification conflicts declined to 3.1 percent, representing a statistically significant reduction with a p-value of less than 0.001.

Despite the targeted modifications and the significant reduction in classification conflicts, the majority of food items retained their original positions on the five-tier letter scale. Specifically, 68.7 percent of the 450 evaluated products maintained the exact same letter grade after recalculation under the updated algorithm.

The researchers concluded that the updated Nutri-Score algorithm is associated with improved alignment with the NOVA processing scale and an effective reduction in favourable ratings granted to ultra-processed foods found in Polish retail stores.

Subject of Research: Nutri-Score algorithm update alignment with NOVA food processing classification

Article Title: Assessment of the alignment between Nutri-Score and the NOVA classification before and after the Nutri-Score algorithm update: analysis of 450 food products from the Polish retail market

Article References: Kobus-Bogucka, A. E., Olszowska, P., Potoniec, P., Sikora, D., Trzcina, K., & Dobrowolski, H. (2026). Assessment of the alignment between Nutri-Score and the NOVA classification before and after the Nutri-Score algorithm update: analysis of 450 food products from the Polish retail market. Archives of Public Health. https://doi.org/10.1186/s13690-026-02140-6

Image Credits: AI Generated

DOI: 10.1186/s13690-026-02140-6

Keywords: Nutri-Score, NOVA classification, front-of-pack nutrition labelling, ultra-processed foods, food processing, nutritional quality, food products, Polish retail market, dietary recommendations, algorithm update, KomPAN questionnaire, supermarket foods, public health nutrition, food labelling, cross-sectional study