A hybrid nutrient profiling–processing framework to better inform healthy food choices

Nutrient profiling models such as the Health Star Rating (HSR) are widely used to guide consumer food choices. However, nutrient-based models may not fully account for food processing, potentially overestimating the healthfulness of ultra-processed foods. One suggested solution was considering both nutrient content and processing in food evaluation. This study aimed to test a combination of The Health Star Rating (HSR) with the Brazilian NOVA food classification system using international food composition databases. This study applied a modification to the HSR system by incorporating food processing information based on the NOVA food processing classification. Food products from multiple databases (German Food Code and Nutrient Database (BLS), USDA FoodData Central, McCance and Widdowson’s CoFID, and the Skurikhin database) were categorized by processing level and food group. Baseline HSR scores were compared with modified HSR scores using non-parametric methods. Descriptive statistics and graphical analyses were used to examine score distributions across processing levels and food categories. The study sample comprised 5503 food products and dishes from four databases divided into 13 main food categories. Incorporation of processing-based deductions resulted in systematic reductions in HSR scores for foods classified as more highly processed. Wilcoxon signed-rank tests indicated statistically significant differences ( p ≤ 0.0001) between baseline and modified HSR scores across most categories. Specifically, accounting for NOVA food processing decreased scores in the following food categories: meat and meat products, sauces and spreads, bread and bakery, convenience foods, dairy, cereal and cereal products, and snack foods. Integrating food processing information into nutrient profiling alters HSR score distributions and improves conceptual alignment with processing-based food classifications. These findings suggest that hybrid approaches combining nutrient composition and processing characteristics may enhance the interpretability of front-of-pack labelling systems. Further research is needed to evaluate such approaches against health outcomes.