Comparison of food processing classification frameworks in the NHANES 2017-2018

Food processing has become an important focus in nutrition science due to its link to adverse health outcomes. Evaluating different classification systems provides valuable insights into how food processing is defined and its implications for nutrition research, cardiometabolic health, and food policy.

We aim to compare the NOVA, International Agency for Research on Cancer (IARC), International Food Information Council (IFIC), and University of North Carolina classification systems by evaluating their characteristics, reliability, and associations with nutrient intake and cardiometabolic risk factors.

Using data from 4392 adult participants in the 2017 to 2018 NHANES cycle, we categorized 4605 food and beverage items according to 4 classification systems and harmonized processing levels into 3 categories: minimally processed, processed, and highly processed/formulated foods. Sociodemographic characteristics, nutrient profiles, and associations with cardiometabolic risk factors were analyzed using linear regression models that accounted for survey weights.

The mean age of the population was 47 y, 52% were female, and 63% were white. Inter-rater agreement between 2 coders ranged from 68% to 86%. IARC classified the largest proportion of food and beverage items as highly processed/formulated, whereas IFIC classified the lowest. Across all systems, the top contributors to highly processed/formulated food intake were grain products; meat, poultry, fish, and mixtures; and sugars, sweets, and beverages. Greater intake of highly processed/formulated foods was consistently associated with younger age, White and Black race, lower education, and lower intake of protein, fiber, and several micronutrients. All systems linked greater highly processed/formulated food consumption to higher body mass index and elevated levels of log-transformed high-sensitivity C-reactive protein.

Despite differences in classification criteria across systems, associations with nutrients, sociodemographic, and cardiometabolic parameters were consistent. A standardized classification system is needed to ensure a unified understanding of ultraprocessed food and its health implications.