Food classification systems that focus on processing and formulation have gained traction in research and dietary policies. Yet, their utility and scientific foundations have been debated. To address criticisms and identify paths forward, the Institute for the Advancement of Food and Nutrition Sciences convened a tripartite Working Group and independent Writing Team. Drawing on expert collaboration, a targeted literature review, and stakeholder feedback, the Writing Team developed these principles to guide the development and application of such food classification systems for research that supports public health: 1) Documentation and definitions that allow for reproducibility, rigor, and transparency should be provided; 2) Properties for which there is evidence of a biological link with a health-related endpoint should be used to differentiate foods; 3) Associations without robust causal evidence should be considered preliminary; 4) The impact that processing steps have on the final composition and structure of the food in terms of a putative effect on a health-related endpoint should be considered; 5) The impact of formulation on the final composition and structure of the food in terms of a putative effect on a health-related endpoint should be considered; 6) Systems should evolve over time to reflect advancements in science and changes in the food supply, with previous versions of a system being distinguishable from updated versions; 7) Current scientific evaluations from scientific bodies with relevant expertise should be consulted for each iteration; 8) The context(s) in which a system was validated should be considered in its application; and 9) The probative value of a research question or proposed food classification system should be considered prior to engaging in analysis or development. As understanding of food processing, formulation, and health evolves, these principles can serve as a foundation for designing classification systems that support impactful research and, through this, public health policy.
Bernstein et al. (Wed,) studied this question.