EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Can existing food composition databases support ultra-processed food research?

Jimmy Chun Yu Louie

Journal of Food Composition and Analysis · 2026-11 · 159 · 109515

DOI: 10.1016/j.jfca.2026.109515 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Food Composition and Analysis, 2026-11. Verfasst von Jimmy Chun Yu Louie. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Much research on ultra-processed food (UPF) assigns foods from dietary surveys, questionnaires, recalls, purchase records and menus to Nova groups using food composition databases (FCDBs) as the coding substrate. This Commentary examines whether existing FCDBs can support that task. Generic composition tables remain the substrate for most epidemiological coding. They provide food names and nutrient values, but usually not ingredients, additive functions or the degree and purpose of processing. These decisive attributes are therefore assumed rather than observed. In one branded database, access to ingredient lists increased the proportion of products classified as Nova Group 4 from 54% to 70%. Branded databases narrow this gap, but rapid reformulation and market turnover limit their value for retrospective coding. Structured systems such as FoodEx2 offer richer facet-based description, yet do not themselves provide validated processing categories. I argue that, in many datasets, a Nova assignment is best understood as an interpreted proxy rather than a measured attribute. Its evidential basis, assumptions and uncertainty should be reported. I propose separating the descriptors stored in a database from the classification schemes applied to them, so that coding decisions become auditable and testable through sensitivity analyses.

Zitieren

APA

Louie, J. C. Y. (2026). Can existing food composition databases support ultra-processed food research?. Journal of Food Composition and Analysis, 159, 109515. https://doi.org/10.1016/j.jfca.2026.109515

Vancouver

Louie JCY. Can existing food composition databases support ultra-processed food research?. Journal of Food Composition and Analysis. 2026;159:109515. doi:10.1016/j.jfca.2026.109515

BibTeX
@article{Louie2026,
  author = {Louie, Jimmy Chun Yu},
  title = {Can existing food composition databases support ultra-processed food research?},
  journal = {Journal of Food Composition and Analysis},
  year = {2026},
  volume = {159},
  pages = {109515},
  doi = {10.1016/j.jfca.2026.109515}
}
Kostenlos starten →