Validity of Food Atlases as Portion-Size Estimation Aids in Dietary Assessment: A Scoping Review

Abstract Accurate portion-size estimation is a key element of practical dietary assessment. Food atlases, visual representations of various portion sizes, have become increasingly popular estimation tools. However, their validity remains underexplored. This scoping review aims to evaluate the validity of food atlases in diverse dietary assessment contexts. A comprehensive search for relevant articles published between 2010 and February 2025 was conducted in 3 primary databases: PubMed, Scopus, and Web of Science. The inclusion criteria were original English-language studies that assessed the validity of food atlases as aids for portion-size estimation, regardless of population characteristics. Data extraction followed the PCC (Population, Concept, Context) framework and PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews) guidelines. Of 241 articles identified, 27 met the eligibility criteria. Most studies (37%) were published between 2018 and 2021 and originated from 25 countries, primarily focusing on adult populations (59.3%). Both print (70.4%) and digital (22.2%) versions of food atlases were evaluated. Most studies reported acceptable validity for food atlases, although estimation accuracy varied widely between individuals. Amorphous foods and liquids exhibited lower validity. Overestimation was more frequent for smaller portions, while larger portions tended to be underestimated. Some studies reported no significant differences in estimation accuracy between printed and digital formats. Food atlases are potentially valid for estimating portion sizes at the population level; however, the accuracy at the individual level is variable. Accuracy may be improved through cultural adaptation, enhanced visual design, and complementary estimation tools. This review highlights key methodological considerations to guide future development and validation of food atlases.