Validation of a photographic atlas of food portions designed as a tool to visually estimate food amounts in Ecuador.

Introduction: the use of real images of food portions constitutes a useful and effective tool to help measure the amount of food consumed. Objective: to validate content and visual perception of the images of a photographic atlas of food portions designed for Ecuador. Methods: first, eight experts assessed the content in an atlas of food portions, using the Delphi technique. Then, 56 adults (aged 18-59) gave an assessment of their visual perception of about 35 portions of nine selected products. The concordance in the estimation using the atlas versus an estimation without the atlas was evaluated through the intra-class correlation coefficient (ICC), the Bland-Altman graphical method, and a hypothesis contrast. The differences between the real amounts and the estimation were assessed using the Wilcoxon test (p < 0.05). For each of the food items, the percentage of participants who chose the correct photograph, the one directly above or below was calculated. Results: the assessment carried out by experts showed that this instrument is relevant and appropriate. ICC values of between 0.576 and 0.956 were obtained using the atlas, as well as significant differences between the real amounts and the estimation without the atlas (p < 0.001). There was a sufficient correlation between the actual image and its perception for all food items except mayonnaise. The correct photograph was chosen in 66% of 500 estimations. Conclusions: the photographic atlas of food portions for Ecuador is an appropriate tool for helping to estimate the amount of food consumed by adults.