Validity and reproducibility of food photographic estimation for evaluating meals in evacuation shelters.

Objective: The primary goal was to evaluate the validity of food photographic estimation for nutritional assessment compared with weighed food record (WFR).

Methods: We evaluated the validity and reproducibility of photographic estimation of foods provided in evacuation shelters. We analyzed 35 meals served at 12 shelters in Kumamoto Prefecture in Japan, affected by a heavy rain disaster in 2020. In this context, we compared 21 senior students' portion size estimation by food photographs to WFR. In addition, we assigned five meals for each of the 21 senior students, and the same meal photograph was estimated by three students to test reproducibility.

Results: No statistically significant difference was detected between the two methods regarding energy, the total grams of meal, the protein, and vitamins B₁, B₂, and C, except for salt. In addition, the students who never self-cooked underestimated the total grams.

Conclusion: Food photographic estimation could simplify the nutritional assessment in evacuation shelters. However, unclear photographs and food items served by weight could weaken the estimation accuracy. According to previous studies and the applied postestimation questionnaire, photographs taken from specified angles and reference food photobooks for portion size estimation may improve accuracy.