An evaluation of a colour food photography atlas as a tool for quantifying food portion sizes in epidemiological dietary surveys

Objective: To test the validity of a colour food photography atlas for quantifying portion size eaten compared with weighed foods. DESIGN: The colour food photography atlas was prepared by cooking, weighing and taking digital photographs of three portion sizes of 434 foods and beverages typical of the Italian diet. SUBJECTS AND INTERVENTIONS: In all, 448 male and female volunteers aged 6-60 y from a wide variety of social backgrounds completed 9075 assessments of food portions eaten at lunch and dinner in relation to a set of colour food photographs during 8 weeks of investigation. The amounts of foods eaten by individuals in five different cafeterias in Pavia, Northern Italy, were weighed by trained investigators at the time of serving and, within 5-10 min of the end of the meal, each subject was asked to quantify all foods consumed with reference to one of the three food photographs or in terms of virtual portions among those shown in the photographs. RESULTS: Multiple regression analysis shows that weights of portion sizes chosen from the set of photographs are significantly associated (P<0.05) to weights of eaten portions (beta=0.81; R(2)=0.70) and are independent of age, gender and BMI. The differences between mean weights of the portions chosen by individuals from photographs and mean weights of eaten foods are significant for all food categories (P<0.05), except for bread. However, because of the very large number of observations, the mean differences are very small (range: from +23.2 g (+11.2%) for first courses to -1.3 g (-2.7%) for bread). Bland-Altman plots show that first courses limits of agreement are wide because the dispersion is increasing while weights are rising. CONCLUSIONS: The use of a series of three photographs and virtual portion sizes being associated with relatively small errors, our findings support the validity of using this colour food photography atlas as a tool for quantifying food portion size in epidemiological dietary surveys on different age groups of Italian subjects.