Evaluation of the precision and accuracy of a food registry survey with digital image support in real time

Abstract

Introduction: Food surveys are used to determine the consumption of food and nutrients at the population and/or individual level, although they underestimate the intake by 37% or more. The introduction of new technologies such as the digital image of the portion of consumption could help to reduce the estimation error for calories and certain macronutrients. Objective: To evaluate the precision and accuracy of the food record method supported with real-time digital imaging in the estimation of caloric intake and macronutrients in young adults, considering the weighed food record (WFR) as a reference method. Methods: Lunch intake was evaluated previously submitted to a WFR of 58 subjects. Before and after the intake, each participant captured an image of their consumption portions and completed a food record. The nutritional status of the subjects was determined. The statistical analysis was performed with medians, Kruskal-Wallis, and concordance with Bland-Altman (p < 0.05). Results: The highest proportion of subjects presented normal nutritional status (72%) and 80% presented average cardiovascular risk. Significant differences were found between the three methods when estimating the intake for calories, proteins, lipids. The underestimation of calories and lipids decreases from 32% to 25.8% and lipids from 47% to 24.4% when using a digital image to support the estimation of intake. Conclusions: The food record complemented with digital image of the consumed portions presents better agreement with the reference method (WFR) in estimation of calories, proteins and lipids, improving the precision and accuracy of the method.