Topical Area: Nutritional Epidemiology
Objectives : To determine the relative validity of a food screener compared with two 24-hour joint parent-child dietary recalls for measuring dietary intakes in low-income Latino children who are overweight/obese.
Dietary intake of children participating in a pediatric obesity weight management program was measured at baseline using both the Spanish-language version of the Block Food Screener for Ages 2-17 (BFS2-17) for the last week and two 24-hour recall interviews (24Hrs) conducted in Spanish on separate days (reference method). Intake data from each 24HR were entered into the Nutritionist Pro software (Axxya Systems) to determine nutrient content of food items. Median intakes were compared for each method by Wilcoxon signed rank test. Reproducibility and relative validity of the BFS2-17 was assessed by Cohen's weighted kappa. Agreement was further assessed using cross-classification by quartiles and Bland-Altman plots.
Results : Complete data on energy and six macronutrients of interest were obtained from 30 participants. Median daily intakes of energy, protein, carbohydrates, and total and saturated fats estimated from the 24Hrs were significantly different than intakes from the BFS2-17 except, for sugar and fiber (Table 1). Cross-classification showed some agreement, with 47%-67% of the participants classified into identical or contiguous quartiles although, none of the results were statistically significant. Bland-Altman plots showed systematic overestimation of the measures assessed by BFS2-17 compared to 24HRs as well as, large amounts of variability around the mean constants.
Conclusions : Results showed poor agreement between the two dietary assessment methods in assessing energy and macronutrient intake among a sample of young Latino children who are overweight/obese. There is a need for culturally-adapted dietary assessment tools for this population.
Funding Sources : Potomac Health Foundation