Abstract
Background: Many studies derive dietary information from child self-reported Food Frequency Questionnaires (FFQs). This may be subjected to misreporting, especially among overweight and obese children.
Aim: The aim of this study was to examine the validity of data acquired from child-reported dietary intake using a semiquantitative FFQ developed for assessing dietary habits of overweight and obese children in Greece, using parental 3-day food records of child intakes.
Methods: Validation analysis was based on 106 (from total 115) children (41% boys and 59% girls). Children were asked to report the frequency of their dietary intake, using the FFQ provided. Parents were asked to keep a 3-day food record for their children's intake. Correlations and significance between methods were assessed via Spearman correlation coefficient and Wilcoxon nonparametric pairwise comparisons, respectively. Agreement between the FFQ and the 3-day record was performed using Bland-Altman method.
Results: Significant correlations, ranging from 0.32 to 1 (all P < .05), were observed between food consumption reported in the FFQ and recorded in the 3-day dietary record. High correlation was found for fruits ([rho] = 0.988), vegetables ([rho] = 0.985), dairy ([rho] = 0.702), meat ([rho] = 0.958), fish ([rho] = 0.841), starchy foods ([rho] = 0.793), sweets ([rho] = 1), and beverages ([rho] = 0.978). Medium correlation was observed only between the consumption of legumes ([rho] = 0.329). No significant differences were found between reported FFQ and 3-day dietary record for most food groups and beverages examined. Mean intake agreement was ranged from 90.6% to 98.1% (Bland-Altman).
Conclusions: The FFQ used appears to be a valid tool for investigating dietary intake of food among overweight and obese children.