Development and Validation of a Food Frequency Questionnaire for Preschool Children Using Multiple Methods

Background: The ability to determine the relationship between diet and health outcomes in children requires reproducible and validated long-term dietary assessment tools such as food frequency questionnaire (FFQ). Objective: To test the reproducibility and relative validity of a FFQ for young children using 24-hour food recalls (24HRs), anthropometric measurements, and a comprehensive feeding practices questionnaire (CFPQ). Methods: Children (aged 5-6) and their mothers were recruited during one school-year (2008) from preschools. Children's anthropometric measurements were obtained. Mothers provided during a personal interview on three occasions a 110-item semiquantitative FFQ, 24HRs and CFPQ. Pearson-correlation coefficients were calculated between the results of the FFQ and 3*24HR. Validity coefficients between the FFQ and the different measurements were calculated. Scores of the 12 factors of the CFPQ were calculated and related to dietary intake. Results: Sixty-six healthy children (47% boys) were recruited. Pearson's correlations between the average of the FFQs and 3*24HRs ranged from 0.3-0.6 (P<0.05). The highest correlation coefficients were 0.59 for total fat intake and 0.56 for energy. Dietary intake of energy and carbohydrates differed significantly (P=0.05, 0.001 respectively) across the three BMI z-score levels (normal-weight, overweight, obese) and the three waist circumference tertiles (0.019, 0.006 respectively). Obesogenic factors from the CFPQ correlated with consumption of empty calories like sweets, snacks, junk foods and sweet drinks. Conclusions: The modified FFQ is a relatively valid instrument to estimate mean energy intake in preschool children. The questionnaire performs reasonably well to rank children with respect to macronutrients intake as well as obesogenic food groups.