Can the EPIC food-frequency questionnaire be used in adolescent populations?

Objective: To assess the validity of the food-frequency questionnaire used in the European Prospective Investigation of Cancer (EPIC FFQ) for estimating nutrient intake in an adolescent population. Design: Sixty-seven schoolchildren (mean age: 12.3 +/- 0.3 years) were recruited to complete a 7-day weighed dietary record (7-day WDR), the EPIC FFQ and supply one 24-hour urine collection. Setting: Harris Academy in Dundee (UK). Results: Fifty subjects completed both dietary assessment methods. Thirteen of these were classified as underreporters with energy intake/basal metabolic rate < 1.14. The EPIC FFQ showed higher estimates than the 7-day WDR for all nutrients. The median Spearman correlation coefficient for the nutrients examined was found to be 0.31 and increased to 0.48 after adjustment for total energy. The limits of agreement were as far apart as 13.4 MJ, 120 g, 270 g, 120 g and 1170 mg for energy, fat, sugar, protein and calcium, respectively. Correlations between urine and 7-day WDR dietary nitrogen and potassium were found to be statistically significant with r = 0.45 (P < 0.05) and r = 0.78 (P < 0.001), respectively. The median proportion of subjects that appeared in the same and opposite third of intake was found to be 45.9% and 10.8%, respectively. Conclusions: The EPIC FFQ seems adequate to correctly classify low, medium and high consumers and might therefore be used to identify adolescent population groups at risk or for differences between populations. However, agreement between the EPIC FFQ and the 7-day WDR was very poor on both a group and an individual basis, and demonstrates that the EPIC FFQ is not an appropriate method for estimating absolute intakes in this age group.