Assessing the diet of adolescent girls in the UK.

Abstract

To assess the ability of a food frequency questionnaire (FFQ1) and a food checklist (FCL) to determine energy and macronutrient intakes of adolescent girls in the general population. Energy and macronutrient intakes determined by FFQ1 and the FCL were compared with those from a 7-day weighed dietary record (WDR). The reproducibility of FFQ1 was assessed by a comparison of intakes with those from a repeated questionnaire (FFQ2) completed a month later. Southampton, UK. Forty-seven 15-year-old girls completed FFQ1 and the WDR and FCL; and 61 girls completed FFQ1 and FFQ2. The broad dietary patterns described by the three methods of assessment were similar, although absolute intakes differed. Energy and macronutrient intakes determined by FFQ1 were higher than those recorded in the WDR (all P < 0.001), but intakes assessed by the FCL and WDR were similar. Only FFQ1-assessed energy intakes appeared consistent with predicted energy requirements. With the exception of protein intake, there was reasonable agreement between FFQ1 and the WDR in their estimation of energy and macronutrient intake (range of correlation coefficients 0.28 for energy to 0.33 for carbohydrate). The poorer agreement between FFQ1 and the WDR in their estimates of protein intake arose principally from the misclassification of meat and fish intake, although there was no obvious explanation for this. Energy and macronutrient intakes were similar for FFQ1 and FFQ2. Adolescent girls give reproducible answers in response to a self-administered FFQ, which yield useful information about their broad dietary patterns. The FFQ may be a more suitable dietary assessment method than prospective records for use in general population studies of girls of this age. We discuss suggestions for its improved performance.