Background/objectives: The objective of this study was to evaluate the performance of a short food frequency questionnaire (FFQ) to assess dietary intake at 4 and 7 years of age, against 3d food diaries (FD) and serum biomarkers, using two methods to convert the FFQ to daily intake in grams and nutrients (standard and z-score method).
Subjects/methods: The present analysis comprises data from 2482 4-year-old children and 3511 7-year-old children, from the birth cohort Generation XXI (Porto, Portugal). To estimate daily consumption from the FFQ, the frequency response was multiplied by a standard mean portion (standard method) or adjusted with data from the FD (z-score method). The dietary intake obtained from the FFQ was compared with the FD and serum biomarkers, using Intra-Class Correlation Coefficients (ICC), de-attenuated Pearson's correlation coefficients and Bland Altman analysis.
Results: In general, the mean daily food intake estimated by the z-score method had a higher agreement with the FD, than the standard method. The highest ICC was obtained for "vegetable soup" (ICC = 0.536), using the z-score method, compared to an ICC of 0.373 using the standard method. Significant correlation coefficients were observed for all nutrients; the average of correlation coefficients was 0.39 at 4 years and 0.42 at 7 years of age. For the majority of nutrients, the correlation between mean and mean difference was lower using the z-score method, in comparison with the standard method.
Conclusions: The results suggest that the FFQ is a reasonably good instrument to estimate dietary intake in children. Moreover, adjusting the FFQ portion size, by using a z-score method, seems to increase the accuracy of dietary data in children.