Abstract
OBJECTIVE: To assess the reliability and validity of a semi-quantitative FFQ designed to evaluate the usual nutrient intake of adults in Quito, Ecuador. DESIGN: Dietary data using 24 h recalls (24hR) were used to design a list of commonly consumed foods. The relative validity of a 111-item FFQ was evaluated by comparing nutrient intakes against three non-consecutive 24hR. All nutrients were energy-adjusted. Reliability was assessed using two FFQ (FFQ1 and FFQ2) and assessed by the intra-class correlation coefficient. The comparisons between the FFQ and the 24hR were assessed by the de-attenuated Pearson correlation coefficient, weighted kappa and by Bland-Altman plots. SETTING: Quito, Ecuador. SUBJECTS: Overall, 345 adults were enrolled in the present study. Two hundred and fifty participated in FFQ development and ninety-five participated in the FFQ validity and reliability. RESULTS: The FFQ produced higher energy and nutrient intakes. Reliability correlation coefficients after adjusting for energy ranged from 0·62 to 0·88 for protein and Ca, respectively. For the validity study, energy-adjusted and de-attenuated correlation coefficients between the questionnaire and the 24hR ranged from 0·21 for fat to 0·65 for Ca. Only 4 % of the participants were grossly misclassified and 46 % had weighted kappa higher than 0·42. The Bland-Altman plot showed a constant bias with a tendency to increase according to the intake level. CONCLUSIONS: The FFQ showed reasonably good relative validity and reliable measurements, especially for nutrients considered protective and risk markers of non-communicable disease, and can be used to assess usual nutrient intake in this population.