Food Frequency Questionnaires (FFQs) are valuable research tools in nutritional epidemiology. This study aimed to develop and validate a new semi-quantitative FFQ, specifically designed for the Italian population and best fitted for self-administration. During the development process, we adapted to Italian needs the validated FFQ proposed by the Fred Hutchinson Cancer Research Center, revising food items, food frequency scale, portion sizes, and time frame. To assess the validity of the proposed FFQ, we compared the estimated daily intake using FFQ with the mean of 3-day food diaries and one 24-hour recall (considered as reference method). The validation process was conducted among a cohort of 51 healthy subjects enrolled in a clinical trial. Four statistical tests were applied on 23 estimated nutrient intakes. Spearman's coefficients ranged from 0.223 (sodium) to 0.748 (alcohol) and were good (≥0.50) and acceptable (0.20-0.49) for 7 and 16 nutrients, respectively. Cross classification showed a good agreement (≥50% in the same tertile or ≤10% in the opposite tertile) for 7 nutrients. The weighted Cohen's kappa values indicated an acceptable outcome (0.20-0.60) for 13 nutrients. Bland Altman plots did not show heteroscedasticity in the error terms, despite the presence of a bias. Our study provided a new Italian semi-quantitative FFQ for self-administration with an acceptable validation level. Its definitive release requires additional refinements and efforts.
Keywords: Bland Altman; Food frequency questionnaire; Nutritional assessment; Nutritional epidemiology; Validation.
Copyright © 2020 Elsevier Inc. All rights reserved.