Development, validity, and reproducibility of a semiquantitative food frequency questionnaire for the assessment of antioxidant vitamins intake in Lebanon.

Abstract

OBJECTIVES\nThe growing interest in the possible role of antioxidant vitamins in many pathologies makes the methods of evaluating the intakes of these vitamins necessary. The food frequency questionnaire is the most widely used tool in epidemiologic studies and large-scale surveys to assess long-term nutritional exposure and to determine the usual intake of specific foods or nutrients. The aims of this study were to develop and investigate the validity and reproducibility of a semiquantitative food frequency questionnaire (SQFFQ) to evaluate the antioxidant vitamins intake in the Lebanese population and for use in clinical trials.\n\n\nMETHODS\nAn SQFFQ draft was built with the major sources of vitamins with antioxidant properties initially selected from food that have the higher percentages of the daily value of vitamins A, C, and E; listed in the US Department of Agriculture tables. We used food intake data that were available through a 24-h dietary recall (3 weekdays and 1 weekend day) among a subsample of 50 participants to complete the SQFFQ. We sought the major sources of the most frequently reported antioxidant vitamins, portion sizes, and percentages of the daily value of the major sources of these vitamins (developed by the US Department of Agriculture) to build a 34-item SQFFQ. The consumption frequency was coded as daily, weekly, monthly, or occasionally and one to three times or more. Five hundred Lebanese participants (the previous 50 who provided the 24 h dietary recall, along with 450 new participants) completed the questionnaire, whereas only 50 and 42 were accepted to be enrolled in the validity and reproducibility studies, respectively. To assess validity, the questionnaire was compared with the 24-h dietary recall performed on a subsample of 50 participants who already completed the SQFFQ. The reproducibility was assessed by comparing baseline SQFFQ with a second administration of the SQFFQ 3 to 5 wk later at the same dietetic clinic.\n\n\nRESULTS\nThe means of daily consumption of vitamins A, C, and E based on the SQFFQ and the 24 h data were as follows: Vitamin A: 252.5 ± SD 958.9 versus 242.6 ± SD 580.5 mcg of retinol activity equivalents (µg RAE), Vitamin C: 86.5 ± SD 59.9 versus 85.5 ± SD 56.2 µg, and Vitamin E: 17 ± SD 7.2 versus 17.1 ± SD 7.9 µg. The Pearson correlation coefficients of the three vitamins between dietary intake estimates derived from the SQFFQ and 24-h data were, respectively, r\u202f=\u202f0.83, 0.98, and 0.9; they were highly significant at P\u2009<\u20090.001. One sample t tests and Bland-Altman plots confirmed this validity. The reproducibility of the SQFFQ was assessed using the Cronbach α index that measures reliability between different components (for vitamin A: 0.994, vitamin C: 0.998, and vitamin E: 0.975). The correlation matrix (interitem correlation) showed that all elements were highly correlated (r > 0.4) and even close to 1 (for vitamin A\u202f=\u202f0.989, vitamin C\u202f=\u202f0.995, and vitamin E\u202f=\u202f0.95).\n\n\nCONCLUSIONS\nIn this study, we described the development of an SQFFQ. Based on the results, we found the SQFFQ to be valid and reproducible to assess intake of antioxidant vitamins in Lebanese adults..