Elaboração e validação da versão reduzida de um questionário de frequência alimentar para gestantes de uma capital do nordeste do Brasil

Abstract

Pregnancy is a critical period for the mother-child binomial, where knowing the food consumption of the pregnant woman is extremely important, given that this assessment makes it possible to characterize the level of risk and vulnerability to nutritional deficiencies and excesses. The food frequency questionnaire is considered a useful, practical and informative tool to carry out the evaluation of habitual consumption, however, most tools are considered extensive, raising the need to develop short tools, aiming to facilitate the applicability, understanding and data analysis, while maintaining the validity of the tool. Thus, this dissertation is structured in three sections, being na introductory chapter, a literature review chapter and a chapter composed of two scientific articles. The first original article entitled “Food frequency questionnaires developed and validated for pregnant women: a systematic review of the literature”, is a systematic review that aimed to evaluate the methodological quality and the correlation coefficient of the questionnaires developed for pregnant women. Of the fourteen studies selected, half of the tools evaluated were classified as having excellent methodological quality, reflecting validation correlation coefficients classified as moderate and high (r≥0.40). The second original article is a validation study entitled “Elaboration and validation of the reduced version of a food frequency questionnaire for pregnant women residing in the northeast of Brazil”, where, to reduce the list of the original questionnaire containing 112 items, the following was used Pearson correlation, multiple linear regression and frequency of consumption. For the validation analysis, gross, energy-adjusted and attenuated values of energy and nutrients of interest were used, where acceptable correlation values ranged from 0.40-0.70, with p<0.05 considered significant. Also, Bland & Altman graphs were constructed to analyze the agreement between the methods. In this way, the food frequency questionnaire was reduced to 84 items, a 25% reduction compared to the original. The correlation coefficient showed acceptable values for energy (0.40, p<0.001), vitamin B12 (0.50, p<0.001), magnesium (0.85, p<0.001), phosphorus (0.88, p<0.001). 0.001), selenium (0.76, p<0.001), cholesterol (0.62, p<0.001), saturated fat (0.74, p<0.001), monounsaturated (0.84, p<0.001), polyunsaturated (0.68, p<0.001) and fibers (0.78, p<0.001). Furthermore, it was demonstrated that the FFQ overestimated the consumption of energy and nutrients in relation to the 24-hour food recall. Thus, the shortened version of the FFQ developed in this study is a useful and reliable tool to assess the food consumption of pregnant women.