The influence of education in the validation process of a food frequency questionnaire for adults in Viçosa, Minas Gerais, Brazil

• Autor(a) principal: Crispim, S. P
• Data de Publicação: 2006
• Outros Autores: Ribeiro, R. C. L, Silva, M. M. S, Rosado, L. E. F. P, Rosado, G. P
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: LOCUS Repositório Institucional da UFV
• Texto Completo: http://dx.doi.org/10.1038/sj.ejcn.1602458; http://www.locus.ufv.br/handle/123456789/22331
• topic: Validation; Dietary measurements; Methods; Education; Food frequency questionnaire

To evaluate the educational influence in the relative validation of a semiquantitative food frequency questionnaire (FFQs) for adults in the city of Viçosa, Brazil.Four 24-h dietary recalls (reference method) were applied to a sample of 94 adults of both genders, at intervals of 1 month. At the end of the study, an FFQs consisting of 58 food items was also applied. Then, the individuals were divided into two main groups according to their educational level (lower and higher). The dietary data were calculated by the Diet Pro 4.0 software and analyzed according to differences of means or medians and Pearson's correlation coefficients. These coefficients were adjusted by the energy and corrected by the within-person variance for each educational group, considering the extreme quartiles of the data distribution.The intake of energy and nutrients, based on the 24 h dietary recalls, was inferior for the lower educational group (P<0.05). For the FFQs, just the protein and calcium intakes were statistically different, suggesting interference of the education variable in this assessment. Overestimations in the FFQs were identified in the analyses of means and medians for vitamin C and retinol intakes in the lower education group and for retinol in the higher education one. However, when evaluated by correlation coefficients non-adjusted and adjusted, they were well correlated. On the other hand, lipid (r=0.34) and calcium (r=0.13) coefficients of the group with less instruction showed beneath the desirable values, suggesting weak consistence of the estimates provided by the FFQs for these nutrients. Correlation means of r=0.65 and 0.54 were found for the higher and lower educational groups, respectively.The results indicate that FFQs showed acceptable performance on evaluating the habitual food consumption for most of the nutrients in the studied population. A tendency for better quantifications in the groups with higher education was observed, inferring its influence in the assessment of the dietary intake.