OBJETIVE: To validate a semi quantitative food frequency questionnaire to assess the food intake of adults in the city of Viçosa, Minas Gerais. METHODS: Ninety-four adults of both genders and representing different educational levels. As a reference method for the validation study four 24 hour dietary recalls were used with intervals of one month in between. A semi quantitative food frequency questionnaire containing 58 food items was applied at the end of the study. The dietetic data was calculated using Diet Pro software and analyzed according to difference of means or medians and Pearson's correlation. The correlation coefficients were adjusted for energy and corrected for intrapersonal variance. Cross classification was used for comparing both methods. RESULTS: Differences of means or medians revealed underestimation of the semi quantitative food frequency questionnaire on determining the consumption of micronutrients (vitamin C, retinol, and calcium). However, when evaluated by the correlation coefficients and corrected for the variance, these micronutrients presented adequate values; lipid was the only nutrient that indicated low consistency in the dietary assessment among the methods (r=0.33); the other nutrients showed a correlation varying from r=0.40 to r=0.76, with a mean of r=0.52. The correct classification of the methods presented a non-perfect agreement, but acceptable among the evaluated methods. CONCLUSION: The performance of the semi quantitative food frequency questionnaire in the assessment of the habitual food intake was acceptable for most of the nutrients in the studied population. The studied questionnaire can be adapted and used in other populations, but a new validation study is needed if the characteristics of the population differ.
Nutritional epidemiology; Dietary surveys; Methods; Questionnaire
