Reproducibility and validity ELSA-Brasil Food Frequency Questionnaire

Abstract

Background: To investigate the association between diet and health effects in population surveys, it is necessary to have precise tools that allow estimating the habitual consumption of the population. Introduction: To investigate the association between diet and health effects in population surveys, it is necessary to have precise tools that allow estimating the habitual consumption of the population. Aim: To evaluate the reproducibility and validity of the ELSA-Brasil food frequency questionnaire (FFQ) to assess macro and micronutrients intake. Material and methods: We collected dietary information of 281 participants which completed two ELSA-Brasil food frequency questionnaires (FFQ) over a year period and three records. To assess the reproducibility of the FFQ, we compared the macro and micronutrients intake from the two FFQ while to assess the validity, we compared the intakes of FFQ with the mean of three records. The intraclass correlation test (ICC) and agreement percentages of nutrient intake were calculated after categorization by tertiles. Results: ICC coefficients for reproducibility ranged from 0.51 (polyunsaturated fat) to 0.70 (magnesium) while the ICC coefficients for validity ranged from 0.14 to 0.61 for omega 3 and magnesium, respectively. The exact concordances between methods ranged from 37% for omega 3 to 50.2% for magnesium (mean = 44.6%). An average disagreement of 13.4% was found. Conclusions: This study suggests that the ELSA-Brasil FFQ is suitable tool to assess dietary intake with a satisfactory reproducibility and relative validity.