OBJECTIVES: to evaluate the influence of level of schooling on the reproducibility of a quantitative food frequency questionnaire (QQFA), previously designed for pregnant women, to estimate nutrient intake pregnancy. METHODS: the QQFA was applied in duplicate (between 15 and 45 days) to a sample of 95 pregnant women. The intra-class correlation coefficient (CCI) was calculated and the pregnant women classified into quartiles of nutrient estimates, stratified by eight years of schooling. RESULTS: after adjustment for energy, the mean CCI was 0.66 (varying between 0.33 for linoleic fatty acid and 0.88 for potassium and folate) and 0.72 (varying from 0.44 for arachidonic fatty acid to 0.84 for potassium) among applications of the QQFA to pregnant women with lower and higher levels of schooling, respectively. On average, 78% and 84% of pregnant women were classified in the same quartile or in adjacent quartiles on subsequent applications of the QQFA to pregnant women with lower and higher levels of schooling, respectively. CONCLUSIONS: the questionnaire presented a high level of precision in estimating nutrient intake during pregnancy, except for some fatty acids. The QQFA was shown to be more accurate among women with more than eight years of schooling.
Food consumption; Reproducibility of results; Pregnant women; Educational status
