Agreement between a new web-based and a legacy paper-based food frequency questionnaires in French-speaking Switzerland

Updated and validated tools are crucial for assessing dietary intake. In French-speaking Switzerland, the only validated food frequency questionnaire (FFQ) was a paper-based version developed using data collected in 1990s. A new Swiss electronic FFQ (eFFQ) has recently been developed based on nationally representative dietary data from 2010s. The eFFQ has been recently validated and is intended for adoption in future studies among French-speaking adults living in Switzerland. Both FFQs differ in format (paper vs. electronic), food lists, frequency categories, portion sizes, and food composition databases. We aimed to compare dietary intake estimates from the recently developed eFFQ vs. the paper-based FFQ. We recruited 75 volunteers aged 18 to 75 years (54.7% female) who completed both FFQs in random order. Intakes of energy, 10 macronutrients, four micronutrients and 28 food groups were estimated using each FFQ and compared using multiple statistical tests. Group-level estimates of energy and most nutrients showed good agreement between FFQs, with differences in median intakes < 10% for 8 of 14 nutrients, including energy, protein, carbohydrates, total sugars, total fat, and saturated fatty acids (SFAs). Largest differences were observed for alcohol (+ 119% in the eFFQ), fibre (+ 37%), polyunsaturated fatty acids (+ 27%), and cholesterol (+ 26%). Spearman’s correlation coefficients (SCCs) were ≥ 0.50 for 12 of 14 nutrients, and misclassification into opposite tertiles (> 10% of participants) was observed only for SFAs (13%). Bland-Altman analysis confirmed acceptable agreement for energy and macronutrient intakes. At the food group level, the eFFQ yielded significantly higher median intakes for 14 of 28 food groups – most notably plant-based protein-rich foods and cream, fatty sauces & other fats – and lower median intakes for four groups, including vegetables and fish. Ranking agreement was acceptable for most food groups (SCCs = 0.57–0.91) but weaker for water (0.07); vegetable oil (0.42); cream, fatty sauces & other fats (0.33); and mixed dishes & soups (0.47). The new eFFQ demonstrated good relative agreement with the previous paper-based FFQ, particularly for ranking individuals. These findings provide valuable information for researchers and policy makers comparing results of both FFQs over time and across populations.