Validity and reproducibility of an interviewer-administered food frequency questionnaire for healthy French-Canadian men and women

OBJECTIVE To evaluate the validity (study 1) and the reproducibility (study 2) of an interviewer-administered food frequency questionnaire (FFQ). METHOD The FFQ was designed at Laval University and contains 91 items and 33 subquestions. STUDY 1: The FFQ was compared against a 3-day food record (2 week-days and 1 weekend-day), at week 0, 6 and 12 of a nutritional intervention. STUDY 2: In order to evaluate the reproducibility of the FFQ, 2 registered dietitians administered the FFQ 4-weeks apart among subjects who were not part of the nutritional intervention. RESULTS STUDY 1: Mean values for intake of most nutrients assessed by the FFQ and by the 3-day food record were not statistically different. Energy-adjusted correlation coefficients for major macronutrients ranged from 0.36 for proteins to 0.60 for carbohydrates (p < or = 0.01). Agreement analysis revealed that on average, 35% of the subjects were classified in the same quartile when nutrients were assessed by either the 3-day food record or the FFQ. STUDY 2: Significant associations were observed between dietary measurements derived from the two FFQs administered 4 weeks apart. Correlation coefficients for the reproducibility of macronutrients ranged from 0.66 for carbohydrates to 0.83 for lipids after energy adjustment. On average, 46% of the subjects were classified in the same quartile when nutrient intakes were assessed by either FFQ. CONCLUSION These data indicated that the FFQ developed has a good validity and is reproducible.