The relative validity of vitamin intakes derived from a food frequency questionnaire compared to 24-hour recalls and biological measurements: Results from the EPIC pilot study in Germany

Background. For the European Prospective investigation into Cancer and Nutrition (EPIC) study Germany. a self-administered food frequency questionnaire (FFQ) was developed and tested for its relative validity and reproducibility in 1991/1992. Study participants were 92 potential cohort members. This paper reports results regarding retinol, carotenoids, tocopherols and ascorbic acid. Methods. Study participants were invited to the study centre in Heidelberg once a month over one year. At each visit, a 24-hour recall was obtained. The FFQ was filled in twice with a 6-month interval (FFQ1, FFQ2). In addition, a questionnaire on general consumption frequencies of 14 broad food groups was completed. This information was combined with estimates derived from FFQ2 and frequency-corrected food and nutrient intakes were calculated (FFQ(corr), Blood specimens were taken in winter and summer 1992. Results. The intraclass correlation of the FFQ ranged from 0.65 to 0.67 for retinol, tocopherols, carotenoids, and ascorbic acid. Intake of carotenoids by FFQ(corr) showed de-attenuated Pearson correlation coefficients with blood values in the order of 0.37, and with recall data of 0.44. Respective correlations for retinol were 0.27 and 0.29, for tocopherols 0.18 and 0.52, and for ascorbic acid 0.36 ana 0.69. Errors of FFQ(corr) and 24-hour diet recalls were not correlated. Conclusions. In general, it was demonstrated that the FFQ was able to rank participants into biologically meaningful categories of intake or blood concentrations for carotenoids and ascorbic acid, but misclassification was higher for tocopherol and retinol.