Reproducibility and relative validity of food group intake in a food frequency questionnaire developed for the German part of the EPIC project

Background. For the German part of the European Prospective Investigation into Cancer and Nutrition (EPIC), a self-administered, optically-readable food frequency questionnaire (FFQ), including 158 food items and 87 coloured portion size photographs, was developed to assess the usual food and nutrient intake of individuals during the past year. In 1991/1992, the reproducibility and validity of the questionnaire measurements were studied according to the EPIC protocol. This article reports the results on reproducibility and relative validity of measurement of food group intake. Methods. A total of 104 men and women aged 35-64 years, who are members of the local health insurance institution, AOK Heidelberg, participated in this study. Reproducibility of the questionnaire measurements was obtained by a repeated administration of the FFQ to the same study subjects at a 8-month interval. The mean of 12 24-hour dietary recalls applied at monthly intervals served as reference method for the estimation of the relative validity of questionnaire measurements. A second version of the FFQ that integrated questions on general food consumption patterns was also investigated. Results. Spearman test-retest correlations of food group intake ranged from 0.49 for bread to 0.89 for alcoholic beverages (median = 0.70). Spearman correlations between food group intake values derived from the 24-hour diet recalls and the FFQ completed in the summer of 1992 varied from 0.14 for legumes to 0.90 for alcoholic beverages (median = 0.45). Correction for attenuation due to within-person error in the reference method as well as the correction for general consumption patterns improved the correlations. Conclusion. The results indicate that our newly developed FFQ gives reproducible estimates of food group intake. Large day-to-day variation in food group intake complicated the evaluation of FFQ validity. Overall, moderate levels of relative validity were observed for estimates of food group intake.