REPRODUCIBILITY OF A FOOD FREQUENCY QUESTIONNAIRE FOR USE IN OCULAR RESEARCH

Purpose. Assessment of nutritional factors was an ancillary component of the Eye Disease Case-Control Study sponsored by the National Eye In stitute, This multicenter study was designed to evaluate the role of p otential risk factors for a number of retinal disorders. The authors e xamined the reproducibility of the food frequency questionnaire used i n this study. Methods. A semiquantitative food frequency questionnaire , designed for use in this study, was self-administered by participant s. To evaluate the reproducibility of this dietary questionnaire, a su bsample of 325 participants completed a second questionnaire within 12 to 18 months of the first. Pearson product-moment coefficient was use d to assess the correlation between the log of calorie-adjusted nutrie nt scores, and Spearman correlations were used for specific food items . Results. The responses for intake of 60 food items in the questionna ire were reasonably consistent. Correlation coefficients for individua l food items ranged from 0.40 to 0.82. Pearson a correlation coefficie nts for logs of calorie-adjusted intake of nutrients ranged from 0.38 to 0.75. Nutrients of interest in eye disease include total protein (r = 0.57), total fat (r = 0.71), saturated fats (r = 0.69), carotene (r = 0.61), vitamin C (with supplements, r = 0.66), vitamin E (with supp lements, r = 0.69), and zinc (with supplements, r = 0.43). Partial cor relations controlling for age, sex, and clinical center were similar. Conclusions. These findings indicate that the food frequency questionn aire used in this study provides reasonably reproducible dietary infor mation.