Comparison of self- and surrogate-reported dietary information.

Studies of diet's role in the etiology of rapidly fatal diseases may utilize data taken from surrogate sources. To assess such sources, 46 subject-spouse pairs were interviewed with a food frequency questionnaire designed to provide an index of vitamin A consumption. Information concerning amount and past pattern of use was also obtained. The frequency and amount information was used to calculate two aggregate indices of vitamin A consumption: one based on frequency alone and the other based on frequency and amount. For single foods, the mean frequencies of consumption reported by subjects and by their spouses for them were similar; for both sexes combined, the average level of exact agreement was 66 per cent, with improvement to 93 per cent for agreement within one category. Similar agreement was found for amount. For the overall daily vitamin A intake of men, the means based on subject data were not significantly different from those calculated from their wives' responses. For women, husbands underreported their total intake. Agreement between subject- and surrogate-based overall vitamin A consumption was less satisfactory than for the individual foods.