A diet history questionnaire ranks nutrient intakes in middle-aged and older men and women similarly to multiple food records.

We evaluated the reproducibility of a modified version of the National Cancer Institute diet history questionnaire and also studied comparability of nutrient estimates from this questionnaire to those from four 2-d food records. Subjects (n = 211) were from a population-based sample of middle-aged and older adults participating in the Beaver Dam Eye Study in south-central Wisconsin. Median age-specific correlation coefficients between the questionnaire, administered twice at a 3-mo interval, were 0.8 in men and 0.7 in women (range = 0.5 to 0.9). Questionnaire estimates of protein, cholesterol, thiamin, niacin, iron and zinc were consistently lower than food record estimates whereas estimates of fat were higher. Correlation coefficients between estimates from the questionnaire and records were generally > 0.5, indicating overall good agreement in ranking. However, the range in correlation coefficients was wide [ranging from 0.06 for iron (without supplements) in middle-aged women to 0.8 for alcohol in middle-aged men and women]. Questionnaire estimates similarly classified persons into the lowest two food record quintiles of intake 50 to 93% of the time depending on nutrient. These results suggest that the questionnaire produces nutrient estimates that rank individuals on the basis of intake of most nutrients similarly to estimates from multiple food records.