Diet data collected using 48-h dietary recall : within—and between-person variation

Abstract

Background and Aims: Forty-eight-hour dietary recall is a valuable source of information regarding food consumption in a population-based sample. This method covers 2 consecutive days in a single interview. Nevertheless, the number of assessments and the sample size necessary to estimate usual intake are unknown. We aimed to assess sources of variation, sample sizes, and numbers of days necessary to estimate usual nutrient intake using the 48-h dietary recall. Methods: This was a population-based cross-sectional study including 237 participants, 11–90 years old, selected using multistage probabilistic sampling to obtain data using 48-h dietary recall. Analysis of variance was used to calculate within- and between-person variation and determine the statistical parameters necessary to calculate sample size and the number of days required to calculate the usual energy and nutrient intake. Results: Within-person variation was generally lower than between-person variation, except for calcium (CV2 w = 40.8; CV2 b = 38.4%), magnesium (CV2 w = 27.4; CV2 b = 18.7%), and monounsaturated fat (CV2 w = 20.0; CV2 b = 17.3%) for the entire group and magnesium for women (CV2 w = 28.3; CV2 b = 91.8%). The number of days and sample size required to determine usual energy and nutrient intake varied substantially with gender and age (e.g., vitamin C in women N = 9, in men N = 1,641). Conclusions: Energy and nutrient intake assessment using the 48-h dietary recall misrepresents within-person variation but can generate acceptable results for between-person variation. The calculation of sample size and number of days required to determine usual energy and nutrient intake might have been affected by inadequate assessment of the within-person variation.