Within- and Between-Individual Variation in Nutrient Intake in Children and Adolescents

BackgroundLittle is known regarding the number of 24-hour recalls required to rank-order children and adolescents on usual intake for diet–disease studies.ObjectiveTo determine the within- to between-individual variance ratios and number of 24-hour recalls required to rank-order individuals on usual intake for select macro- and micronutrients in a large, multiracial/ethnic sample of children and adolescents.DesignCross-sectional survey.Participants/settingChildren and adolescents ages 6 to 17 years participating in the 2007-2008 and 2009-2010 National Health and Nutrition Examination Survey (NHANES).Main outcome measuresVariance ratios for predefined sex, age (children age 6 to 11 years, adolescents age 12 to 17 years), and racial/ethnic groups (Mexican American/Hispanic, non-Hispanic black, and non-Hispanic white).Statistical analysisMixed-effects linear regression models were used to estimate within- and between-individual variance components for selected nutrients. The number of 24-hour recalls required to rank-order participants on usual intake (absolute values and energy-adjusted) was obtained from the nutrient variance ratios for various levels of accuracy.ResultsVariance ratios were more than 1 for all nutrients examined. High values (variance ratio >3) were observed for protein, saturated and unsaturated fatty acids, cholesterol, and several micronutrients. Variance ratios for absolute nutrient intakes were similar for both sexes within age groups, but higher for children than for adolescents. A total of six to nine and three to six 24-hour recalls were typically sufficient to rank-order children and adolescents, respectively, on usual intake with an accuracy of r=0.8. Additional recalls were required to achieve the same accuracy for energy-adjusted nutrients. Variance ratios were similar for adolescents across racial/ethnic groups, but highly variable in children.ConclusionsA total of six to nine 24-hour recalls may represent a reasonable trade-off between accuracy and participant burden for rank-ordering nutrient intakes in children and adolescents. Additional research is required to determine whether this may be reduced using statistical modeling–based approaches and the number of recalls children and adolescents will reliably complete.