OBJECTIVE: To examine the utility of using external estimates of within-person variation to adjust usual nutrient intake distributions. DESIGN: Analyses of the prevalence of inadequate intake of an example nutrient by the Estimated Average Requirement (EAR) cut-point method using three different methods of statistical adjustment of the usual intake distribution of a single 24-hour recall in Russian children in 1996, using the Iowa State University method for adjustment of the distribution. First, adjusting the usual intake distribution with day 2 recalls from the same 1996 sample (the correct method); second, adjusting the distribution using external variance estimates derived from US children in 1996; and third, adjusting the distribution using external estimates derived from Russian children of the same age in 2000. We also present prevalence estimates based on naive statistical analysis of the unadjusted distribution of intakes. SETTING/ SUBJECTS: Children drawn from the Russia Longitudinal Monitoring Survey in 1996 and 2000 and from the 1996 Continuing Survey of Food Intakes by Individuals. RESULTS: When the EAR cut-point method is applied to a single recall, the resulting prevalence estimate in this study is inflated by 100-1300%. When the intake distribution is adjusted using an external variance estimate, the prevalence estimate is much less biased, suggesting that any adjustment may give less biased estimates than no adjustment. CONCLUSIONS: In moderately large samples, adjusting distributions with external estimates of variances results in more reliable prevalence estimates than using 1-day data.
OBJECTIVE: To examine the utility of using external estimates of within-person variation to adjust usual nutrient intake distributions. DESIGN: Analyses of the prevalence of inadequate intake of an example nutrient by the Estimated Average Requirement (EAR) cut-point method using three different methods of statistical adjustment of the usual intake distribution of a single 24-hour recall in Russian children in 1996, using the Iowa State University method for adjustment of the distribution. First, adjusting the usual intake distribution with day 2 recalls from the same 1996 sample (the correct method); second, adjusting the distribution using external variance estimates derived from US children in 1996; and third, adjusting the distribution using external estimates derived from Russian children of the same age in 2000. We also present prevalence estimates based on naive statistical analysis of the unadjusted distribution of intakes. SETTING/ SUBJECTS:Children drawn from the Russia Longitudinal Monitoring Survey in 1996 and 2000 and from the 1996 Continuing Survey of Food Intakes by Individuals. RESULTS: When the EAR cut-point method is applied to a single recall, the resulting prevalence estimate in this study is inflated by 100-1300%. When the intake distribution is adjusted using an external variance estimate, the prevalence estimate is much less biased, suggesting that any adjustment may give less biased estimates than no adjustment. CONCLUSIONS: In moderately large samples, adjusting distributions with external estimates of variances results in more reliable prevalence estimates than using 1-day data.
Authors: Amanda L Thompson; Khristopher M Nicholas; Elijah Watson; Enrique Terán; Margaret E Bentley Journal: Am J Hum Biol Date: 2019-10-23 Impact factor: 1.937
Authors: Caitlin D French; Joanne E Arsenault; Charles D Arnold; Demewoz Haile; Hanqi Luo; Kevin W Dodd; Stephen A Vosti; Carolyn M Slupsky; Reina Engle-Stone Journal: Adv Nutr Date: 2021-03-31 Impact factor: 8.701