Poster Topical Area: Global Nutrition

Location: Hall D

Poster Board Number: 563

P12-033 - Application of External Variance Estimates for Modeling Usual Nutrient Intake Distributions: Literature Review and Simulation Analysis

Sunday, Jun 10
8:00 AM – 6:00 PM

Objectives: Estimating population distributions of usual nutrient intakes is a crucial step in planning and evaluating nutrition interventions. Where replicate dietary recalls are not available, external estimates of within-individual variation (WIV) can adjust for day-to-day variation in nutrient intake to estimate usual intake distributions. However, WIV may vary by population. This study aims to: 1) describe variability in WIV estimates across a range of populations and settings for selected nutrients; and 2) test the effect of using different external WIV values on estimated prevalence of inadequate usual intakes.

A literature search was conducted to compile reported WIV estimates. Included papers described studies that collected 2 or more replicate 24 hour recalls or food records and reported WIV of nutrient intakes. To assess the effect of variability in assumed WIV on prevalence of inadequate intake (intake below the estimated average requirement), usual nutrient intake distributions were modeled using IMAPP software using a national survey conducted in Cameroon (n=537) and a survey in 2 rural districts in Bangladesh (n=460). Simulations of intake distributions for vitamin A, folate, and zinc in non-pregnant, non-lactating women aged 19-50 years were performed using single-day intake data and external WIV estimates (expressed as proportion of total variance) ranging from 0.5 to 0.9.

Of 76 publications reporting on replicate dietary measures, 7 reported WIV values, and these varied in the reported unit, nutrients, and age and sex groups. In simulations, varying WIV had a small or no effect on prevalence of dietary folate inadequacy, which ranged from 59% to 62% for Cameroon and was consistently 100% for Bangladesh. Simulated prevalence of inadequacy ranged from 92-99% and 47-63% for vitamin A, and 72-85% and 6-30% for zinc, for Bangladesh and Cameroon respectively.

Estimates of WIV in nutrient intake are rarely reported in the literature, and calculation methods vary. The effect of assumed WIV on prevalence of inadequate intake varied by nutrient and setting. This may be important when determining if a potential public health problem exists, such as when the prevalence of inadequacy is in a low to moderate range, as was seen with zinc in Cameroon.

Funding Source: Funding for this research was provided by the Bill & Melinda Gates Foundation. The survey in Cameroon was funded by the Michael and Susan Dell Foundation, and the survey in Bangladesh was funded by HarvestPlus.

CoAuthors: Joanne Arsenault, Ph.D., M.P.H., R.D. – University of California, Davis; Charles Arnold, M.S. – University of California, Davis; Hanqi Luo, Ph.D. – University of California, Davis; Reina Engle-Stone, Ph.D. – University of California, Davis

Caitlin D. French

Doctoral Student
University of California, Davis
Davis, California