Background: Observed-Weighed Food Record Surveys (OWFR) are regarded as the most precise dietary assessment methodology, despite their recognized shortcomings, which include limited availability, high cost, small samples with uncertain external validity that rarely include all household members, Hawthorne effects, and using only 1 or 2 days to identify "usual intake." Although Household Consumption and Expenditures Surveys (HCES) also have significant limitations, they are increasingly being used to inform nutrition policy
Objective: To investigate differences in fortification simulations based on OWFR and HCES from Bangladesh.
Methods: The pre- and postfortification nutrient intake levels from the two surveys were compared.
Results: The total population-based rank orderings of oil, wheat flour, and sugar coverage were identical for the two surveys. OWFR found differences in women's and children's coverage rates and average quantities consumed for all three foods that were not detected by HCES. Guided by the Food Fortification Formulator, we found that these differences did not result in differences in recommended fortification levels. Differences were found, however, in estimated impacts: although both surveys found that oil would be effective in reducing the prevalence of inadequate vitamin A intake among both subpopulations, only OWFR also found that sugar and wheat flour fortification would significantly reduce inadequate vitamin A intake among children.
Conclusions: Despite the less precise measure of food consumption from HCES, the two surveys provide similar guidance for designing a fortification program. The external validity of these findings is limited. With relatively minor modifications, the precision of HCES in dietary assessment and the use ofHCES in fortification programming could be strengthened.