Case study based on: Fiedler, J., & Lividini, K. (2014). Managing the vitamin A program portfolio: a case study of Zambia, 2013-2042. Food Nutr Bulletin.
Zambia is a large and landlocked middle-income country located in the center of southern Africa. Today, Zambia has some of the highest levels of poverty and inequality in the world (World Bank, 2019). Undernutrition and micronutrient deficiencies remain a major public health concern and are among the top causes of death and disability (WHO, 2019). In the past decades, supplementation and fortification programs have been put in place to combat micronutrient deficiencies in Zambia (Fiedler & Lividini, 2014). However, there is a need to understand which combination of programs is most effective and cost-effective to address these deficiencies both presently and in the future. The following case study, based on work by Fiedler & Lividini (2014), illustrates the use of national and regionally- representative Household Consumption and Expenditure Surveys (HCES) data from Zambia to estimate nutrient level intakes and project the potential coverage and impact of potential food vehicles used in current and possible future vitamin A programs. The following paragraphs describe the methods and findings from Fiedler & Lividini (2014) and highlight strengths and weaknesses of using HCES data to assess household food consumption and nutrient intakes.
Fiedler & Lividini (2014) quantified apparent food consumption, nutrient intake, and adequacy levels from HCES data in order to identify likely consumers of fortified flour, sugar, oil, and maize meal for potential coverage of vitamin A supplementation through the Child Health Weeks program. This information was modeled to project the potential effectiveness and cost-effectiveness of different combinations of interventions over a 30-year period.
To undertake the assessment, Fiedler & Lividini (2014) relied on the Zambia 2006 Living Conditions Monitoring Survey (LCMS), a type of HCES, to obtain data on the quantity of food purchased and consumed and the mode of acquisition (purchased, home-produced or received as an in-kind transfer). Because the survey measured food acquisition, rather than intake, the authors assumed that all food that was acquired during the reference period was consumed and referred to this variable as “apparent consumption.” The food purchase module asked questions for two recall periods: the previous two weeks and the previous four weeks (depending on the frequency of purchase). For analysis, food expenditures for all food items were converted to a uniform reference period (last 30 days) and then aggregated. The authors then used the Zambian National Food and Nutrition Commission’s Food Composition Table (FCT) (National Food and Nutrition Commission, 2009) to derive nutrient information for each food item in order to estimate household average dietary energy consumption and adequacy of micronutrient intake of vitamin A, iron and zinc. To estimate intake and inadequacy for individual household members, the authors assumed that all food was consumed within the household in an amount equivalent to each member’s relative energy requirement. The relative energy requirement was derived by comparing the relative share of energy required for each household member of a given age, sex, and physiological status compared to an adult male (FAO, 2002). Individuals’ micronutrient intake levels derived through this “AME method” were then compared with age- and sex-specific estimated average requirements (EARs) to classify the adequacy of each individual’s intake. The EARs are “the average daily intake level for a specific nutrient estimated to meet the requirements of half of the healthy individuals of a specific age, sex, and life-stage” (IOM, 2006). For vitamin A and zinc, the authors used the “cut-point” method for estimating apparent adequacy and for iron, they used the “full probability method” because the physiological requirements for this micronutrient is often not normally distributed. The authors estimated the potential coverage of fortifiable flour, sugar, oil, and maize meal by gauging whether, and how much, households consumed these industrially fortifiable food vehicles. The likely vitamin A intake from fortified foods was estimated based on regulatory data pertaining to required fortification levels for a given food vehicle, adjusted for pre-consumption degradation, and the quantity of this food apparently consumed by households. Vitamin A supplementation coverage was estimated from the HCES survey, which captured vitamin A supplements received by children in the household over the preceding 6 months. They used HCES agricultural production data from the HCES along with data from a supplementary food security survey to simulate the adoption, production, disposition, consumption and dissemination of biofortified maize. Using primary cost data, the authors examined the cost of each of the three program interventions and calculated cost-effectiveness by combining cost data with the number of disability-adjusted-life years (DALYS) likely saved through the intervention. The combination of this information yielded a picture of optimal coverage, cost, health impacts, cost-effectiveness, and affordability of different intervention portfolios for the year of the study (2013). These parameters were then integrated into IFPRI’s IMPACT model to project how changing production and consumption patterns through 2042 would affect the results.
Their analysis found a high prevalence of inadequate apparent vitamin A intake in the Zambian population. In rural areas, 65% of people were affected while in urban areas the prevalence was about half (35%). In some rural provinces, prevalence of inadequacy affected 99% of the population. Preschool children (12-59 months) and women (15-49 years) from the Eastern and Southern provinces were the most affected, with apparent intakes reaching only ~15% of the relevant EARs. The portfolio analysis showed a reduction in the prevalence of apparent inadequate intake over time for each package of interventions analyzed. Fiedler & Lividini (2014) recommended that the country consider introducing fortified vegetable oil to the existing vitamin A supplementation and sugar fortification program, followed by phasing in fortification of wheat flour and maize meal. While the introduction of maize meal fortification was found to have the lowest incremental cost-per-DALY saved, the authors underscore the fact that all the combinations of interventions examined were highly cost-effective and would merit country investment in the full portfolio.
This case study is an example of how HCES data can be used to estimate the percentage of the population at apparent risk of inadequate intake of vitamin A and simulate the coverage, effectiveness and cost-effectiveness of packages of interventions to help address micronutrient deficiencies in low-middle income countries in the long term. HCES data have been shown to be a valuable source of information for diet and nutrition-related analyses (Russel et al., 2018; Fiedler et al., 2016; Smith et al., 2014; Fiedler et al., 2012) for many reasons, including: (1) HCES are typically conducted on a large nationally representative sample; (2) the data are collected fairly frequently (e.g. every 3-5 years); and (3) these multipurpose surveys typically include a wide range of other variables (e.g. socioeconomic status, health, education, production), allowing for an examination of diets in relation to agriculture, schooling, and other outcomes of interest. However, HCES data also have weaknesses that make the data less optimal for certain food security and nutrition analyses. For instance, the Zambia HCES only measures “apparent consumption,” which is based on acquisition data, not actual consumption. Food lists of many HCES are often not disaggregated enough to make an appropriate match to a Food Composition Table. The Zambia 2006 HCES collected data on only 39 items (World Bank, 2012), which was adequate for the purposes of this study but would not accurately reflect a complete diet. Recall periods across many HCES surveys extend up to 365 days, raising concern about recall bias, though in this instance the recalls were more reasonable at two weeks and 30 days. Many HCES surveys, like the one in Zambia, lack information on food consumed away from home, meaning that the estimates of apparent consumption are likely to be biased downward. Even more significantly, HCES generate household level data, without information on consumption patterns of groups of individuals, such as women or young children. The authors of the Zambia study assumed that food and micronutrients in the household were allocated to individuals according to their relative physiological energy needs. Though this assumption has been shown to hold true in some settings for certain age groups, the assumption is context and nutrient-specific and is least accurate for the youngest children (Fiedler et al., 2015; Coates et al., 2017). Intra-household allocation patterns are unknown with HCES data, so extrapolating to individuals is risky without triangulating with individual-level dietary survey data. Data from 24-hour Dietary Recalls and Food Frequency Questionnaires offer more accurate estimates of individual food and nutrient intake, and should be used when possible to assess dietary patterns trends. As Zambia did not have national 24-hour dietary data at the time of the study, the authors made creative use of the HCES, recognizing its limitations and acknowledging that the results were estimates of “apparent intake” only.
The authors took advantage of available national and regional level data to model a useful range of criteria of interest to policy-makers in considering how best to tackle widespread vitamin A deficiency in Zambia. These results are extremely useful for near-term as well as long-range nutrition planning.
- Coates, J., Rogers, B., Blau, A., Lauer, J., & Roba, A. (2017). Filling a dietary data gap? Validation of the adult male equivalent method of estimating individual nutrient intakes from household-level data in Ethiopia and Bangladesh. Food Policy, 72:27-42.
- Food and Agriculture Organization. (2003). Food energy–methods of analysis and conversion factors. Report of a technical workshop. Rome, 3-6 December 2002. Food and Nutrition Paper 77. Rome: FAO, 2003.
- Fiedler, J., & Mwangi, D.M. (2016). Using Household Consumption and Expenditure Surveys to Make Inferences about Food Consumption, Nutrient Intakes and Nutrition Status: How Important Is It to Adjust for Meal Partakers? IFPRI Discussion Paper No. 1571.
- Fiedler, J., Lividini, K., & Bermudez, O. (2015). Estimating the impact of vitamin A-fortified vegetable oil in Bangladesh in the absence of dietary assessment data. Public Health Nutrition, 18(3):414-420.
- Fiedler, J., & Lividini, K. (2014). Managing the vitamin A program portfolio: a case study of Zambia, 2013-2042. Food Nutr Bull, 35(1):105-125.
- Fiedler, J., Lividini, K., Bemudez, O., & Smith, M. (2012). Household Consumption and Expenditure Surveys (HCES): A primer for food and nutrition analysts in low- and middle-income countries. Food and Nutrition Bulletin, 33(3) suppl, S170-S184.
- International Food Policy Research Institute. (n.d.). IFPRI’s Food Security Portal: IMPACT Model. Retrieved January 29, 2020, from https://www.ifpri.org/content/ifpris-food-security-portal-impact-model
- Otten, J.J., Pitzi Hellwig, J., & Meyers, L.D, eds. Institute of Medicine of the Unites States National Academy of Sciences. (2006). DRI Reference Intakes. The Essential Guide to Nutrient Requirements. Washington, D.C: The National Academies Press.
- National Food and Nutrition Commission (NFNC) (2009) Ministry of Health, Government of the Republic of Zambia. Food Composition Table, 4rd ed. Lusaka: NFNC.
- Rosegrant, M.W., Ringer, C., Msangi, S., Cline, S.A., & Sulser, T.B. (2008). International Model for Policy Analysis of Agricultural Commodities and Trade (IMPAACT-WATER): Model description. Washington, DC: International Food Policy Research Institute.
- Russel, J., Lechner, A., Hanich, W., & Delisle, A. (2018). Assessing food security using household consumption expenditure surveys (HCES): a scoping literature review. Public Health Nutrition, 21(12), 2200-2210. https://doi.org/10.1017/S136898001800068X
- Smith, L.C., Dupriez, O., & Troubat, N. (2014). Assessment of the Reliability and Relevance of the Food Data Collected in National Household Consumption and Expenditure Surveys. IHSN Working Paper N0. 008.
- World Bank. (2012). Zambia Poverty Assessment. Stagnant Poverty and Inequality in a Natural Resource-Based Economy. Report No. 81001 – ZM. Poverty Reduction and Economic Management, Africa Region.
- World Bank. (2019). The World Bank in Zambia. Retrieved from https://www.worldbank.org/en/country/zambia/overview
- World Health Organization. (2019). Micronutrient deficiencies: Vitamin A deficiency. Retrieved from https://www.who.int/nutrition/topics/vad/en/