Estimating the cost and cost‐effectiveness of adding zinc to, and improving the performance of, Burkina Faso's mandatory wheat flour fortification programme

Abstract Zinc is an essential micronutrient that promotes normal growth, development and immune function. In the context of persistent dietary zinc inadequacies, large‐scale food fortification can help fill the gap between intake and requirements. Burkina Faso mandates wheat flour fortification with iron and folic acid. We used activity‐based cost modelling to estimate the cost of adding zinc to the country's wheat flour fortification standard assuming (1) no change in compliance with the national standard, and (2) a substantial improvement in compliance. We used household food consumption data to model effective coverage, that is, the number of women of reproductive age (WRA) predicted to achieve adequate zinc density (zinc intake/1000 kcal) with the addition of fortification to diets. Without interventions, the prevalence of inadequate dietary zinc density was ~35.5%. With no change in compliance, the annual average incremental cost of adding zinc to fortified wheat flour was $10,347, which would effectively cover <1% of WRA at an incremental cost of ~$0.54/WRA effectively covered. Improving compliance added ~$300,000/year to the cost of the fortification programme without zinc; including zinc added another ~$78,000/year but only reduced inadequate intake among WRA by 3.6% at an incremental cost of ~$0.45/WRA effectively covered. Although the incremental cost of adding zinc to wheat flour is low ($0.01/wheat flour consumer/year), given low levels of wheat flour consumption, zinc fortification of wheat flour alone contributes marginally to, but will not fully close, the dietary zinc gap. Future research should explore potential contributions of zinc to a broader set of delivery vehicles.


| INTRODUCTION
Micronutrient deficiencies, or "hidden hunger", remain a significant global health concern, particularly among subpopulations with relatively high micronutrient requirements, including children under 5 years of age and women of reproductive age (WRA) (Stevens et al., 2022;Victora et al., 2021). Adequate zinc intake is necessary for normal growth, development, and immune function, and zinc deficiency is associated with increased incidence of diarrhoea and respiratory infections among children under 5 years of age (King et al., 2016). The global prevalence of risk for inadequate zinc intake has been estimated at 17.3% (Wessells & Brown, 2012), and among 25 low-and middle-income countries (LMICs) with plasma/serum zinc concentration data, the prevalence of zinc deficiency is >20% for at least one physiological group in all but two countries (Gupta et al., 2020). In Burkina Faso, an estimated 35.2% of the population is at risk of inadequate zinc intake (K. R. Wessells, personal communication, December 1, 2022), and a study in rural Burkina Faso found low serum zinc concentrations among 39.4% of women and 63.7% of children 36-59 months of age (Martin-Prevel et al., 2016).
Alongside interventions to promote dietary diversification and modification, biofortification, and supplementation, large-scale food fortification (LSFF) is one of the primary intervention strategies identified by the World Health Organization (WHO) and the United Nations Food and Agriculture Organization (FAO) to address micronutrient malnutrition (Osendarp et al., 2018). LSFF, which is the addition of vitamins and/or minerals to staple foods or condiments during industrial processing (Olson et al., 2021;World Health Organization & Food and Agricultural Organization, 2006), does not require changes in dietary patterns to be effective and is often touted as relatively low-cost. As such, LSFF is widely recognized as a potentially cost-effective intervention to improve micronutrient intake and reduce the prevalence of deficiency (Horton et al., 2008;Keats et al., 2021;Osendarp et al., 2018;World Health Organization & Food and Agricultural Organization, 2006).
In the past two decades, LMICs have increasingly utilized the fortification of staple foods, including wheat flour, refined oils and rice, as a strategy to improve the micronutrient adequacy of diets (Osendarp et al., 2018). A recent systematic review and meta-analysis found that zinc fortification with or without other micronutrients reduced the prevalence of zinc deficiency and may provide health and functional health benefits, such as lowering the incidence of child diarrhoea (Tsang et al., 2021). Based on this evidence, the recently  Grant et al., 2018). However, despite the estimated high burden of zinc deficiency in the country, zinc was not included in the fortification standard because, at the time the regional standard was developed, there was considered to be a lack of evidence of the efficacy and effectiveness of zinc fortification (Tarini et al., 2021). As such, and because wheat flour is already required to be fortified with iron and folic acid, revising the wheat flour fortification standard in Burkina Faso to also include zinc may be a relatively straightforward and low-cost strategy to increase dietary zinc intakes.
The success of a LSFF programme in terms of its ability to costeffectively improve micronutrient intake and status depends on appropriate programme design (or redesign) and effective implementation and management Mkambula et al., 2020). This includes determining whether industrially processed forms of the candidate food vehicle are regularly consumed in sufficient quantities, such that fortification would help close the gap between dietary intake and requirements by population subgroups at greatest risk for deficiency. It also includes understanding the range of costs associated with implementing and managing an LSFF programme as well as planning for how different components of the cost will be paid (e.g., paid for or subsidized by government, borne by industry, passed on to consumers, etc.). Including cost as fundamental input into the overall evidence base for decision-making can help

Key messages
• Large-scale food fortification can be an effective strategy for reducing zinc deficiency.
• Dietary zinc inadequacy is a public health problem in Burkina Faso. We modelled the potential effectiveness and cost-effectiveness of fortifying wheat flour with zinc to help address this issue.
• Although the incremental cost of adding zinc to wheat flour would be low ($0.01/wheat flour consumer/year), given the severity of inadequacy and low levels of wheat flour consumption, zinc fortification of wheat flour alone contributes marginally to, but will not fully close, the dietary zinc gap.
• A broader set of delivery vehicles is needed to effectively address zinc inadequacy in Burkina Faso. prevent situations in which an intervention with great potential to help address micronutrient deficiencies is either never implemented or underperforms due to high and/or unexpected costs. Given the well-documented challenges of achieving the full potential of LSFF due, in part, to weak monitoring and enforcement systems (Luthringer et al., 2015), understanding the investments that must be made to strengthen and sustain these systems is also a critical component of this evidence base (Heidkamp et al., 2021).
The Micronutrient Intervention Modelling (MINIMOD) project has developed a set of modelling tools to provide policymakers with estimates of the nutritional benefits, costs, and cost-effectiveness of alternative micronutrient intervention programmes (Adams, Luo, et al., 2022;Engle-Stone, Nankap, et al., 2015;Vosti et al., 2015).
Our objectives were to apply the MINIMOD modelling framework and its underlying data to estimate the incremental nutritional benefits, costs and cost-effectiveness of adding zinc to Burkina Faso's existing wheat flour fortification programme. As described below, Burkina Faso's wheat flour fortification programme is underperforming relative to the goals set out in the country's national wheat flour fortification standard. Given this, we also estimated the costs associated with improving compliance with the national standard and the incremental cost and cost-effectiveness of adding zinc to the wheat flour fortification standard under this improved compliance scenario.

| Cost modelling scenarios
We estimated the incremental cost of adding zinc (in the form of zinc oxide) to Burkina Faso's wheat flour fortification programme. More specifically, we modelled the costs of moving from the status quo of wheat flour fortified with only iron and folic acid to the addition of zinc included in the country's wheat flour fortification standard. Our analysis considered two levels of programme compliance vis-à-vis the standard (current compliance and improved compliance); in all, we modelled four distinct wheat flour fortification scenarios (Table 1)  compliance of the current programme (i.e., fortification of wheat flour with iron and folic acid) along two dimensions: (1) the percentage of fortifiable wheat flour that is fortified to any extent, and (2) among fortified wheat flour, the average fortification level relative to the national standard. Finally, the expanded programme with an improved compliance scenario combined the costs of the expanded current programme (i.e., the hypothetical addition of zinc to the national standard) with the costs associated with improving compliance.

| Costing approach
We took an activity-based (or microcosting) approach to estimating the economic cost of each of the scenarios outlined above (World Health Organization, 2003), and we defined costs from a societal perspective, that is, including the costs paid by all stakeholders

| Cost data sources
We tapped multiple sources of data to inform the underlying cost model assumptions, as well as unit cost and quantity parameters values. Our

| Effectiveness modelling
Our metric of effectiveness was effective coverage, defined as the number or percentage of a target population with inadequate micronutrient intake from dietary intake without any intervention to increase micronutrient intakes who achieve adequate intake after the introduction of a micronutrient intervention programme (Adams, Luo, et al., 2022). We used these food consumption data to estimate household daily apparent consumption of each food item, where apparent emphasizes assumptions inherent in using household-level data, including that all food was consumed by the household without food waste or loss and was distributed within the household in proportion to each household member's age-and sex-specific energy requirements. We estimated daily apparent food consumption by adjusting the total quantity of food apparently consumed by the household to reflect the edible portion and yield factor from cooking, where appropriate, and dividing the resulting quantity by 7 days of recall.
We identified extreme outliers in reported quantities of food apparently consumed as the food-and region-specific 95th percentile of apparent household consumption per adult male equivalentand replaced outliers above the 95th percentile with the value at the 95th percentile .
To estimate daily apparent household nutrient intake, we then matched each food item to a food composition proportion of WRA current pregnant and not pregnant according to the most recent Demographic and Health Survey data (ICF, 2012).
Because nutrient density is a metric of dietary quality, adequacy was based on the assumption that WRA were meeting their age-and sex-specific energy requirements (Vossenaar, Solomons, Muslimatun, et al., 2019).
We modelled the contribution of wheat flour fortified with zinc by multiplying daily apparent household consumption of wheat flour (including reported consumption of wheat flour and wheat flour equivalents from products containing wheat flour, such as bread and biscuits) by the average hypothetical zinc fortification levels under the expanded current programme and expanded current programme with improved compliance scenarios (Table 1) and recalculating the nutrient density of the household diet and prevalence of inadequacy.
Finally, effective coverage was calculated as the percentage of WRA moving from inadequate zinc density from dietary intake without any intervention to increase zinc intakes to adequate zinc density with the introduction of zinc to fortified wheat flour.

| Cost-effectiveness
We

| Sensitivity analyses
We conducted sensitivity analyses to assess the influence of several key parameter values on our cost and cost-effectiveness results. In particular, we reduced and increased the price of zinc oxide by 20%, from $7.00 per kg in the primary analysis to $5.60 and $8.40. We also modelled a 50% increase in the assumed zinc fortification level, from 95 to 142.5 mg/kg; this higher fortification level is similar to countries with wheat fortification standards that include zinc at levels above the WHO recommendations relative to wheat flour consumption (e.g., Ethiopia). We did not assess the sensitivity of our estimates to potential errors in the estimated nutrient composition of diets.

| The wheat flour industry in Burkina Faso
Burkina Faso does not grow any wheat in-country, but two industrial-

| The cost of wheat flour fortification
The estimated annual average cost of the current wheat flour fortification programme (which is functioning at a low level of compliance) in Burkina Faso was $26,601 (Table 2, Table S1).
Improving compliance of the existing programme (including only iron and folic acid), from the current state of 61.5% of wheat flour being fortified at, on average, 12% of the national standard to 90% of wheat flour being fortified to 100% of the national standard, would cost an additional~$3 million over 10 years, or an annual average increase of $301,411 in the total cost of the programme (

| The potential effectiveness and cost-effectiveness of zinc-fortified wheat flour
Based on the most recent household survey data, 44% of households in Burkina Faso report consuming any wheat flour (or product containing wheat flour) in the 7 days preceding the survey, and among those, average apparent consumption of wheat flour among consumers was 21 g/WRA/day (

| Sensitivity analysis
If the price of zinc oxide were 20% higher or 20% lower than our assumed cost of $7.00/kg, the incremental cost per WRA effectively T A B L E 2 Estimated cost of the current wheat flour fortification programme in Burkina Faso and the cost of expanding the programme to include zinc.  ). e Note that the cost per kg of premix is lower because, to maintain the correct ratio of micronutrient fortificant to excipient, the addition rate (the rate at which premix gets added to unfortified wheat flour) jumps from 250 g of premix per MT of wheat flour without zinc to 400 g of premix per MT of wheat flour with zinc, so the concentration of micronutrient fortificants is lower with zinc (and hence the price of the premix per kg is lower) but more premix is added per MT of wheat flour. f Average annual quantity of fortified wheat flour is 50,338 metric tons.
T A B L E 3 Estimated cost of improving the performance of the current wheat flour fortification programme in Burkina Faso and the cost of expanding the improved programme to include zinc. Under the current programme scenario, industry compliance modelled as 61.5% of wheat flour fortified with iron and folic acid at 12% of the national standard. National standard: 60 mg/kg iron and 2.5 mg/kg folic acid.
b Under the improved compliance scenario, industry compliance modelled as 90% of wheat flour fortified with iron and folic acid at 100% of the national standard. National standard: 60 mg/kg iron and 2.5 mg/ kg folic acid.
c Under the expanded programme with improved compliance scenario, industry compliance modelled as 90% of wheat flour fortified with iron, folic acid, and zinc at 100% of the hypothetical national standard. Hypothetical national standard: 60 mg/kg iron, 2.5 mg/kg folic acid, 95 mg/kg zinc. f Note that the cost per kg of premix is lower because, to maintain the correct ratio of micronutrient fortificant to excipient, the addition rate (the rate at which premix gets added to unfortified wheat flour) jumps from 250 g of premix per MT of wheat flour without zinc to 400 g of premix per MT of wheat flour with zinc, so the concentration of micronutrient fortificants is lower with zinc (and hence the price of the premix per kg is lower) but more premix is added per MT of wheat flour.
g Average annual quantity of fortified wheat flour is 69,565 metric tons.
T A B L E 4 Effectiveness of expanding Burkina Faso's wheat flour fortification programme to include zinc.   zinc was $4.40 (Fiedler et al., 2013), although the cost of maize flour fortification included zinc plus eight other generally more expensive micronutrients. In another analysis, Vosti et al. (2023)   Our study had several limitations. One important limitation was that our measures of effectiveness were based on household food consumption data, which are less accurate than individual dietary recall data and require assumptions about the intrahousehold distribution of food . Although estimates of average apparent intake of energy and zinc among WRA were plausible (2135 kcal/day and 11.4 mg zinc/day), it is still likely that the accuracy of our estimates of zinc inadequacy and the impact of fortification were affected, to some degree, by measurement error in the underlying data. We assessed zinc adequacy based on the zinc density of the household diet to try and account for some of the measurement error inherent in household food consumption data, but these energy-adjusted measures do not help correct systematic under or over-reporting of foods that may vary in nutrient contents from the typical household diet, such as foods consumed away from home or foods not shared among household members. Related, because food consumed away from home was not adequately captured in the data, if food items containing wheat flour are commonly consumed outside the home, our estimates of the impact of wheat flour fortification may be low. Also, because breastfeeding status was not collected as part of the household survey, we were unable to account for zinc requirements during lactation, which are higher than during pregnancy. As a result, our estimates of the prevalence of zinc inadequacy among all WRA are likely underestimated, and our estimates of the impact of wheat flour fortification may be overestimated. As with most dietary modelling studies, it is also possible that the nutrient composition values we used (primarily from the 2019 West African FCT) may not be accurate for Burkina Faso. Finally, we were unable to model potential impacts on plasma/serum zinc concentrations, cases of diarrhoea averted, or other health outcomes. However, given low effective coverage of modelled programmes, their effects on these health outcomes would likely also be low.
Our study also had several strengths. First, we developed comprehensive, activity-based cost models that allowed for comparison of the absolute and relative contributions of different cost components to the total cost of wheat flour fortification. Another strength was the use of those cost models, along with estimates of effectiveness, to predict cost-effectiveness based on current compliance with the national standards, but also based on a scenario of improved compliance.