The prospective base-case intervention scenario used in this economic evaluation was conceptualised to be pragmatic and implementable within the South African healthcare system. The intervention design was informed by consultations with key stakeholders, including the NDoH. The design was also based on the oral iron preparations and current packaging available on the NDoH Master Health Product List. We envisaged delivery as occurring concurrently with other routine clinic visits within the South African PHC system (e.g. for childhood immunisations). We also considered the current implementation of routine preventive vitamin A supplementation in South Africa.

We used the WHO source guideline recommendation to inform dose, frequency and duration, that is infants and young children aged 6–23 months receive 10 mg – 12.5 mg elemental iron daily for 3 consecutive months in a year.⁶ This daily dosage is equivalent to between 83.3 mg and 104.2 mg of ferrous gluconate.⁶ The NDoH Master Product List provides for 100 mL ferrous gluconate syrup, with every 5 mL of syrup providing 350 mg of ferrous gluconate.²⁹

The intervention scenario used for this evaluation was that each child would receive one 100 mL bottle of ferrous gluconate syrup per year, for 2 years. To distribute the iron supplement, the caregiver would collect a bottle of syrup from the clinic during the child’s routine 6- and 18-month immunisation visits. Each year, the child would receive 1.5 mL of syrup daily for 5 days of the week for 13 weeks.

Each dose of 1.5 mL of syrup contains 105 mg of ferrous gluconate and 12.6 mg elemental iron. The deviation from the WHO’s recommended dosing regimen from 7 to 5 days per week was made for several well-considered reasons related to wastage and safety. Based on the WHO regimen, a child requires 137 mL of ferrous gluconate syrup over 3 months each year. Since the current bottle size of ferrous gluconate available in South Africa is 100 mL, caregivers must receive two 100 mL bottles per dosing cycle to adhere to the WHO regimen. This results in 63 mL of syrup remaining at the end of the treatment period. This additional wasted syrup increases the costs of the programme over the 2 years by, in South African Rand, R38 408 803.00 ($2 081 778.00) (the cost of an additional bottle and syringe per child for the total population for 2024 and 2025). It would also be potentially problematic given the toxic nature of excessive iron, if consumed incorrectly. This has been shown with maternal iron tablets, where researchers reported that accidental iron poisoning in children is associated with significant mortality.³⁰ The dosing regimen was further substantiated by findings from the update of an effectiveness systematic review of randomised controlled trials (RCTs) showing that frequent dosages (5–7 times per week) of iron supplement were effective in reducing iron-deficiency anaemia in children aged 6 months – 23 months.³¹ Furthermore, if the caregivers were to incorrectly give 1.5 mL of syrup for 7 days instead of 5 days, there would be no harm to the child, and the child would still benefit from the iron supplementation.

Most of the South African population (approximately 85%) accesses healthcare through the government-funded public healthcare sector.^32,33 The remainder of the population accesses private healthcare funded by their monthly contributions to medical schemes. For simplicity in implementation and to remove the need for additional clinic visits, caregivers would collect the oral iron supplementation when the children receive their immunisations. However, not all children receive all their immunisations.³⁴ The costing for the base-case intervention was based on the entire South African population of children within the age category. This is regardless of whether they received their immunisations and whether they accessed care in the public or private sector, because South Africa provides free PHC for mothers and children under the age of 6 years. Using the entire relevant population also provides a worst-case scenario from a cost perspective, as it assumes the maximum expense for the South African health system in planning for all eligible intervention recipients. The population size for South African children aged 6–23 months was based on Statistics South Africa (Stats SA) projections of the population size for children aged less than 2 years for 2024 and 2025 (Statistics South Africa, personal communication).

The costing was based on the provider perspective, whereby the costs were considered from the viewpoint of the public health system, in this case, the South African Government. The costs were based on relevant inputs that would be required to integrate preventive oral iron supplementation into routine PHC services for infants and young children in South Africa, in line with the pragmatic scenario previously described. This approach allows for significant cost savings as staff costs usually drive total health care costs. However, the intervention scenario relies on existing staff capacity to distribute syrup to the caregivers during scheduled clinic appointments. Discounting was not needed as the time horizon for both costs and effects was a year; annual costs for 2025 were adjusted for an estimated 6% inflation, as this value is the upper bound of the country’s inflation target band. All costs were estimated in South African Rand (R [ZAR]) and converted to United States Dollar ($ [USD]) using the average exchange rate for 2023 of R18.45 to USD. Table 1 provides the items necessary for the intervention and the source of the costing information.

Costing of the intervention did not include costs relevant to adverse outcomes. The effectiveness systematic review (Brand et al. 2026, forthcoming) found only a small number of studies reported on infection outcomes (diarrhoea and respiratory illness), with pooled effect estimates showing oral iron supplementation having little to no effect on these outcomes. Furthermore, none of the studies included in the systematic review reported the need for health care because of infections. The cost analysis also did not include any savings arising from averted anaemia-related clinic visits because of the intervention.

We classified anaemia into the WHO-defined severity levels based on haemoglobin levels, that is non-anaemic (≥ 11.0 g/dL), mild (10.0 g/dL to 10.9 g/dL), moderate (7.0 g/dL to 9.9 g/dL), or severe (< 7 g/dL) anaemia.²⁹ The percentage of the population within the various severity levels of anaemia at baseline (pre-intervention) is based on estimates for children under the age of 24 months for Southern Africa. The estimates are from a study combining data from Southern African countries. The South Africa Demographic and Health Survey has anaemia rates or South Africa only, but the sample size of children in each group is small, which may influence the analysis. The additional analysis for South African children aged < 2 years is provided in the sensitivity analysis.⁹ Using these values, we calculated a population mean and standard deviation (s.d.) for a normal distribution, using z-scores that corresponded with the proportional area under the curve.

To quantify the size of the effect, we used a mean difference in circulating haemoglobin levels of 0.55 g/dL (95% confidence interval [CI] 0.42 to 0.68) with iron supplementation compared to no supplementation over a single treatment period. This effect estimate was obtained from the related systematic review of 41 RCTs involving 10 298 infants and children, produced for the GELA project (Brand et al. 2026, forthcoming). The systematic review took the approach of assessing the effect of assignment to intervention, rather than assessing the effect of adherence to intervention. We applied this average effect estimate to the haemoglobin levels, defining anaemia categories to determine the changes in haemoglobin values with the intervention and calculated changes in the proportion of the population in each of the anaemia severity categories using the population mean and s.d. (see Figure 1). The analysis included children with severe anaemia to cover those who had undiagnosed severe anaemia and would benefit from the intervention.

The health outcomes in the CEA were based on three main assumptions. Firstly, we assumed haemoglobin levels are normally distributed. Secondly, we assumed that effects on haemoglobin levels are non-cumulative across the two years of the intervention, that is, haemoglobin levels are anticipated to return to the baseline over the course of the nine months without oral iron supplementation. This assumption was necessary as there is no literature showing the cumulative effects of intermittent oral iron supplementation. In the absence of such evidence, assuming non-cumulative effects for intermittent supplementation also provides a worst-case scenario from an effectiveness perspective, as it assumes the minimal effectiveness benefit. Any potential bias from this assumption would result in the intervention being less cost-effective. Thirdly, we assumed the effect size is equivalent across the categories of anaemia severity.

For the CEA, we drew on the approach from Akpan and colleagues’ CEA of iron supplementation and iron-containing micronutrient powders in Bangladesh.²⁸ We estimated disability-adjusted life years (DALYs) per year for no oral iron supplementation and for iron supplementation. One DALY represents the equivalent of 1 year of full healthy life that is lost because of a specific health condition. Disability-adjusted life years for a health condition are calculated as the sum of years lost to disability (YLD) and years of life lost (YLL) because of the causes of the health condition. Because we do not anticipate clinically meaningful effects on mortality from the iron supplementation only, our DALY focused on YLD only. Years lost to disability because of anaemia were calculated by multiplying the number of children in each anaemia category by the corresponding disability weights from the ‘… Global Burden of Disease (GBD) study (note that Akpan28 includes diarrhoea in the YLD calculation as there is a small chance of diarrhoea from iron supplementation. However, the effectiveness analysis updating Andersen et al. 2023 found little to no effect of oral iron supplementation on diarrhoea, so we have excluded this from our calculations). World Health Organization-defined severity levels based on haemoglobin levels were assigned a corresponding GBD disability weight.^35,36

The incremental cost-effectiveness ratio (ICER) was calculated by taking the difference in costs divided by the difference in DALYs, taking incremental differences between iron supplementation and no oral iron supplementation. The ICER was then compared to ‘benchmark values’ or ‘cost-effectiveness thresholds’ to determine if the new intervention that requires additional costs represented good value for money. An ICER per DALY averted that is lower than a threshold suggests that the intervention likely generates positive net health benefits for the population. No specific cost-effectiveness threshold has been adopted for South Africa’s National Essential Medicines List decisions, which this work informed. As such, we considered two publicly available thresholds developed for South Africa. Edoka and Stacey developed a threshold that estimates the opportunity cost of health spending in South Africa.³⁷ The threshold was estimated to be 53% of per capita gross domestic product (GDP).³⁷ Pichon-Riviere and colleagues³⁸ followed a slightly different approach by using growth in life expectancy and health expenditure to determine cost-effectiveness thresholds for 174 countries. These country-specific thresholds consider the country’s unique health needs and resources; these were developed following the withdrawal of WHO’s GDP-based rule on thresholds.^37,38 For South Africa, the authors determined a threshold as a proportion of per capita GDP (and range) to be 0.68 (0.50–1.20).³⁹ Given that Edoka and Stacey estimated the opportunity cost of healthcare spending resulting in a more conservative threshold, we deemed this as the primary threshold appropriate for this study.³⁷ However, we also compared the ICER to the lower bound (50% of per capita GDP) reported by Pichon-Riviere and colleagues to further substantiate our findings.³⁸ We calculated the thresholds using per capita GDP for 2024 of R118709.00 ($6434.00).

A BIA is usually performed in addition to a CEA. Budget impact analysis takes the unit cost of an intervention and multiplies it by the number of people affected by the intervention to provide an understanding of the total budget required to fund the intervention. The impact on the national budget was calculated annually for 2 years (2024 and 2025). The total impact for 2024 included all the once-off start-up costs, which were excluded from 2025 (and onwards). The BIA did not include potential savings to the health system that were realised beyond the 2-year intervention period.

We were uncertain about whether the ZAR would weaken compared to the USD and whether inflation would be higher than in our original model. A weakening ZAR and higher inflation would increase the cost of the syrup, syringe, and staff start-up costs, meaning that we underestimated the total cost of the intervention. To address this uncertainty, Sensitivity analysis 1 is a multiway sensitivity analysis where higher estimates of costs were used in a cost-effectiveness and BIA. For the cost of the syrup, we assumed that the ZAR weakened by 20% to the USD each year (the assumption was based on the South African Rand having a history of volatility. During 2023, the Rand weakened by nearly 16% before recovering again, therefore we use 20% to capture a potential sustained weakening of the Rand). We also assumed that the cost of the syringes is double what was initially estimated. Finally, we estimated that the salary of the person responsible for updating all the guidelines and facilitating the training was at a higher pay scale.

To accommodate uncertainty in the effects of preventive oral iron supplementation, we incorporated the lower bound 95% confidence interval on the effect estimate and combined it with the higher cost from the first sensitivity analysis in a multiway sensitivity analysis (Sensitivity analysis 2). This approach allows us to determine a worst-case scenario of higher cost for a smaller effect for the CEA, no BIA was performed for this sensitivity analysis.

We were also uncertain about the proportion of South African children aged 6–23 months falling in each category of anaemia severity at baseline (‘pre-intervention’) in South Africa. To address this uncertainty, Sensitivity analysis 3 used a one-way sensitivity analysis based on estimates of anaemia severity levels from the South Africa Demographic and Health Survey and the original estimates of effects.⁴⁰ The sample size of children in each of the severity levels is smaller than the Southern African totals we used in the main analysis, which may influence the results.

Finally, because our intervention scenario would require caregivers to collect the syrup at their child’s 6- and 18-month immunisation clinic visit, this implementation strategy would result in children who miss these immunisation visits also not receiving the syrup. The NDoH estimates that 83.9% of children up to a year old and 76.8% of children up to 18 months receive all basic vaccinations.³⁴ To account for this uncertainty, Sensitivity analysis 4 is a one-way sensitivity analysis that adjusts the total population size to include only the number of children who receive their immunisations, using the original costs and effects estimates for a CEA and BIA.

This article followed all ethical standards for research without direct contact with human or animal subjects.

The population size for South African children between 6 and 23 months of age is estimated to be 2 272 576 in 2024 and 2 263 752 in 2025 (estimated figures [not rounded], Statistics South Africa, personal communication). We estimated the total immunised population to be 1 829 296 (80%) in 2024 and 1 822 358 (81%) in 2025. These estimates are based on the Department of Health ‘Expanded programme on immunisation national coverage survey report’ indicating 83.9% of children in the survey received all basic vaccinations up to 1 year old and 76.8% were fully vaccinated up to 18 months.

Table 2 provides the breakdown of costs for the intervention between 2024 and 2025. The first section of the table provides the breakdown of once-off start-up costs, thus only reflecting in 2024. The second section provides the annual unit costs for each year. The last two rows provide the total cost per child of the intervention, both per year and over 2 years, excluding the start-up costs.

Table 3 provides the percentage of children within the various anaemia severity levels before and after iron supplementation based on the mean difference of 0.55 g/dL in haemoglobin levels obtained from the systematic review of RCTs comparing iron supplementation to no supplementation (Brand et al. 2026, forthcoming). It also provides the YLD (95% confidence intervals) for before and after the intervention. The estimated proportion of children with moderate anaemia decreases noticeably from 35.75% to 23.82%, while the estimate of severe anaemia decreases from 2.57% to 1.64% of the total population when the effect of iron supplementation on haemoglobin levels is applied.

Table 4 provides the ICER for the total population of children under the age of 24 months.

The analysis reveals that the intervention is more effective than no oral iron supplementation, as illustrated by the reduced YLD after the intervention. However, the intervention costs more than no oral iron supplementation. The incremental cost is R1077.00 ($58.40) (95% confidence interval [CI]: R741.00 ($40.16) to R1643.00 ($89.06)) per DALY averted. This is below the Edoka and Stacey’s context-specific threshold of 53% of per capita GDP or R62 916.00 ($3410.00) for 2024, even if YLDs averted are assessed at the lower bound (the per capita GDP for 2024 was R118 709.00 [$6434.00] in 2024).^37,39 These ICERs represent between 1.71% and 2.61% of this threshold. The ICER was also lower than Pichon-Riviere and colleagues’ lower range estimate for the threshold of 50% of per capita GDP;³⁸ estimated at R59 355.00 ($3217.00) (2024).

Given that the CEA showed that iron supplementation can be considered cost-effective in the South African setting, the next step is to estimate the cost of the intervention to the fiscus. The BIA shows that, in 2024, the total cost per year for iron supplementation would amount to R18 699 908.00 ($1 013 544.00), comprising once-off start-up costs (R114 781.00 [$6221.00]) as well as the cost of 100 mL bottles of ferrous gluconate syrup (R16 778 429.00 [$909 399.00]) and syringes (R1 806 698.00 [$97 924.00]). For 2025, the BIA shows a total cost per year for iron supplementation would amount to R19 823 676.00 ($1 074 454.00), comprising the cost of 100 mL bottles of ferrous gluconate syrup (R17 916 012.00 [$971 058.00]) and syringes (R1 907 664.00 [$103 396.00]). These results translate to an annual total cost for 3 months of iron supplementation per child aged less than 2 years of R8.18 ($0.44) in 2024 (excluding start-up costs), and R8.76 ($0.47) in 2025.

Table 5 provides the ICERs for the four sensitivity analyses carried out for the cost-effectiveness analyses.

While there were slight differences in the ICER for the four sensitivity analyses, all four are in line with the initial analysis indicating that iron supplementation is cost-effective when compared to the context-relevant threshold of 53% of per capita GDP or R62 916 ($3410.00), as well as Pichon-Riviere and colleagues’ lower range estimate of 50% of per capita GDP (R59 355 [$3217.00]) based on an increase in life expectancy and health expenditure in South Africa (per capita GDP was extracted for 2024 from the South African Reserve Bank Quarterly Bulletin).^37,38,39 However, assuming higher input costs and reduced effects have the greatest influence on the ICER. Table 6 provides the sensitivity analyses 1 and 4 for the BIA for 2024 and 2025.

Our findings show that a preventive oral iron supplementation intervention has additional costs compared to no oral iron supplementation. The cost per DALY averted is R1077.00 ($58.40). For interpretation, we considered two thresholds that were developed for the South African context. Both suggested that iron supplementation was cost-effective as the ICER was well below these thresholds, even when the most conservative lower bound of the threshold based on growth in life expectancy and health expenditure was used.³⁸ Our findings suggest that diverting money to iron supplementation would likely have a net economic and health benefit. This is unlike the finding of Akpan and colleagues, that iron supplements were not cost-effective in the rural Bangladeshi setting but could become so with lower delivery costs of < $5.60/child.²⁸ The delivery costs in Bangladesh were derived from literature from previous studies, and their base case used the Home Fortification Technical Advisory Group programme delivery estimate.²⁸ The South African intervention was designed without additional delivery costs as the intervention would align with established pharmaceutical distribution networks. The cost-effectiveness of iron supplementation remained robust in sensitivity analyses that tested higher input costs and South African-specific anaemia severity levels, indicating that the intervention is likely to confer stable net economic and health benefit under a range of conditions. However, sensitivity analyses did indicate that input costs, including distribution costs, have the greatest influence on the ICER; therefore, the delivery platform differences between Bangladesh and South Africa may account for some of the observed differences in cost-effectiveness.

The anticipated annual total cost of 3 months of iron supplementation per child aged less than two years (excluding start-up costs) is estimated to be R8.18 ($0.44) in 2024. Our BIA estimated that, in the first year, iron supplementation will cost R18 699 908.00 ($1 013 545.00) for the entire population of children aged less than two years in 2024 (including start-up costs), representing 0.007% of the total R271.9 billion ($14.74 billion) budget allocated to health for 2024.⁴¹ For 2025, the intervention will cost R19 823 676.00 ($1 074 453.00).

We designed and costed a pragmatic scenario for the intervention in South Africa that we deemed implementable within the current primary healthcare system. The model comprehensively incorporated estimated health system costs related to iron supplementation procurement and implementation. Furthermore, the model drew on the effects of iron supplementation compared to no supplementation ascertained from an updated systematic review of RCTs (Brand et al. 2026, forthcoming) that was conducted specifically for GELA.

A limitation of this study is that the benefits from iron supplementation beyond the year of the intervention were not included. We assumed that the benefits from iron supplementation were not cumulative from 1 year to the next, which is a conservative assumption that would potentially bias results in the direction of underestimating cost-effectiveness. In the Benefits and Risks of Iron Interventions in Children (BRISC) trial, Akpan found that children who had received iron supplementation had at least a 20% reduction in anaemia presence at the end of the study period.²⁸ However, it was unclear how long these benefits would last, making it difficult to estimate DALYs averted over a longer period.²⁸ Furthermore, other longer-term benefits, such as cognitive improvements for school age children who did not have anaemia when they were younger, were not calculated. Incorporating these benefits would reduce the ICER, making the intervention more cost-effective, although it becomes increasingly more difficult to attribute outcomes to the intervention as these become more temporally distal. Therefore, the omission of these improvements also represents a conservative approach that would potentially underestimate cost-effectiveness.

Within the provider perspective taken in this study, we did not factor in potential savings to the health system that were realised through reduced healthcare utilisation from fewer anaemia-related clinic visits, both during the year of the study and beyond. These savings arise from fewer anaemic children with compromised immune functions visiting clinics and the financial benefits from fewer children with cognitive deficits by the time they attend school. Additionally, having taken a societal perspective may have highlighted that additional benefits to caregivers would have lower costs and productivity gains from fewer anaemia-related clinic visits. Caregivers would also not experience increased costs (including transport and loss of earnings) from collecting iron supplementation, as they would receive the supplementation during their child’s regular clinic visit. The only additional burden for the caregiver would be that they would need to ensure that they administer supplementation correctly.

The study does not include the costs of the current intervention, as administering iron supplementation is a new intervention. The BIA was also not compared to an allocated budget, as there is no specific maternal and child health budget in South Africa.

Our findings may support South African decision makers tasked with the challenge of allocating scarce health funds across many competing priorities in the public sector. While other factors would need to be considered in planning, implementing and managing a preventive oral iron supplementation programme, such as those highlighted in the related qualitative evidence synthesis,²⁵ these results provide a consequential signal of potential net economic and health benefits for a very small proportion of the total national health budget.

Furthermore, our findings offer insights into implementation considerations and the value for money of preventive oral iron supplementation in a developing country setting. This is noteworthy since the cost-effectiveness of preventive interventions for iron deficiency in developing countries have not been well-documented to date.