Malaria Consortium — Co-Delivery of Vitamin A Supplementation and Seasonal Malaria Chemoprevention in Nigeria (October 2023)

Note: This page summarizes the rationale behind a GiveWell-recommended grant to Malaria Consortium. Malaria Consortium staff reviewed this page prior to publication.

Published: January 2024

1. Summary

1.1 Background

Vitamin A deficiency (VAD) is a common condition in low-income countries that can lead to blindness, increased susceptibility to infection, and death.1 The World Health Organization (WHO) recommends that 6 - 59 month old children in areas of high VAD receive vitamin A supplementation (VAS) two to three times per year to reduce child morbidity and mortality.2

VAS is one of GiveWell’s top-recommended programs, and we have previously funded two other organizations (Helen Keller International and Nutrition International) to support VAS campaigns.3 Our full research report for VAS is available here.

1.2 What we think this grant will do

This grant will fund Malaria Consortium to deliver VAS once per year in two states in Nigeria (Bauchi and Niger states) in 2024 (in both states) and 2025 (in Niger state only). Malaria Consortium already supports another child health program, seasonal malaria chemoprevention (SMC), via door-to-door campaigns in these states. This grant will fund the additional costs of delivering VAS as part of these campaigns (e.g., additional training for distributors, and the extra costs of distributors reaching fewer children per day because they need to deliver an extra service).4

We think this grant will increase the number of children receiving VAS, and in turn avert a higher number of child deaths.

1.3 Why we made this grant

Our best guess is that this grant will be 17x (Niger state) to 54x (Bauchi state) as cost-effective as unconditional cash transfers (GiveWell’s benchmark for comparing different programs). At the time of writing this page, GiveWell’s funding bar is to fund grants that we estimate to be ~10x or more as cost-effective as cash transfers.

In simple terms, we think this grant will be cost-effective because:

• We think it will increase the number of children who receive VAS. In Nigeria, children can access VAS during routine healthcare appointments and biannual campaigns called Maternal, Newborn and Child Health Weeks (MNCHW), but we think (based on evidence from household surveys) that only ~35% to ~40% of children access VAS through these routes in Bauchi and Niger states. Based on Malaria Consortium’s previous pilots and surveys of its SMC program, we estimate that Malaria Consortium will reach 67% of children with this grant (albeit only once a year, while children will continue to have access to VAS through the other routes for the second recommended annual dose). (More)
• Child mortality is very high in northern Nigeria, where these states are located. We estimate the annual mortality rate for 6 - 59 month children is roughly 2% in Bauchi state and 1% in Niger state. (More)
• VAS probably reduces child mortality by a modest but meaningful amount. We estimate that receiving VAS will reduce a child’s mortality risk by ~3 - 5% in these states. This is based on a meta-analysis which finds that VAS reduces child mortality by 24%. We adjust this estimate down to account for improved child health since the underlying studies were conducted, our uncertainty about the reliability of the studies, and because this grant will only fund one round of VAS per year (rather than the normal two). (More)
• Co-delivering VAS alongside SMC is very cheap. Malaria Consortium’s SMC campaign platform already exists in both states, so the additional cost of delivering VAS alongside it is low. We estimate that it costs Malaria Consortium around $0.45 to reach each child with one round of VAS as part of this program. This is somewhat lower than our estimate of the one-round cost of reaching a child with VAS through Helen Keller International (~$0.65), which is already low relative to the other direct delivery programs GiveWell funds. (More)

Here is a summary of our cost-effectiveness analysis, using estimates for one state (Bauchi) as an example.

You can see the simple cost-effectiveness analysis for this grant here and the full version here.

The other factors informing our decision to make this grant are:

• Malaria Consortium as an implementer (more). GiveWell has previously funded Malaria Consortium to deliver SMC in a number of countries, including Nigeria. Although this would be GiveWell’s first grant to Malaria Consortium for VAS, we expect that Malaria Consortium is well placed to deliver the program and will implement it to a high quality.
  • Malaria Consortium has a track record of supporting SMC campaigns that achieve high coverage in Nigeria.
  • It conducted two recent pilot studies testing co-delivery of VAS and SMC in Nigeria. We would expect this experience to inform its implementation and help it deliver the program to a high standard.
  • Our qualitative assessment of Malaria Consortium as an organization is highly positive, even compared to GiveWell’s other top charities.

• Learning value (more). This approach to delivering VAS has the potential to be expanded to other locations where there are also SMC campaigns, but so far it has only been tested in two small pilots. Making this grant could help us understand how promising the approach is for scale-up elsewhere (e.g. whether Malaria Consortium is able to achieve high coverage at scale, and whether there are unanticipated problems we haven’t considered). Malaria Consortium is also planning to conduct additional monitoring alongside the program to help us understand what impact it has on the delivery of existing health services (e.g. by tracking coverage of interventions like deworming—that are normally delivered alongside VAS—over time).

1.4 Main reservations

• Will this grant reduce access to existing child health services? During our grant investigation we heard some feedback that this approach to delivering VAS could reduce access to existing child health platforms. This could happen via reducing VAS coverage during the other campaign round (e.g. if caregivers start to expect to receive VAS at home rather than at clinics), or by reducing coverage of other child and maternal health services that are normally co-delivered alongside VAS, such as deworming, if caregivers are less motivated to go to health clinics because they’ve already received VAS. We try to account for these risks in our analysis, but we could be underestimating the risk. Malaria Consortium is planning to conduct additional monitoring as part of the grant to gather more evidence about how big a problem this is. (More)
• Should we fund another VAS implementer in these locations using existing platforms? Considering the risks of a relatively new approach, an alternative would be to fund another organization (e.g. Helen Keller International or Nutrition International) to support VAS delivery using the existing VAS platform in these states. This would also mean we could fund two rounds of VAS per year rather than just one. (More)
• Wider uncertainties about VAS. We have a number of other questions about VAS programs that are not specific to this grant. These include (more):
  • The impact of VAS on mortality. While our analysis of VAS is underpinned by a meta-analysis of 19 randomized controlled trials, we have some doubts about this evidence (e.g. some studies finding little or no impact for unclear reasons, most studies being conducted in the 1980s and 1990s when disease environments were different). This means we’re unsure what impact VAS will have on mortality today.
  • Vitamin A deficiency rates. We expect that vitamin A status is the main mediator of the impact of VAS on mortality, but data on deficiency rates today is relatively limited. In recent years many countries have introduced programs to fortify staple crops with vitamin A, and it’s possible this means that deficiency rates have fallen faster than we expect. Anecdotally, we have heard that a recent survey in Nigeria found low prevalence of deficiency. Our understanding is that the survey hasn’t yet been published, and so we haven’t yet been able to review the results. If the survey finds deficiency is lower than we currently estimate, this would reduce our estimate of the impact of the program. We plan to review data from this survey if and when it becomes available.
  • Are there drawbacks to VAS campaigns we are missing? We think that VAS is one of the most cost-effective programs donors can support. However, our understanding is that there is relatively little investment in VAS from other global health funders. This raises a concern that there are drawbacks to VAS that we are missing.

Funding for the grant

This grant was funded by the Rauch Family Foundation, on GiveWell's recommendation.

2. Planned activities and budget

2.1 Background

In Nigeria, VAS is normally distributed with other maternal and child health services (e.g. deworming, nutrition screening, iron & folic acid supplementation)5 in campaigns called Maternal, Newborn and Child Health Weeks (MNCHW). These campaigns are organized by the health authorities in each Nigerian state.6 These are meant to take place twice a year and involve delivery of VAS at health clinics, with outreach events to more remote areas.7

GiveWell currently funds another organization, Helen Keller International, to support VAS through these campaigns. In 2023, GiveWell funded Helen Keller International to support VAS through this platform in five states.8 However, in many non-GiveWell supported states in Nigeria, we believe that the proportion of children receiving VAS through this platform is relatively low (more below).

2.2 What we think this grant will do

This ~$1.4m9 grant will fund Malaria Consortium to deliver VAS to 6 - 59 month olds in 2 of Nigeria’s 36 states (Bauchi state and Niger state). The grant funds delivery of VAS in Bauchi state in 2024 only, and Niger in both 2024 and 2025 (reasoning in footnote).10

Niger and Bauchi states are highlighted in blue.11

Malaria Consortium’s seasonal malaria chemoprevention (SMC) program already operates in both states. To deliver SMC, community distributors funded and trained by Malaria Consortium go door-to-door each month for 4-5 consecutive months to deliver antimalarial drugs to young children.12

We expect this grant will fund the costs of delivering 1 round of VAS to children alongside SMC. In particular, the grant will fund training for distributors to administer vitamin A supplements and additional labor time that distributors will require to provide an extra service on top of SMC.13 Malaria Consortium expects that it will deliver VAS as part of the final SMC cycle in each state (in October).14 Community distributors will be instructed to give the SMC drugs to children first, and then deliver VAS after a 30 minute wait (to ensure the child does not vomit up the drugs).15 VAS is delivered by cutting open a capsule containing vitamin A in solution and squeezing the contents into children's mouths.16

Malaria Consortium has conducted 2 previous pilot studies testing this approach. These were in Sokoto state in 2019 and Bauchi state in 2021. Malaria Consortium told us that VAS was removed from the services available in the closest MNCHW campaign round in the locations where the pilots took place. In the other MNCHW campaign round, VAS was provided as normal. Malaria Consortium told us that it expects the same model to apply for this grant.17

This grant will solely cover the additional costs of delivering VAS alongside SMC. It includes the cost of procuring VAS capsules, although Malaria Consortium may be able to get these donated for free (details in footnote).18 Malaria Consortium is also planning to conduct some additional monitoring as part of this program, but this will be funded by a separate research funding pot and is not included in this grant. (More)

3. The case for the grant

3.1 Cost-effectiveness

Our best guess is that this grant will be 17x (Niger state) to 54x (Bauchi state) as cost-effective as unconditional cash transfers (GiveWell’s benchmark for comparing different programs). At the time of writing this page, GiveWell’s funding bar is to fund grants that we estimate to be ~10x or more as cost-effective as cash transfers.19 The main benefit that we expect from this grant is reduced child mortality. Our best guess is that the grant will avert ~960 child deaths20 from reduced infectious disease.

We set out the reasons why we think VAS is generally a cost-effective intervention in our separate report on VAS. The main reasons we expect this specific grant to be cost-effective are discussed below.

Impact on VAS coverage

We think this grant will increase the number of children who receive VAS. This is based on our analysis of two factors:

• The proportion of children who would receive VAS without the program
  • In Nigeria, children can access VAS during routine healthcare appointments and biannual campaigns called Maternal, Newborn and Child Health Weeks (MNCHW).21
  • As part of our grant investigation, we estimated the proportion of children who would be reached through these routes if we did not provide funding for Malaria Consortium’s program. We refer to this estimate as “counterfactual VAS coverage.” In general, the higher this proportion, the less additional impact we think our funding will have.
  • Our estimates of counterfactual VAS coverage are based on 4 household surveys conducted in northern Nigeria between 2018 and 2023 that asked caregivers whether their child received VAS in the prior 6 months (compiled in this spreadsheet). Using our preferred assumptions, we estimate that 39% (Bauchi state) / 37% (Niger state) of children in the overall target population would receive VAS in the absence of GiveWell funding, and 43% (Bauchi state) / 40% (Niger state) of children that will be reached by Malaria Consortium with GiveWell funding would receive VAS in the absence of this funding (discussion in footnote).22
  • We have higher confidence in our estimates for Bauchi state (where we have 2023 data from surveys conducted by Malaria Consortium as part of its regular SMC program monitoring) than Niger state (where we don’t). Our Niger state estimates are based solely on two household surveys conducted in 2018. We decided not to adjust these estimates because (a) they are in the same range as other states and (b) like many of the other states included in our analysis, we think Niger state is one of the states receiving the least support from external funders for VAS. (details in footnote).23
• The proportion of children who will receive VAS with the program
  • Our best guess is that 67% of targeted children will receive VAS through Malaria Consortium’s program (albeit only once a year, while children will continue to have access to VAS through the other routes for the second recommended annual dose).
  • This estimate is based on post-campaign survey data from Malaria Consortium’s 2 earlier pilots (in Sokoto state in 2019 and Bauchi state in 2021) testing co-delivery of VAS and SMC. These surveys found 59% and 82% VAS coverage respectively. We also put some weight on coverage data from its SMC program in Nigeria between 2018 and 2021, which found 78% coverage overall.24
  • Note that this estimate reflects only the proportion of children that we think will be reached by Malaria Consortium’s campaigns, not the overall proportion of children who will receive VAS in these states. We expect this latter proportion will be higher, since some children will still have access to VAS through other routes (e.g., as part of Nigeria’s routine immunization schedule).25 Discussion in footnote on how we account for this distinction quantitatively.26

High child mortality in these states

Child mortality is very high in northern Nigeria, where Niger state and Bauchi state are located. Our estimates of child mortality are based on estimates from the Global Burden of Disease Project (GBD), to which we apply several adjustments.27 Overall, we estimate that the annual child mortality rate for 6 - 59 month old children is around 1% in Niger state and 2% in Bauchi state. These mortality rates are relatively high compared to other locations where GiveWell funds VAS programs (range of 0.67% to 1.76%).28 While we’re uncertain about the GBD estimates,29 this fits with our impression from other sources that child mortality in northern Nigeria is particularly high.30

Impact on all-cause child mortality

VAS probably reduces child mortality by a modest but meaningful amount. We estimate that receiving one round of VAS reduces a child’s mortality risk by ~3% in Niger state and ~5% in Bauchi state.

Our main estimate of the mortality impact of VAS is based on a meta-analysis of 19 randomized controlled trials (Imdad et. al. 2017) that finds VAS reduces child mortality by 24%.31 We adjust this estimate down to account for improved child health since the underlying studies were conducted and our uncertainty about the reliability of the studies. See these rows in our main VAS cost-effectiveness analysis for more details on our reasoning.

For this grant, we produced a new estimate of the impact of a single round of VAS on mortality. To date, GiveWell has only funded programs that aim to deliver two rounds of VAS per year, and most studies in the meta-analysis we rely on tested two or three annual rounds of VAS.32 By contrast, Malaria Consortium’s program will only deliver a single round of VAS per year.

Our best guess is that a single round of VAS is 57% as effective at reducing mortality as two rounds per year. We then adjust this downward to 53% to account for Malaria Consortium’s program altering the interval between campaigns (details in footnote).33 Our analysis is based on two main elements:

• Prior estimate: A best guess that one round of VAS is slightly more than half as effective (55% impact on mortality) as two rounds, due to a diminishing returns effect (i.e., the first dose provides strong protection against mortality, with each successive dose providing relatively less protection).
• Evidence from VAS trials: We also assign a small amount of weight (10%) to two studies in the meta-analysis that tested one round of VAS and had 12 month follow-up periods. We use these studies because we’d expect these studies to be the closest guides to the impact of Malaria Consortium’s program (which will also deliver one round of VAS per year).34 The overall weighted reduction in mortality in these studies was larger than we’d expect (-18%) based on our prior (-13%), leading us to update our overall estimate upward. However, we only assign a small amount of weight to these studies because we would expect data from just 2 trials to be noisy, and so together they comprise less than 10% of the overall meta-analysis weight.35

See this spreadsheet for our full reasoning and calculations. Overall, we’re uncertain about this estimate, which is based on limited empirical data and only a light-touch review. We hope to improve our analysis on this question in the future (e.g., through conversations with VAS experts).

Co-delivering VAS alongside SMC is very cheap

We estimate that it costs Malaria Consortium around $0.45 to reach each child with one round of VAS as part of this program. This is somewhat lower than our estimate of the one-round cost of reaching a child with VAS through Helen Keller International (~$0.65)36 , which is already lower than most other direct delivery programs GiveWell funds.

Our estimate of the cost per child reached with VAS is based on a cost projection of $0.30 per child targeted shared by Malaria Consortium and our estimate that the program will reach 67% of these children.37 Our understanding is that the $0.30 cost projection was based on a costing study conducted by Malaria Consortium as part of its earlier Bauchi pilot of the program in 2021, that found the program cost was $0.24 per child reached.38 It then increased this estimate to account for increasing costs over time39 (a rough GiveWell analysis suggests that this was approximately a 40% increase in costs, details in footnote).40 We think that Malaria Consortium’s estimates seem reasonable, although we didn’t closely investigate the assumptions used or try to understand the methodology of the costing study in detail.

Our intuition for why this program is so cheap is that Malaria Consortium’s SMC campaign platform already exists in both states. The fixed costs of setting up and administering this platform have already been incurred, and so the additional cost of delivering VAS alongside it is low.

Other cost-effectiveness analysis updates

As part of our grant investigation, we also made a number of other updates to our cost-effectiveness analysis. These are discussed in a footnote.41

3.2 Malaria Consortium as an implementer

GiveWell has previously funded Malaria Consortium to support SMC campaigns in a number of countries, including Nigeria (more details on our separate report on its SMC program). Although this would be GiveWell’s first grant to Malaria Consortium for VAS, we expect that Malaria Consortium is well placed to deliver the program and will implement it to a high quality. Reasons to think this include:

• Malaria Consortium has a track record of supporting SMC campaigns that achieve high coverage in Nigeria. Our analysis of its surveys found an overall coverage rate of 78% between 2018 and 2021. We think that its door-to-door delivery platform for SMC is a key reason for this high coverage rate, and will have similar benefits for coverage of VAS.
• Malaria Consortium conducted two recent pilot studies testing co-delivery of VAS and SMC in Nigeria. We would expect this experience to inform its implementation and help it deliver the program to a high standard. For example, Malaria Consortium made a number of changes to the program in response to the Sokoto pilot’s findings (e.g. reducing community distributors’ daily targets for the number of children reached and adding an additional day to the distribution period because of concerns about overwork).42
• Our qualitative assessment of Malaria Consortium as an organization is highly positive. We think it is transparent, thoughtful, and has staff with a high level of operational and technical expertise. When we have requested feedback on Malaria Consortium from national malaria programs and other malaria contacts, we have generally heard positive (and often very strongly expressed) comments. More in this section of our report on Malaria Consortium’s SMC program.

3.3 Learning value

This approach to delivering VAS has the potential to be expanded to other locations where there are also SMC campaigns, but so far it has only been tested at small scale in Malaria Consortium’s initial pilots. We see this grant as an opportunity to understand how effective the program is at scale.

Specific questions we hope to answer as a result of this grant are:

• What impact will the program have on VAS coverage? This will be collected via a VAS question43 in surveys conducted by Malaria Consortium before and after the campaign is delivered.
• What impact will the program have on other interventions normally delivered alongside VAS? As we discuss below, one of our key reservations about this grant is that it could reduce coverage of other health services (e.g., deworming) normally co-delivered with VAS in the MNCHW campaigns.
• How do the Nigeria health authorities (at the federal and state level) respond to the program? We have asked Malaria Consortium to share ad hoc feedback on how the authorities and other partners are responding to the program to understand this. We may also seek out conversations with these partners.

Malaria Consortium is planning to conduct additional monitoring as part of the program to help us answer these questions. Before we made the grant, we agreed with Malaria Consortium that this work would be funded out of a separate research budget it holds and so is not included in this grant decision. At the time of writing, we have not yet agreed on a monitoring approach with Malaria Consortium, but our initial proposal to Malaria Consortium includes44 :

• A baseline survey to gather data on coverage of VAS and other interventions co-delivered alongside VAS (e.g., deworming).
• 3 subsequent survey waves to gather data on coverage of the same interventions immediately after the Maternal and Newborn Child Health Weeks.
• 3 waves of qualitative interviews with health clinic staff to understand their perspectives on the program and its impact on the Maternal and Newborn Child Health Weeks.

We hope that this approach will provide us with evidence on the impact of the program on existing health services. However, it’s possible that the monitoring will be less informative than we hope (e.g., because a before-and-after design is not able to isolate the impact of Malaria Consortium’s program from other factors that affect MNCHW coverage).

4. Risks and reservations

Risk of reducing access to existing child health services. In the external conversations we had while investigating this grant, we heard some concerns that this approach to delivering VAS does not adhere to the national strategy for delivering vitamin A, and could reduce access to other child health services delivered through the MNCHW. VAS campaigns are meant to take place twice a year in Nigeria, and this program would only deliver VAS once a year, using a different campaign platform (more above).

This raises concerns about the program including:

• It could reduce VAS coverage during the other annual campaign round (e.g. if caregivers start to expect to receive VAS at home and are less motivated to travel to clinics).
• It could reduce coverage of the other child and maternal health services that are normally co-delivered alongside VAS, such as deworming, if caregivers are less motivated to go to health clinics because they’ve already received VAS.45
• It could change the interval between VAS rounds. Malaria Consortium expects to deliver VAS in October, whereas the second VAS campaign round may take place in the months after this.46 This means that children who receive both annual VAS rounds are likely to receive VAS ~4-5 months after the previous round, followed by a ~7-8 month gap (rather than every 6 months). We’re unsure what impact this could have on the effect of VAS on mortality.

We have attempted to account for these concerns in three ways, although we’re unsure about each of them:

• Adjustment for negatively impacting other programs. Our cost-effectiveness analysis includes a -15% adjustment to account for the risk of harming coverage of other interventions. This is because we think that this concern is plausible. However, the specific adjustment we use should be considered a very rough guess, and is not based on any empirical evidence.
• Monitoring the impact of the program on other interventions. Alongside this grant, Malaria Consortium has agreed to conduct additional monitoring to understand whether this funding is reducing the proportion of caregivers accessing other health services (more above). We expect that this will provide data to update our -15% best guess (e.g., via household surveys to understand coverage of other services before and after the program is delivered). However, at the time of writing, we have not finalized the details of this monitoring with Malaria Consortium. It’s possible that the evidence it gathers will be less informative in answering this question than we expect.
• Adjustment for changing the interval between VAS campaigns. We incorporate a -5% adjustment into our estimate of the impact of VAS on child mortality. This is designed to account for the risk that this program will alter the interval between VAS campaigns, and that this will make VAS less effective at averting mortality. Our adjustment is based on (i) our estimate of the impact of each round of VAS (discussed above), (ii) guesses about the impact of changing the interval between rounds on mortality and (iii) an estimate of the proportion of children who will receive VAS outside this program (details in footnote).47 This adjustment should also be considered a rough estimate, and is not based on any conversations with experts.

Should we fund another VAS implementer in these locations using existing platforms? Co-delivery with SMC is a relatively new approach to delivering VAS. Considering the risks we describe above, an alternative would be to fund another organization (e.g. Helen Keller International or Nutrition International) to support VAS delivery using the existing MNCHW platform in these states. This would also mean we could fund two rounds of VAS per year rather than just one.48

On balance, we have decided against doing this because:

• We think this program is likely to be a particularly effective use of funding, because the costs of layering VAS onto the SMC platform are very low (more above).
• The learning value of trying a new approach and gathering evidence about how well it works. This could help inform decisions about whether to scale up the program in more locations in the future.
• Our qualitative assessment of Malaria Consortium is particularly positive (more above).

Other reservations about VAS. We have a number of other questions about VAS programs that are not specific to this grant. These include:

• The impact of VAS on mortality. We estimate that receiving VAS as part of this grant will reduce all-cause child mortality by 5% (more above). This is based on a meta-analysis of 19 randomized controlled trials (Imdad et. al. 2017). However, we have a number of reservations about this evidence, including:
  • How to interpret variation between studies of VAS. Most of the studies in the meta-analysis took place in the 1980s and 1990s. Two out of the three more recent studies found lower effect sizes than the older literature.49 This includes the DEVTA trial, whose sample size is larger than the combined sample size of the remaining studies.50 We don’t know how to account for this variation between studies.
  • We don’t have a clear understanding of how vitamin A reduces the risk of infectious disease mortality.51 This makes it hard to know what we should expect the mortality impact of VAS should be today, in a different disease environment to when the studies were originally conducted.
• Vitamin A deficiency rates. We think vitamin A status is the main mediator of the impact of VAS on mortality. This means our analysis is sensitive to the rate of deficiency in locations where we fund VAS today. But we’re uncertain about contemporary deficiency rates and have relatively little high-quality data with which to estimate them. In Nigeria, the latest survey data we have seen is from a national survey conducted in 2001. We then adjust these to account for change over time and to create state-level estimates using more recent data on indicators we think may be proxies for deficiency (e.g. poverty rates, anemia rates).52 We would expect this method to be relatively imprecise. Since the survey was conducted, Nigeria has introduced programs to fortify staple crops53 , but we’ve seen little evidence on how effective this has been. It’s possible that these programs mean deficiency rates are lower than we expect, implying we’re overestimating the impact of VAS.
  • Another reservation is that, before we finalized this grant, we heard an anecdotal report of a new vitamin A deficiency survey in Nigeria finding low VAD prevalence. Our understanding is that this survey has not been published yet, and so we haven’t yet been able to review the results. If this survey finds deficiency is lower than we currently estimate, this could reduce our estimate of the impact of the program. We plan to review data from this survey if and when they become available.
• Are there drawbacks to VAS campaigns we are missing? We think that VAS is one of the most cost-effective programs donors can support. However, our understanding is that there is relatively little investment in VAS from other global health funders. This raises a concern that there are drawbacks to VAS that we are missing.

5. Plans for follow up

• We will have occasional calls with Malaria Consortium to discuss its work, including progress on its negotiations with the health authorities and partners in each state and the process for applying for vitamin A capsule donations.
• We will discuss and agree details for monitoring of the program (discussed above). We expect this to include discussions about GiveWell’s research objectives, Malaria Consortium submitting a research protocol, and GiveWell giving feedback on that protocol.
• We will use data from Malaria Consortium’s post-campaign survey (expected October 2024) to decide whether or not to disburse the 2025 funding for Niger state. We also expect to make a decision about whether to renew funding for Bauchi state around this time.

6. Internal forecasts

7. Our process

We identified this grant opportunity through conversations with Malaria Consortium. As part of our grant investigation process, we had conversations with Malaria Consortium and two other VAS implementing organizations: Helen Keller International and Nutrition International.

To generate a grant-specific cost-effectiveness estimate, we used our existing cost-effectiveness model for VAS and updated various parameters to match the specifics of this program.

For internal review, a Senior Program Officer gave feedback on various aspects of the grant investigation, and a Senior Research Associate gave feedback on our estimate of the impact of one round of VAS on mortality.

Sources

• 1

  "Vitamin A deficiency is a major public health problem affecting an estimated 190 million preschool-age children, mostly from the World Health Organization (WHO) regions of Africa and South-East Asia (1). Infants and children have increased vitamin A requirements to promote rapid growth and to help combat infections. Inadequate intakes of vitamin A at this age could lead to vitamin A deficiency, which, when severe, may cause visual impairment (night blindness) or increase the risk of illness and mortality from childhood infections such as measles and those causing diarrhoea (2)." World Health Organization, "Guideline: Vitamin A supplementation in infants and children 6-59 months of age," 2011, pg. 2.

• 2
  • In settings where vitamin A deficiency is a public health problem, vitamin A supplementation is recommended in infants and children 6–59 months of age as a public health intervention to reduce child morbidity and mortality (strong recommendation). The quality of the available evidence for all-cause mortality was high, whereas for all other critical outcomes it was moderate to very low. The quality of the available evidence for outcomes in human immunodeficiency virus (HIV)- positive children was moderate for all-cause mortality." World Health Organization, "Guideline: Vitamin A supplementation in infants and children 6-59 months of age," 2011, pg. 1.
  • One dose of 100,000 IU of vitamin A is recommended for infants aged 6 to 11 months of age, and a 200,000 IU dose of vitamin A is recommended for children 12 to 59 months of age every four to six months. Table 1, World Health Organization, "Guideline: Vitamin A supplementation in infants and children 6-59 months of age," 2011, pg. 5.

• 3

  Read about our most recent grant to Helen Keller International for VAS here and our most recent grant to Nutrition International for VAS here.

• 4

  This understanding of what the funding will be used for is based on Malaria Consortium, comment on a draft of this page, January 8th, 2024.

• 5

  See this spreadsheet for the services co-delivered alongside VAS in states where GiveWell funded Helen Keller International in 2021.

• 6

  “The National Primary Health Care Development Agency (NPHCDA) is the public institution with primary responsibility of implementing the MNCHW. The agency provides guidance for implementation of the weeks. The implementation of the weeks at state level is by the State PHCDA with funding predominantly from the State Ministry of Health and development partners.” UNICEF, Evaluation of the Maternal Neonatal and Child Health Week in Nigeria, November 2016, pg. 19.

• 7
  • “The distribution campaigns, known as MNCHW (Maternal Newborn and Child Health Weeks), span over a week and involve delivery at the primary health care facility level combined with outreach events organized in remote areas.” Helen Keller International, 2023 room for more funding report, pg. 24.
  • “In Nigeria, VAS campaigns are carried out in May/June and October/November whereas SMC administration typically happens from July to October each year. This timing shows an overlap only in October, therefore making integration possible only once a year.” Oresanya et. al., 2022

• 8

  See this section of GiveWell’s renewal grant page for Helen Keller International’s VAS program.

• 9

  See the budget breakdown here. Note: The funding is conditional on Malaria Consortium obtaining support from the health authorities in each state, and the 2025 funding in Niger state is conditional on Malaria Consortium achieving 55% coverage or more in the Niger 2024 campaign, according to the 2024 post-campaign survey.

  Note: the specific value of this grant was $1,439,330, but we plan to transfer $1,440,000 to Malaria Consortium. This is because we received exactly $1,440,000 from a donor to fund this grant.

• 10

  During our grant investigation, we heard from Malaria Consortium that it may be involved in a program to distribute azithromycin to children in the same age range in some states in northern Nigeria from 2024 onwards. If the program went ahead, Malaria Consortium expects that it would be delivered in Bauchi state starting in 2025, but would not be delivered in Niger state in 2025. At the time we made the grant, we were unsure what impact delivering VAS and azithromycin to the same population at around the same time would have, since our understanding is that azithromycin reduces mortality through similar mechanisms to VAS (e.g., averting diarrhea mortality), and it’s possible that VAS would have less additional benefit.

  To account for this, this grant only provides funding for Bauchi state in 2024. We are planning to consider this question in more detail in the future and consider renewal of funding in Bauchi state at a later stage.
  This understanding is based on an email from Malaria Consortium, October 10 2023 (unpublished) and a conversation with Malaria Consortium, October 12 2023 (unpublished).

• 11

  Created using https://www.mapchart.net/.

• 12
  • “The community-based distribution of SMC medicines is at the heart of SMC. Some community distributors are CHWs who are trained to provide basic health services at the community level. In most countries, however, most community distributors are volunteers recruited and trained specifically for the SMC campaign. Community distributors typically work in pairs, going door-to-door in the communities with whom they work to identify eligible children and administer SMC medicines.” Malaria Consortium, SMC philanthropy report 2022, pg. 20.
  • The number of SMC cycles delivered each year in Nigeria differs depending on the state. As of 2023, 4 cycles were delivered annually in Niger state and 5 cycles in Bauchi state. This is based on:
    • Bauchi state: Malaria Consortium, comment on a draft of this page, January 8th, 2024 and email to GiveWell, January 11th, 2024 (unpublished).
    • Niger state: email from Malaria Consortium, October 5 2023 (unpublished).

• 13

  This understanding of what the funding will be used for is based on Malaria Consortium, comment on a draft of this page, January 8th, 2024.

• 14

  Conversation with Malaria Consortium, September 20 2023 (unpublished).

• 15

  This understanding is based on Oresanya et. al. 2022, a study that reports the findings from Malaria Consortium’s initial pilot of the VAS and SMC co-implementation program in Sokoto state in 2019.
  “All three participant groups expressed that workload and role of the CDDs was affected by the integration of VAS with SMC. Caregivers reported that some CDDs were not patient enough to wait 30 min, which is required between the administration of SMC and VAS to see if the child vomits.”

• 16
  • "Step 3: Administering a Vitamin A supplement to a child
    • Make sure the child is calm and request the caregiver to hold the child.
    • Hold the vitamin A capsule upright with the nipple pointing up.
    • Cut the nipple with a clean pair of scissors to open the capsule while it is still in the upright position
    • Squeeze all the contents (oil drops) of the vitamin A capsule into child’s mouth.
    • Throw the empty capsule into a dustbin."

  Helen Keller International’s Tanzania social mobilization toolkit: VAS administration guide.

  • Note that this description of how VAS is delivered comes from Helen Keller International's VAS program. We assume that Malaria Consortium uses the same method of delivering VAS.

• 17

  Malaria Consortium, call with GiveWell, September 20 2023 (unpublished).

• 18

  Malaria Consortium informed us that it obtained vitamin A capsule donations in its earlier pilots from the NGO Vitamin Angels. At the time we made the grant, Malaria Consortium had not yet confirmed it would be able to obtain these capsules. The grant budget therefore includes the cost of these capsules (see here). If Malaria Consortium is able to obtain donated capsules, we think it’s likely that there will be some leftover funding from the program. We have agreed with Malaria Consortium that this funding could be used either for a future VAS program, or as part of its larger SMC budget.

• 19

  This benchmark is based on ‘moral weights’, a system we use to quantify the benefits of different impacts (e.g. increased income vs reduced deaths). We benchmark to a value of 1, which we define as the value of doubling someone’s consumption for one year. Our estimate of the value of direct cash transfers is 0.00335 per $. For more on how we use moral weights, see this document.

• 20

  See this row in our cost-effectiveness analysis. We estimate that the program will avert 683 deaths in Bauchi state and 139 in Niger state in 2024. The grant funds the program in Niger state for 2 years. 683 + (139 x 2) = ~960.

• 21
  • Vitamin A supplements are recommended at 6 and 12 months in Nigeria’s routine immunization schedule, although we’re unsure what proportion of children actually get VAS at each. See this page for more details.
  • “The distribution campaigns, known as MNCHW (Maternal Newborn and Child Health Weeks), span over a week and involve delivery at the primary health care facility level combined with outreach events organized in remote areas.” Helen Keller International, 2023 room for more funding report, pg. 24.

• 22

  Key assumptions:

  • Source weights: We use the following weights.
    • 65% weight to Malaria Consortium’s recent VAS survey, since it is the most recent source and captures the outcome we care about directly (VAS coverage among 6 - 59 month olds).
    • 25% weight to the New Incentives rapid assessment data.
    • 5% to each of the DHS 2018 and the National Nutrition and Health Survey (NNHS) 2018.
    • We assigned sources other than the Malaria Consortium surveys some weight because VAS coverage may fluctuate over time, e.g., if a campaign took place for one round in a given year but not two rounds.
  • Self-report bias: We assume the true proportion of children who received VAS is 10% lower than the proportion of children reported to receive VAS by caregivers. This factors in (i) the impact of social desirability bias and (ii) the risk that caregivers might accidentally identify another commodity as vitamin A. This is a rough guess.
  • Children reached by Malaria Consortium are more likely to receive VAS in any case: The estimates above refer to the proportion of children in the total target population who we think received VAS in the prior 6 months, based on survey data. We also adjust these figures upwards by 10% to estimate the proportion of children reached in Malaria Consortium's program who we think would otherwise have received VAS in the absence of its program. This is because we expect that children treated by Malaria Consortium will be disproportionately likely to seek out VAS in the absence of its program. In particular, we assume that households where Malaria Consortium successfully delivers VAS will be more likely to have positive attitudes toward VAS. We also think it will be easier to deliver VAS to households in less remote areas, which means that children reached by the program may be more likely to have better access to VAS already because of greater proximity to health clinics. . The 10% estimate we use is a rough guess.

• 23

  Our understanding of the funding situation in Niger state is based primarily on an unpublished analysis from Helen Keller International, suggesting that Niger state receives limited support for VAS from external partners.

• 24

  We assign 20% weight to the Sokoto pilot estimate, 50% weight to the Bauchi pilot estimate, and 30% weight to the SMC coverage data. We also apply a -10% adjustment for self-report bias (the same adjustment that we use for counterfactual VAS coverage). See this section and this section in our cost-effectiveness analysis for further details and sources.

• 25

  Vitamin A supplements are recommended at 6 and 12 months in Nigeria’s routine immunization schedule, although we’re unsure what proportion of children actually get VAS at each. See this page for more details.

• 26

  Our understanding is that the Sokoto and Bauchi pilots conducted by Malaria Consortium asked caregivers whether their child received VAS in the prior 6 months (from any source), not just in Malaria Consortium’s campaign. However, VAS coverage at baseline in these pilots was extremely low (2% in the Sokoto pilot and 1% in the Bauchi pilot). We therefore think it is reasonable to attribute virtually all of the increase in coverage in these studies to Malaria Consortium’s campaign.

  This implies 67% will not be a good guide to VAS coverage at scale, since we think ~35% to ~40% of children would receive VAS through other sources in the absence of the program (more above), and we’d expect some children not reached by Malaria Consortium would continue to receive VAS from one of these sources.

  We account for this by:

  • Using 67% as our best guess of VAS coverage that will be achieved through Malaria Consortium’s program only.
  • Estimating the proportion of children reached by Malaria Consortium who would have been reached with VAS through other sources in the absence of the program (43% in Bauchi, and 40% in Niger). This is distinct from the proportion of children in the target population who would have been reached with VAS in the absence of the program.
  • Using these two estimates to calculate the impact of the program on coverage (e.g., in Bauchi state, coverage = the target population x 67% x (100% - 43%). See this section of our analysis for our calculations.

  Note that this is different from deducting counterfactual VAS coverage (43% in Bauchi) from Malaria Consortium coverage (67%) to estimate the impact of the program. This is because we expect some children not reached by Malaria Consortium would continue to receive VAS from one of these sources (e.g., routine immunizations), meaning this method would produce an underestimate of the program’s impact on coverage.

  “Coverage of VAS at baseline was low (2 percent; 95 percent CI = 0.4—7.0) and increased at endline (59 percent; 95 percent CI = 47.0 – 70.7; p = 0.001) (Table 5).” Oresanya et. al. 2022.

  “VAS coverage increased from 1.2 percent at baseline (without SMC integration) to 82.3 percent at endline (with SMC integration), in both project LGAs.” Malaria Consortium, "Integrating seasonal malaria chemoprevention and vitamin A supplementation: Lessons learnt from Nigeria", pg. 3.

• 27

  The adjustments we use are:

  • Converting the GBD mortality estimates (which use under 7 days old, 7 - 27 days old, 28 - 364 days old, and under 5 years old) into a mortality estimate for 6 - 59 month old children.
  • An adjustment to account for the impact of SMC on mortality (i.e., mortality being lower than the GBD estimates suggest because children who receive the program also benefit from receiving SMC). Our analysis is based on our cost-effectiveness analysis of SMC and our calculations are available here. Overall, this reduces our estimate of child mortality by ~0.2 percentage points in both Bauchi and Niger state.
  • An adjustment to account for existing VAS programs. This is because the GBD mortality estimates that we use aim to capture the overall mortality risk in each location. Our best guess is that these estimates already incorporate the benefits of existing VAS delivery (taking into account that some people are already protected by VAS and therefore have lower morality). To estimate mortality among children who would not be protected by VAS without this grant (and who we would expect to have higher mortality rates than average), we apply an adjustment using (i) the impact of VAS on mortality, (ii) the GBD mortality estimates we previously calculated, and (iii) the proportion of children who we think would receive VAS in the absence of the program. Our calculations are available here.

• 28

  This range is based on our latest (unpublished at the time of writing) VAS cost-effectiveness analysis, which we expect to be published in Jan-Feb 2024.

• 29

  Our understanding is that the GBD estimates rely on a number of modeling assumptions. We have not investigated all of the modeling assumptions underlying these estimates in detail, and when we have investigated certain assumptions we have found them to be less well-founded and robust than we had initially assumed.

• 30

  For example, IGME (the UN Inter-agency Group for Child Mortality Estimation) estimates that the under-five mortality rate (the number of deaths before age 5 per 1,000 live births) was ~137 per 1,000 in Niger state in 2021 and ~154 per 1,000 in Bauchi state. Data available here. This compares to an average in sub Saharan Africa of 74 deaths per 1,000 live births.

  “Children around the world face vastly different chances of survival. Globally, the under-five mortality rate was 38 deaths per 1,000 live births in 2021, but in sub-Saharan Africa, the rate was 74 deaths per 1,000 live births.” IGME, Levels and trends in child mortality, 2022, pg. 4.

• 31

  The main finding of the review was that VAS reduced all-cause mortality by 12% using a “fixed-effects” statistical analysis. As a sensitivity check, the authors also conducted a “random-effects” analysis that found a mortality reduction of 24%. The main difference between the estimates is the amount of weight put on the DEVTA study, a very large trial of VAS that found lower-than-average impacts of VAS on mortality.

  We use the 24% “random-effects” estimate as the main input into our cost-effectiveness analysis because we think that the impact of VAS might vary substantially across study contexts. See the cell note in this row in our VAS cost-effectiveness analysis for more details.

  “A meta‐analysis for all‐cause mortality included 19 trials (1,202,382 children). At longest follow‐up, there was a 12% observed reduction in the risk of all‐cause mortality for vitamin A compared with control using a fixed‐effect model (risk ratio (RR) 0.88, 95% confidence interval (CI) 0.83 to 0.93; high‐quality evidence). This result was sensitive to choice of model, and a random‐effects meta‐analysis showed a different summary estimate (24% reduction: RR 0.76, 95% CI 0.66 to 0.88); however, the confidence intervals overlapped with that of the fixed‐effect model.” Imdad et. al. 2017, abstract.

• 32

  See this section of our supplementary analysis spreadsheet.

• 33

  We expect that Malaria Consortium will deliver VAS in October in each location. VAS is usually delivered in May / June and October - December in Nigeria through the Maternal, Newborn and Child Health Weeks.

  This estimate assumes that Malaria Consortium's program will change the gap between dose 1 and dose 2 from 6 months to ~7.5 months, and the gap between dose 2 and the following dose the next year from 6 months to ~4.5 months (i.e., dose 2 protects children for ~25% less time—(6 - 4.5) / 6 = 0.25)). We expect that this will somewhat reduce the impact of VAS on mortality.

• 34

  Fisker et. al. 2014 and Venkataro et al. 1996. See this section of our analysis for further details on these studies.

• 35

  See this section of our analysis for further details on these studies.

• 36
  • In particular, we estimate that the cost of supplying two rounds of VAS through Helen Keller International's program in Nigeria is $1.28 per child reached. This suggests that the cost of delivering one round of VAS to a child covered by Helen Keller International's program would be about $0.65. ($1.28 / 2 = ~$0.65.)
  • This cost estimate is based on our latest (unpublished at the time of writing) VAS cost-effectiveness analysis, which we expect to be published in Jan-Feb 2024. Note that this estimate is based on a weighted average from other countries and is not based on Helen Keller’s program in Nigeria, so we have lower confidence in this estimate than the Malaria Consortium estimate.

• 37
  • Source for the $0.30 estimate: Email from Malaria Consortium, 17 August 2023 (unpublished).
  • If the program costs $0.30 per child targeted, and 67% of targeted children are reached by the program, then the program will have a cost of $0.30 / 67% = ~$0.45 per child reached.

• 38

  “The total economic and financial cost per child that received SMC during the baseline cycle was $0.94, while the cost per child receiving both VAS and SMC during the endline cycle was $1.18. Integrating VAS into the usual SMC cycle introduced an additional minimal cost of $0.24 per child.” Malaria Consortium, "Integrating seasonal malaria chemoprevention and vitamin A supplementation: Lessons learnt from Nigeria", pg. 3.

• 39

  Email from Malaria Consortium, 27 October 2023 (unpublished).

• 40

  This estimate is based on the following logic:

  • Malaria Consortium’s costing study estimated a cost per child reached of $0.24. 82% of children were reached with VAS in the study, implying a cost per child targeted of $0.20 ($0.24 * 82% = $0.20).
  • Malaria Consortium’s updated cost estimate per child targeted for this program was $0.30. This is a 50% increase in costs ($0.30 - $0.20) / $0.20 = 0.5.
  • However, Malaria Consortium told us its costing estimate for this study included the cost of capsules (approximately $0.02 per capsule), in case it was not able to obtain donated capsules for the program. Our understanding is that capsules were donated in the earlier program, implying the total cost to Malaria Consortium per child targeted will be lower ($0.30 - $0.02 = $0.28).
  • ($0.28 - $0.20) / $0.20 = ~40% increase in costs.

• 41

  The updates we made include:
  Adjustment to account for the impact of SMC:

  • By necessity, the locations where this program will be delivered also receive SMC. This means mortality is likely to be somewhat lower than we would expect based on GBD estimates alone.
  • We included an adjustment to account for this, using assumptions drawn from our cost-effectiveness analysis of SMC. This adjustment reduced our best guess of the annual 6 - 59 month old mortality rate from 2.23% to 2.01% in Bauchi and from 1.15% to 0.93% in Niger.
  • See this section of our analysis for our calculations.

  Update to the probability that other actors would fund the program in our absence:

  • Our analysis includes adjustments to account for the probability that another funder would fund a program in our absence (“funging”). The intuition for this is that, if a program would have been funded in GiveWell’s absence anyway, the impact of our funding is likely to be lower than we think (for more on this concept, see here). Our full calculations for our funging adjustment are available in this sheet of our cost-effectiveness analysis.
  • As part of this grant investigation, we decided to change our adjustment for the chance that another funder would fund the program in GiveWell’s absence from 25% (our previous value for Nigeria) to 15% (Bauchi state) and 20% (Niger state). See these rows in our cost-effectiveness analysis. Factors feeding into this change included:
    • Helen Keller International told us that 11 states in Nigeria receive no or limited external support for VAS. This suggests to us that the funding gap for VAS in Nigeria in general is large.
    • During the grant investigation, we heard from Nutrition International that the cost of capsules has risen substantially in the past ~5 years (Nutrition International procures capsules for ~60 countries each year to ensure that supply is not an impediment to coverage). This leaves less Nutrition International funding available to support VAS distribution in some countries. We interpreted this to lower the chance that Nutrition International would be able to fund this program in the absence of GiveWell funding.
    • As part of this investigation, we learned that some states in Nigeria have been receiving World Bank funding for a program called ANRIN (Accelerating Nutrition Results in Nigeria). Our understanding is that this project provided $232m of funding to 12 states in Nigeria between 2018 and 2023 (including Niger state). However, the project appears to be finishing in 2023 and our guess is that the funding is not being renewed. The reason we use a different probability in Niger state compared to Bauchi state is we think there might still be a small chance of additional funding from the program.

  “The objective of this component is to scale up a basic package of nutrition-related interventions in the following 12 States: Abia, Akwa Ibom, Gombe, Kaduna, Kano, Katsina, Kogi, Kwara, Nasarawa, Niger, Oyo, and Plateau.” Accelerating Nutrition Results in Nigeria (P162069) project summary, pg. 18 - 19. The $232m figure appears on page 2.
  Other sources:

  • Helen Keller International, 2023 room for more funding report.
  • Conversation with Nutrition International, September 29 2023 (unpublished) and email from Nutrition International, January 15 2024 (unpublished).

  Reduced adjustment for averted anemia: Our main VAS cost-effectiveness analysis includes an unmodeled +9% adjustment to account for VAS averting anemia. We reduced this to +2% during this grant investigation. This is because we think SMC is also likely to reduce anemia through averted malaria, and so our best guess is that the additional reduction in anemia from VAS would be low.

• 42

  “The adoption of a co-design approach to develop the strategy that informed the study protocol provided the opportunity to identify and address potential bottlenecks to the integration. We adjusted community distributors’ daily targets by factoring in a 30-minute wait after administering SMC medicines, before giving vitamin A. We reduced daily targets from 70 children per day to 56 children per day and added an implementation day to avoid compromising quality.” Malaria Consortium, "Integrating seasonal malaria chemoprevention and vitamin A supplementation: Lessons learnt from Nigeria", pg. 3.

• 43

  We have not yet agreed on the specific survey question(s) with Malaria Consortium, but we plan to isolate the impact of Malaria Consortium’s program on coverage by asking whether children received VAS specifically in the recent campaign (in addition to / instead of asking about VAS coverage in the prior 6 months).

• 44

  GiveWell, VAS / SMC Co-delivery grant: Learning and monitoring plan, October 2023 (unpublished).

• 45

  In one of these external conversations, we heard two specific reasons why caregivers might be more motivated to get VAS than the other health services normally delivered alongside it at the MNCHW. We have not investigated these in detail, but we think they are plausible:

  • High acceptability. VAS has few side effects and little intimidation factor compared to other health services (e.g., deworming pills can cause vomiting, vaccines involve needles).
  • Reliable supply. Caregivers have learned to expect that VAS will be available at MNCHW, which encourages them to seek it out.

• 46

  Malaria Consortium reports that it expects VAS campaigns to take place in October or November. However, some campaigns we have seen have taken place later than this (e.g., in December or January). See this spreadsheet.

  “In Nigeria, VAS campaigns are carried out in May/June and October/November whereas SMC administration typically happens from July to October each year. This timing shows an overlap only in October, therefore making integration possible only once a year.” Oresanya et. al., 2022

• 47

  See our calculations here. Key assumptions in the calculations for this adjustment are:

  • Dose 2 is somewhat less effective than dose 1 at averting mortality because of a diminishing returns effect. This is the key assumption in our estimate for the impact of 1 round of VAS, discussed above.
  • Dose 2 is ~20% less effective if delivered after 7-8 months rather than 6 months. This is based on the following logic:
    • Assume that Malaria Consortium's program will change the gap between dose 1 (delivered by Malaria Consortium) and dose 2 from 6 months to ~7.5 months, and the gap between dose 2 and the subsequent dose the following year from 6 months to ~4.5 months (i.e., dose 2 protects children for ~25% less time—(6 - 4.5 / 6 = 0.25)).
    • We use a -20% rather than a -25% adjustment because we’d expect the impact of VAS on mortality to be non-linear (i.e., having the largest effect soon after it is taken, with diminishing returns afterwards).
  • Only ~50% of children who receive dose 1 will receive dose 2. This is based on our estimates of counterfactual coverage for all children (~40%). We adjust this upwards because we’d expect that children who receive the first dose are disproportionately likely to seek out the second (e.g., because of factors including proximity to health clinics, positive attitudes towards VAS etc.).

• 48
  • This is because SMC campaigns only take place for a few consecutive months of the year in these states, whereas MNCHW is intended to take place twice a year.
  • “In Nigeria, VAS campaigns are carried out in May/June and October/November whereas SMC administration typically happens from July to October each year. This timing shows an overlap only in October, therefore making integration possible only once a year.” Oresanya et. al., 2022
  • Note that as of 2023, 5 cycles of SMC were delivered annually in Bauchi state from June to October. This is based on Malaria Consortium, comment on a draft of this page, January 8th, 2024, and email to GiveWell, January 11th, 2024 (unpublished).

• 49

  See this sheet of our supplementary analysis. The two most recent trials are DEVTA (2013) and Fisker et. al. 2014.

• 50

  See this section of our report on VAS.

• 51

  According to the WHO’s guidance on VAS, “The mechanisms by which vitamin A reduces mortality are not fully understood, and it is not clear whether its action is mediated through the correction of underlying deficiencies or through adjuvant therapeutic effects.” WHO guideline: vitamin A supplementation in infants and children 6-59 months of age, 2011, p. 2.

• 52

  See this section of our cost-effectiveness analysis.

• 53

  See this spreadsheet for further details.