Helen Keller Intl — Vitamin A Supplementation (December 2024)

Note: This page summarizes the rationale behind a GiveWell grant to Helen Keller Intl. Helen Keller staff reviewed this page prior to publication.

Published: October 2025

1. Summary

1.1 Background

Helen Keller Intl (Helen Keller)'s vitamin A supplementation (VAS) program is a GiveWell top charity. We think VAS is an effective way to reduce mortality among children under 5. Helen Keller provides financial and technical support to governments for twice-yearly VAS targeting children between 6-59 months old.1

Helen Keller has been a GiveWell top charity since 2017.2

1.2 What we think this grant will do

This grant will fund Helen Keller to provide technical and financial support to government-run VAS programs in six countries and three Nigerian states. We think Helen Keller's support addresses barriers to governments achieving high VAS coverage and leads to more children receiving VAS. (More)

Previous GiveWell grants provided funding for Helen Keller's VAS program in Burkina Faso, Cameroon, Côte d'Ivoire, DRC, Guinea, Mali, Niger, and five Nigerian states (Nasarawa, Benue, Adamawa, Ebonyi, Taraba) through June 2026.3 With this grant, we recommend:

• Renewing funding for one year (July 2026 to June 2027) in Burkina Faso, Cameroon, DRC, Guinea, Mali, Niger, and Adamawa and Taraba states, Nigeria.
• Providing funding for three years (January 2025 to December 2027) for one new state in Nigeria (Niger State)
• Exiting our support for Côte d'Ivoire and two Nigerian states: Benue and Ebonyi. We have not yet made a decision on renewing or exiting support for Nasarawa State, Nigeria.4 (More)

1.3 Why we made this grant

Our best guess is that, overall, this grant will be 25 times (25x) as cost-effective as unconditional cash transfers. We use cash transfers as a benchmark for comparing different programs. The cost-effectiveness of the specific locations funded by this grant varies from 7x to 79x (more). At the time of writing this page, GiveWell’s funding bar for top charities is to fund grants that we estimate to be ~8x or more as cost-effective as cash transfers.

We think this grant will be cost-effective because, in the locations where GiveWell funds VAS:

• The cost per vitamin A supplement is low (around $0.90 per capsule delivered, based on a weighted average of grant size, and $0.80 per capsule delivered, based on a weighted average of the target population served) (more)
• Helen Keller's support increases VAS coverage (more)
• Child mortality in geographies this grant is funding is high (~0.8% to 1.4% risk of death per child per year, depending on location). Our best guess is that VAS results in a meaningful reduction in child mortality (we estimate ~3% to 11%, depending on location) (more).

Here is a summary of our cost-effectiveness analysis, using estimates for one country (Guinea) as an example.5

You can see the simple cost-effectiveness analysis for this grant here and the full version here. We write about our cost-effectiveness model in more detail below.

We also considered qualitative factors outside our model as part of this investigation. We discuss these in detail below.

1.4 Main reservations

• We are uncertain about key CEA parameters: We're most uncertain about the effect of VAS on mortality, current vitamin A deficiency rates, the size of the populations that Helen Keller targets, and counterfactual coverage (routine coverage and the likelihood that domestic governments or other philanthropic actors would step in if we didn't fund the program). (More)
• We are recommending funding two locations that are below our 8x funding bar (Burkina Faso and Adamawa State, Nigeria) on the basis of qualitative reasoning. This is a subjective decision. (More)
• We are uncertain if we should be funding routine delivery of VAS instead of campaigns in some locations. We expect this grant to largely support VAS campaigns. However, Helen Keller has told us it sees strengthening countries' routine delivery of VAS as a critical need to ensure sustainable programs. Going forward, we’re planning to discuss with Helen Keller what the program model should be. We are recommending going ahead with this grant rather than waiting until we're more confident in the delivery model we should support to maintain two-plus years of funding visibility for Helen Keller, for their planning purposes. However, we're not confident this is the right choice. (More)

2. Planned activities and budget

2.1 Activities supported by the grant

Helen Keller typically provides financial and technical support for government-run VAS campaigns targeting children between 6-59 months old twice a year.6 See our review of Helen Keller's VAS program for a detailed summary of their full set of technical support activities. Much of our understanding comes from discussions with Helen Keller, which we've triangulated through conversations with government officials in some of the countries where Helen Keller works.

We think that Helen Keller's program increases VAS coverage in the locations where it operates. See the section below for details.

On this page, we focus on describing VAS campaigns, which are the most common way that governments distribute VAS in the countries where we support Helen Keller. However, Helen Keller also supports routine delivery in some locations.7 We are uncertain about whether we should be funding routine delivery rather than campaigns going forward (see below).

2.2 Budget for grant activities

We're recommending a $23.6 million grant ($22.7 million for renewal and $900,000 for expansion):

These amounts take into account funding received from another donor and approximately $2.1 million in rollover funds that Helen Keller had available from a previous GiveWell grant.8

What we're not funding

We modeled four renewal locations and two proposed locations for expansion to be either significantly below our 8x funding bar or slightly below our funding bar without a strong qualitative case for support.9 We did not fund these opportunities with this grant.

We previously informed Helen Keller that our August 2024 grant would serve as exit funding for any locations we decided to discontinue funding as part of this grant investigation. We are exiting our support for Côte d'Ivoire and two Nigerian states: Benue and Ebonyi.

We also modeled three potential expansion locations as above our 8x funding bar, but we still had significant open questions we wanted to answer before deciding whether to support these locations. Additional details below.

3. The case for the grant

3.1 Cost-effectiveness

Our best guess is that this grant will be, on average, 25 times ("25x") as cost-effective as unconditional cash transfers, GiveWell’s benchmark for comparing different programs.10 Estimated cost-effectiveness varies significantly by location, ranging from 7x (in Burkina Faso) to 79x (in Niger). At the time of writing this page, GiveWell’s funding bar for top charities is to fund grants that we estimate to be 8x or more as cost-effective as cash transfers. The main benefit that we expect from this grant is reduced child mortality. Our best guess is that the grant will avert ~7,700 child deaths from reduced infectious disease that wouldn't have been averted otherwise.11

Our full cost-effectiveness model can be found here.

In the sections below, we set out the main factors informing the cost-effectiveness of this grant. We explain why we think VAS is generally a cost-effective intervention in more detail in our intervention report on VAS.12

Low cost per supplement

Our estimated cost per supplement delivered ranges from approximately $0.50 in Niger to $1.50 in Burkina Faso, with a weighted average value of ~$0.90 based on grant size and ~$0.80 based on target population. Our estimates are based on spending data, coverage estimates, and target populations provided by Helen Keller. We put 75% weight on the estimated cost per supplement based on data from 2022 to mid-2024, and 25% weight on Helen Keller's projected cost and population data for mid-2024 to mid-2025.13 We place more weight on past data based on Helen Keller's feedback that its future costs are likely overestimates, and the fact that Helen Keller had rollover funds in 2024.14

For more detail on how we calculate cost per supplement, see our full cost per supplement analysis.

Increased VAS coverage

We think that Helen Keller's financial and technical assistance for government-run VAS programs increases VAS coverage in the locations where it operates. The barriers to high coverage that Helen Keller's support addresses include:

1. Lack of funding to conduct VAS campaigns. Without Helen Keller's support, we think it's unlikely that governments would implement robust VAS campaigns. For example, they may rely on a less expensive, less intensive routine delivery system for VAS,15 or they may run a more limited campaign,16 each of which we intuitively expect to achieve lower coverage than an intensive door-to-door campaign. Helen Keller alleviates this barrier by providing financial support for VAS campaigns.17
2. The need for sufficient coordination and planning. Our understanding is that limited or late planning for VAS campaigns can lead to lower coverage.18 Helen Keller alleviates this by beginning the planning process for VAS campaigns well before campaigns are set to begin and by serving as part of a national coordinating body for campaigns.19 This coordination role supports improved VAS coverage nationally, not just in the subnational regions where Helen Keller works.20
3. The need for supportive supervision of VAS campaigns. Supervision refers to checking in real time how campaign implementation is going. This includes reviewing data collected during the campaign and confirming that VAS distributors are adhering to protocols. We think governments usually lack the funding to conduct this type of supervision.21 Without supervision, we expect VAS coverage would be lower, as it would be more difficult to identify and promptly address issues such as VAS capsule stockouts or suboptimal communication about the campaign.22 Helen Keller funds supervision of VAS distribution.23 Our understanding is that this is salary support or per-diem/travel support for ministry of health staff.24 Helen Keller teams also directly conduct supervision every day during a VAS campaign.25

Our understanding was informed by conversations with government officials. See footnote for details.26

Counterfactual VAS coverage

Our estimate of the increase in VAS coverage due to Helen Keller's program relies on our best guess of VAS coverage in the absence of campaigns.27 This is accounted for in two parameters: baseline coverage and risk of funging adjustment. At the moment, those parameters capture somewhat different considerations for Nigeria than for other countries. We plan to align our approaches in the future.

Outside Nigeria

Baseline coverage accounts for the chance that children would receive vitamin A supplementation through routine visits to health facilities. We base our estimates on three sources:28

• Helen Keller's best guess (which we give 40% or 50% weight)
• GiveWell's prior based on the delivery method and our understanding of the strength of the health system (40% or 50% weight)
• If available, Helen Keller's coverage survey data on VAS received outside of campaigns (20% weight where available). .

In the countries supported by this grant, this proportion ranges from 15% (Niger) to 29% (Burkina Faso).

The funging adjustment accounts for the chance that, without our support, children would receive VAS through campaigns.29 This is based on a subjective guess, which relies on:

• (i) whether UNICEF, which is the other main implementer of VAS in the countries where we support Helen Keller, is likely to support VAS programs in these locations in the absence of Helen Keller's program,30 and
• (ii) whether it’s likely there will be polio or measles campaigns with which VAS campaigns may be co-delivered. Due to the lower cost of co-delivered VAS campaigns, we think this increases the likelihood another funder would support VAS.

In the countries supported by this grant (other than Nigeria), this proportion ranges from 22% (in Cameroon) to 43% (in Burkina Faso).

Nigeria

In Nigeria, baseline coverage accounts for the chance that children would receive vitamin A supplementation through either routine coverage or lower-coverage campaigns. We employ recent data from New Incentives and Malaria Consortium, summarized here. Where available, we use state-specific data; otherwise, we use an average from the Helen Keller states we do have information about. We estimate an average baseline coverage rate of 44% in the Nigerian states supported by this grant.31

The funging adjustment accounts for the chance that high-coverage campaigns would have been funded by a different funder. In Nigeria, we estimate this ranges from 17% (in Adamawa State) to 22% (in Niger State).This is based on input from HKI and Nutrition International, and our understanding that there is a somewhat elevated likelihood of VAS being funded through polio campaigns supported by other actors in the near term.32

We are uncertain about both our baseline coverage estimates and funging adjustments (see below).

VAS's impact on child mortality

We estimate a baseline all-cause mortality rate ranging from ~0.8% to 1.4% in the locations we're funding.33 Our estimate of the effect size of VAS treatment on mortality is based on a meta-analysis of VAS trials,34 which we then adjust to account for internal and external validity concerns, including our best guess of current rates of vitamin A deficiency (VAD) prevalence in the locations where Helen Keller operates.35 After these adjustments, we estimate that receiving VAS treatment reduces a child’s mortality risk by ~3% to 11%, varying by location.36 You can read more about our approach to estimating the impact of VAS on mortality in our intervention report here.37

The effect of VAS on mortality is one of our key uncertainties about the CEA (more below).

3.2 Outside-the-model factors

We've seen significant swings in our cost-effectiveness estimates for VAS as we've updated our CEA over the past several years.38 During the course of this investigation, changes in our model that we feel uncertain about, such as updating our estimates of routine coverage in the absence of VAS campaigns, had a significant impact on our bottom line.39 Because of this, we put weight in our grant recommendation on outside-the-model qualitative factors for locations where we estimate cost-effectiveness within 2x of our 8x funding bar. This led us to recommend funding two locations, Burkina Faso and Adamawa State, Nigeria, that are slightly below our bar. We have not yet made a funding decision on whether to renew support for Nasarawa State, which is also slightly below our bar. We also decided to maintain our decision to fund two locations, Cameroon and Niger State, Nigeria, that are slightly above our bar, as we did not see a compelling qualitative case that outweighed our cost-effectiveness estimates.40

Burkina Faso

In Burkina Faso, we feel there is a strong qualitative case to continue funding despite our cost-effectiveness estimate of 6.9x falling below our funding bar. The main reasons for this are:

• There is compelling recent evidence of high vitamin A deficiency (VAD) in Burkina Faso. For most other countries, we rely on outdated survey data for VAD, or, in the case of Nigeria, have seen recent survey data with conflicting results. In contrast, a 2020 national survey in Burkina Faso found high VAD prevalence across two measures of VAD.41
• We have open questions about several key parameters in our model. If we took different approaches then Burkina Faso would be above our bar.42 We plan on making progress on these questions in future investigations.
• The size of the grant in Burkina Faso is relatively modest ($1.4 million, 6% of the grant total). Given this, it seems reasonable to continue support rather than exiting and deciding to resume support at a later date.

However, we have not recently spoken with a government official in Burkina Faso to triangulate our understanding of the program's impact. We plan to do this in the future.

Adamawa State

We think there's a moderately strong case for renewing support in Adamawa State, which we model very close to our bar (7.8x). This is primarily based on the proximity to our bar, the small size of the grant ($1m, 4% of grant total), and the potential difficulty of exiting and restarting support later. However, we have not recently spoken with a government official in Adamawa State to triangulate our understanding of the program's impact.

Nasarawa State

In Nasarawa State (which we model at 6.7x), we don't have sufficient information about the qualitative case for the program there to make a decision at this time. This is because we model the program as a bit further from our bar and we haven't spoken with a government official in Nasarawa State to triangulate our understanding of the program's impact. We plan to do this prior to finalizing a separate decision on whether to provide renewal funding for Nasarawa State.

Expansion states

Outside-the-model considerations also led us not to recommend funding three proposed expansion states in Nigeria (Kaduna State, Sokoto State, and Kebbi State) that are well above our funding bar (10x+).43 Specifically, we did not feel we had a strong enough understanding of the role of different implementers in these states. We plan to consider funding for these states in the future, once we improve our understanding.

4. Reservations and uncertainties

4.1 Uncertainty about CEA parameters

We are highly uncertain about several key parameters in our CEA. We discuss our concerns about these parameters below:

• The effect of VAS on mortality. We are more uncertain about the effect of VAS on mortality than we are about GiveWell’s other top recommended programs. We discuss this in detail in our VAS intervention report.44
• Current vitamin A deficiency rates. These inform our best guess of the impact of VAS on mortality in each location in which Helen Keller operates.
  • Outside of Burkina Faso and Nigeria. In many countries, we rely on outdated or limited survey data on vitamin A deficiency (VAD) rates.45 We make adjustments based on proxies to extrapolate a best guess of current VAD rates, but we remain highly uncertain about VAD prevalence in the locations we're funding.46 GiveWell is currently considering funding VAD surveys to address this information gap.

    We also give some weight to estimates from the Institute for Health Metrics and Evaluation's Global Burden of Disease (GBD) study. We're very uncertain about the GBD estimates we rely on, and use an average of three years to account for this uncertainty.47

  • Burkina Faso. There is a nationally representative VAD survey from 2020.48 We give this survey 75% weight because it is recent and two VAD measures (serum retinol and modified relative dose-response, or MRDR) point in the same direction.49 We give GBD some weight (25%) because we think it's possible that GBD may be capturing additional inputs that are not reflected in the survey.
  • Nigeria. There is a nationally representative VAD survey from 2021.50 However, there is a significant discrepancy between two measures of VAD in Nigeria (serum retinol and MRDR) that we don't fully understand.51 While we adjust our estimate to reflect our best guess, we are also considering funding additional research, such as an MRDR validation trial that compares MRDR and serum retinol versus gold-standard retinol isotope dilution in Nigeria or new VAD surveys that include MRDR.
• Cost per supplement
  • The number of children targeted. We currently make a -10% adjustment to Helen Keller's reported target populations in all locations.52 This was based on a report on alternative target population data shared by IDinsight and a comparison of Helen Keller and Malaria Consortium's target population estimates.53 We've considered updating our approach to population estimates to use external sources, but have not yet investigated this fully.
  • Weight given to past spending data versus future budgets. We currently give past spending data 75% weight and future budgets 25% weight.54 However, Helen Keller has told us that future budgets are poor predictors of its future spending, due to changes in current exchange rates, fuel costs, inflation, partner contributions, and ad hoc opportunities to layer VAS campaigns with immunization campaigns.55 We may revisit this weighting going forward.
• Counterfactual coverage. This refers to the proportion of children that would receive VAS in the absence of our support for Helen Keller and the likelihood that other actors would fund the program in our absence.
  • The proportion of children that would receive VAS in the absence of our support. Our estimate is based on a weighted average of different sources (see above). We're very uncertain about how much weight to place on each of these inputs, and are also uncertain about the accuracy of GiveWell's priors and Helen Keller's best guess.
  • The likelihood that other actors would fund the program in our absence. This is based on a subjective guess, based on factors described above. We're fairly uncertain about these estimates, and we plan to try to better understand UNICEF's funding and priorities going forward.
• Potential adverse effects of receiving multiple doses of VAS. Many countries offer some routine VAS delivery outside of campaigns. We're concerned about the potential health risks of children receiving two doses in a very short period of time (e.g. 24 hours), based on a conversation we had with Sherry Tanumihardjo, a vitamin A expert at the University of Wisconsin-Madison,56 and previous research by GiveWell.57 This could occur if a campaign happens shortly after a child has received vitamin A at a routine health visit, or vice versa.

  We don't account for this in our cost-effectiveness model now, but this question is a high priority on our vitamin A research agenda going forward.

4.2 Uncertainty about our qualitative decision to fund two locations below our funding bar

We are recommending funding two locations that are below our 8x funding bar (Burkina Faso and Adamawa State, Nigeria) on the basis of qualitative reasoning. We discuss these subjective decisions in the “Outside-the-model factors” section above.

4.3 Uncertainty about VAS delivery model

Helen Keller has told us that while it currently supports VAS campaigns as the most cost-effective delivery model based on available evidence, it sees gathering evidence on routine delivery approaches as critical for ensuring sustainable VAS delivery given uncertainties around future funding for campaigns.58 We understand "routine delivery" to typically refer to VAS received through health facilities as part of routine health care—for example, when a child goes to a health facility to receive immunizations—in contrast to mass campaigns. Helen Keller told us they have been focused on campaigns in part because of GiveWell's preference.59

The funding gaps Helen Keller pitched us as part of this grant are generally to: a) renew our funding for mass campaigns, or b) expand our support for mass campaigns in Nigeria. We think it's possible that we would want to support routine delivery rather than mass campaigns, though we haven't modeled this yet. Because of this, we might prefer to wait to make a grant until we:

• Have a stronger understanding of the current opportunities for and budgets for routine delivery support in these countries, and
• Have an organizational view of the cost-effectiveness of switching to routine delivery support.

We're recommending going ahead with this grant now because we think it will take some time to develop our understanding of the points above, and we'd prefer to maintain two-plus years of funding visibility for Helen Keller. We also think this is reasonable because this grant funding could be used to transition to routine coverage in future years.

However, we are uncertain about this decision because we're recommending funding based on our understanding of the cost to support campaigns (which could differ from the cost to support routine delivery), and it may be more difficult to change course in the future if we've already funded campaigns.

5. Plans for follow up

• We plan to receive data from Helen Keller annually on coverage, target populations, spending, and funding gaps in the locations where it operates.
• We plan to consider funding additional research that could influence our funding for Helen Keller, including a study to improve our understanding of VAS impact on mortality and VAD prevalence surveys.
• We plan to further research some of the key parameters in our model that we are uncertain about. This may include refining our target population estimates, investigating estimates of dietary vitamin A inadequacy rates to see if they can inform our VAD rate estimates, revisiting our funging estimates by looking into past grants and UNICEF's funding priorities, and improving our understanding of the risks of double-dosing.

6. Internal forecasts

7. Our process

• We had 10 conversations with Helen Keller and 12 conversations with other funders and implementers of VAS programs, government officials in the countries where Helen Keller operates. We also spoke with polio immunization campaign experts to refine our understanding of the likelihood that VAS campaigns could be paired with polio immunization campaigns.
• We updated our CEA and cost per supplement analysis with new program data and updated our approach to estimating several parameters and adjustments.

8. Sources

• 1

  See our intervention report on Vitamin A Supplementation for more details. Note, as of the time of publication, we are in the process of updating our VAS intervention report to reflect our latest methodology, and the details in our public intervention report from April 2024 may no longer accurately describe the methodology used in this grant investigation, although our overall impression of VAS has not changed.

• 2

  "We rank providing funding to our two new top charities, Helen Keller Intl (Helen Keller)’s VAS program and No Lean Season, next." The GiveWell Blog, "Our top charities for giving season 2017," November 2017

• 3

  See GiveWell, “Helen Keller Intl — Vitamin A Supplementation (January 2023).”

• 4

  We previously informed Helen Keller that our August 2024 grant would serve as exit funding for any locations we decided to discontinue funding as part of this grant investigation.

• 5

  See here for a copy of our cost-effectiveness analysis where we've calculated Guinea-specific estimates for this grant.

• 6
  • “Mass campaigns involve large-scale distribution of VAS to households in a short period of time. In these campaigns, VAS is often co-delivered with other public health interventions by community health workers, including deworming, polio vaccination, "mop-up" immunizations (for children who have missed scheduled immunizations), and screening for severe acute malnutrition and moderate acute malnutrition. The campaigns we have seen occur twice a year. GiveWell, “Helen Keller International.”
  • Vitamin A supplementation (VAS) involves getting 6-59 month old children to swallow a small amount of fluid containing vitamin A, delivered via a single-use capsule. GiveWell, “Vitamin A Supplementation.”

• 7

  Helen Keller’s VAS delivery models include:

  • Door-to-door campaigns: These typically occur twice per year in May and October. These campaigns usually take place over 5 days to 2 weeks. During campaigns, community health workers go door-to-door to provide VAS at home. They often also provide other services such as deworming and malnutrition screening.
    • "This model involves the rapid delivery of vitamin A Supplements to children over a short period, typically lasting 5 days to 2 weeks, usually in May and October each year." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 6
    • "Campaigns often include additional services like deworming, acute malnutrition screening, promotion of health and nutrition best practices, and even birth registration to enhance cost-effectiveness." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 6
  • Fixed site + outreach: These are similar to door-to-door campaigns, in that they occur over a short period of time, typically twice per year. The key difference is that VAS is offered in health facilities and "outreach stations," rather than going door-to-door.
    • "Distribution occurs through either door-to-door approaches or fixed sites such as health facilities and outreach stations in remote locations." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 6
  • Routine delivery: VAS is offered on an ongoing basis through health clinics. Twice per year, mop-up campaigns are held to reach any children who did not receive VAS through routine activities.
    • "... combination of routine activities with mop-up campaigns at the end of each semester in DRC, Sierra Leone and part of Côte d’Ivoire. In these three countries, Helen Keller actively supports the transition towards full integration of VAS into facility-based routine service delivery. This often involves establishing a network of community-based volunteers to inform communities about opportunities and to support distribution activities in remote areas." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 6
    • "Helen Keller is actively supporting countries to transition from campaign-based VAS delivery to routine service models. Routine delivery is more sustainable in the long term and better integrated within existing health systems. In countries like Kenya, Senegal, and Côte d’Ivoire, Helen Keller has begun working with Ministries of Health to scale up routine VAS services, with some regions already seeing promising results." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2024, p. 5.

  See Table 1, Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2024, p. 7 for a breakdown of delivery model by countries supported by GiveWell.

• 8
  • "Please find below a table showing the spending for FY24 with addition of the spending that spilled over between July and October due to campaign delays. It shows that we are anticipating an underspending of around USD M.2.1on the FY24, which can be rolled over for the coming fiscal year. Considering this adjustment, it would be reasonable to deduct this amount from the potential FY27 donation." Correspondence from Helen Keller to GiveWell, November 7, 2024
  • In September 2024, we learned that Helen Keller received a donation of 3.5m GBP (4.6m USD) from Founders Pledge. This donation is restricted to Burkina Faso, Cameroon, Cote d’Ivoire, DRC, Guinea, Mali, and Niger. In order to account for this, we proportionally reduced the size of our funding in the countries supported by the donation. See here in our supplementary calculations. Source: Correspondence from Helen Keller Intl to GiveWell, September 14, 2024 (unpublished).

• 9
  • One-year renewals:
    • Côte d'Ivoire ($4.7m) - 5.7x
    • Benue State ($0.9m) - 3.2x
    • Ebonyi State ($0.9m) - 3.4x
    • Nasarawa State ($0.7m) - 6.7x. We're still considering whether to provide additional funding to Nasarawa; details here.
  • Expansion states in Nigeria (two-year budgets):
    • Plateau ($1.3m) - 3.6x
    • Abuja FCT ($0.7m) - 2.1x

• 10

  To date, GiveWell has used GiveDirectly's unconditional cash transfers as a benchmark for comparing the cost-effectiveness of different funding opportunities, which we describe in multiples of "cash” (more). In 2024, we re-evaluated the cost effectiveness of direct cash transfers as implemented by GiveDirectly and we now estimate that their cash program is 3 to 4 times more cost-effective than we’d previously estimated. (more)

  • For the time being, we continue to use our estimate of the effectiveness of unconditional cash transfers prior to the update to preserve our ability to compare across programs, while we reevaluate the benchmark we want to use to measure and communicate cost-effectiveness.
  • Note that a) our cost-effectiveness analyses are simplified models that are highly uncertain, and b) our cost-effectiveness threshold for directing funding to particular programs changes periodically. See GiveWell’s Cost-Effectiveness Analyses webpage for more information about how we use cost-effectiveness estimates in our grantmaking.

• 11

  In this copy of our cost-effectiveness analysis, we input the grant amount for each location supported by this grant, with management costs distributed proportionally and Nigeria national support costs allocated proportionally to Adamawa and Taraba states (the funding for Niger State already includes support costs). See the grant budget above. Using these amounts, we estimate the total lives saved by this funding that wouldn't have been saved otherwise here.

• 12

  Note: as of the time of publication, we are in the process of updating our VAS intervention report to reflect our latest methodology, and the details in our public intervention report from April 2024 may no longer accurately describe the methodology used in this grant investigation, although our overall impression of VAS has not changed.

• 13

  We make several assumptions in projecting these future cost estimates, including that coverage will remain the same in each geography and that other actors' costs will increase by the same amount as Helen Keller's costs. See this sheet in our cost per supplement analysis for how we've estimated forward-looking costs, and this row for how we incorporate them into our overall estimates.

• 14

  Correspondence from Helen Keller Intl to GiveWell, October 31, 2024. (unpublished)

• 15
  • GiveWell in conversation with Helen Keller Intl, October 16, 2024. (unpublished)
  • Correspondence from Helen Keller Intl to GiveWell, September 2024. (unpublished)

• 16

  GiveWell in conversation with Helen Keller Intl, October 16, 2024. (unpublished)

• 17

  Helen Keller plays an important role in the financial planning and resource allocation process for vitamin A supplementation.

  • "Budgeting support: Using the microplans developed, Helen Keller works with government teams to develop activities budgets. These budgets include specific areas to be supported by Helen Keller and, in some cases, cover the entire country. The resulting budget, along with the detailed information from the microplans, form the basis for sub-agreements signed with government entities.
  • "These sub-agreements define in detail the roles and responsibilities of Helen Keller and the governments in the implementation of VAS activities and are mainly used for campaigns. Sub-agreement funds for government activities and staff represent approximately 70% of the total implementation costs and are established with local administrative entities. They describe the legal obligations of the government and all aspects of campaign implementation such as staff training, social mobilization, distribution, and data management. Sub-agreements also describe payment terms, milestones and deliverables associated with each payment.
  • "Helen Keller also provides direct financial support to other activities such as micro-planning workshops, supervision by national ministry teams, communication activities and production of materials, independent monitoring of the campaign and coverage surveys. The proportion of funding through sub-agreements and through direct funding may varies between countries" Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 9

• 18

  Our understanding that insufficient planning leads to lower coverage is largely based on information received by Helen Keller, which we have not triangulated with other sources. Reasons cited by Helen Keller include:

  • Health workers may not receive training in time for a campaign. GiveWell in conversation with Helen Keller Intl, October 9, 2024. (unpublished)
  • Communications may not be rolled out, so caregivers are not aware of the campaign;
    • "The success of VAS activities depends heavily on effective communication strategies and channels to create widespread awareness among the population on the importance of VAS and the organization of an upcoming campaign. Coverage surveys have shown that one of the main reasons for children not to receive VAS is that caregivers were not aware of the campaign taking place around them. Helen Keller supports the development of communication plans and communication tools and the roll out of the communication campaign." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 10
  • VAS capsules may not be procured in time (leading to inadequate supplies). Source: GiveWell in conversation with Helen Keller Intl, October 9, 2024. (unpublished)

• 19
  • GiveWell in conversation with Helen Keller, October 9, 2024. (unpublished)
  • "The Ministry of Health leads the national coordination body, which includes key partners involved in VAS delivery, such as UNICEF, Helen Keller, and Nutrition Intl." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2024, p. 8

• 20

  Helen Keller sees its comparative advantage among the three main implementers of VAS (MOH, UNICEF, and Helen Keller), each of whom jointly participates in planning for VAS at the national level, as having a strong technical understanding of how the program is implemented at the local level and how to organize and analyze coverage surveys to improve future implementation. Source: GiveWell in conversation with Helen Keller Intl, October 9, 2024. (unpublished). This was partly triangulated in conversation with an additional stakeholder.

• 21

  A government official in Taraba State, Nigeria, told us that supervision started being done when Helen Keller started supporting VAS, implying that there was insufficient funding and/or interest prior to Helen Keller's involvement. They also indicated the importance of supportive supervision to increasing the program's reach. A government official in Cameroon told us that in the absence of additional support, they only have the funding to weakly carry out supervision activities. We have not triangulated these claims against other sources (for example, asking to see records of supervision teams deployed or government spending on supervision in the absence of Helen Keller).

• 22

  "Distribution of supervisors by district/region and development of supervision plan. Helen Keller supports the Ministry of Health in developing a supervision plan that ensures each supervisor oversees at least two distribution teams daily, focusing on areas like hygiene standards, communication, and accurate tally sheet completion. Ideally, teams should not be informed in advance about a supervisor visit… During supervision, the teams work closely with government counterparts to ensure adherence to all aspects of implementation protocols. Every evening, supervisors report to the district Ministry of Health addressing any implementation issues encountered and the actions taken to address them." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2024, p. 13-14

• 23

  "Cost Structure: The majority of campaign expenses were attributed to personnel costs, with 75% of the total budget going to staffing. Distribution was the largest cost category, accounting for over 40% of the budget, while supervision and logistics were other significant cost drivers. Efforts were made to optimize resources, such as streamlining training and supervision efforts, particularly in countries where recurring staff are employed across campaign rounds." Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 4-5

• 24

  GiveWell in conversation with Helen Keller Intl, October 9, 2024. (unpublished)

• 25

  "Supervision of distribution activities is a joint endeavor by supervisors from Ministries of Health (central, regional, district, and health facility levels) and Helen Keller teams. During campaign implementation, supervisors follow predefined circuits with districts and health facilities to monitor the campaign implementation. Daily synthesis meetings are conducted at each level during the campaign to assess its strengths and weaknesses, identify areas for improvement, and address any urgent challenge" Helen Keller Intl, “GiveWell Annual Narrative Report,” October 2023, p. 11

• 26
  • Justina Bitrus, State Nutrition Officer, Taraba State, Nigeria
  • Mamman Bello Kakulu, Director, Primary health care department, Primary Health Care Development Agency, Taraba State, Nigeria
  • Henry Masari, Assistant State Nutrition Officer, Taraba State, Nigeria, October 11, 2024
  • Dr. Hassan Ben bachire, Directeur du Département de Promotion de la Santé, Cameroon
  • Dr. Facely Camara, Directeur Adjoint de la Santé Familiale et de la Nutrition, Guinea

  These conversations are unpublished.

• 27

  See this row in our CEA

• 28

  See GiveWell, Updates to counterfactual coverage of VAS, 2024 for detail.

• 29

  See this row in our CEA for our funging adjustments by country.

  • Our funging adjustment consists of the likelihood that domestic governments and other philanthropic actors would support programs in our absence.
    • For the likelihood of domestic government funging, we use a rough guess of 10% for the countries supported by this grant, with the exception of Burkina Faso, where we make a higher guess of 25% based on our understanding that the government there provides support for VAS via a World Bank loan. This understanding is based on conversations with Helen Keller and Helen Keller's reported government spending in Burkina Faso compared to other countries. See the spending by MoH column for Burkina Faso in our cost per supplement analysis here. We make a lower guess of 5% for Nigeria, because we've already accounted for non-Helen Keller campaigns in our counterfactual coverage estimate (see paragraph above).
    • For the likelihood of funging by other philanthropic actors, we base our estimates on our understanding of UNICEF's prioritization (UNICEF is the main other implementer of VAS in the countries where Helen Keller operates), as well as the likelihood of vaccination campaigns for polio and measles, because VAS campaigns are often layered onto these vaccination campaigns.

• 30

  We rely on three proxies for assessing whether a country is a UNICEF priority: (1) UNICEF's spending per child in each country, (2) whether Helen Keller told us the country is a UNICEF priority, and (3) how many regions UNICEF covers, as reported by Helen Keller.

• 31

  See here for our state-specific estimates of counterfactual coverage for Nigeria. For states where we do not have state-specific estimates, we use an average of the states for which we have data. See this row in our CEA. For the states supported by this grant, we use an estimate of 40% for Niger State, and the average value of 46% for Adamawa and Taraba States. For the value reported here, we average 46%, 46%, and 40% = 44%.

• 32

  See this cell note for details.

• 33

  Our estimates of all-cause mortality are based on:

  • Mortality estimates from the Global Burden of Disease (GBD) Model, UN IGME, and (where available) households surveys (MICS/DHS). See this row of our CEA.
  • The proportion of people under age 5 represented in GBD 2021 disease burden estimates who received VAS, based on an analysis of VAS coverage compiled by Rethink Priorities.
  • The treatment effects of VAS. We estimate the treatment effect of VAS in program contexts by comparing the share of VAS-preventable deaths when the VAS trials we rely on were conducted to the share of VAS-preventable deaths today.

  You can see our calculations in the CEA here.

• 34

  See this row in our CEA, and this document for more on our meta-analysis.

• 35

  See these rows in our cost-effectiveness analysis. The adjustments we make are for internal validity, external validity, and for the difference in supplementation frequency between the trials we use in our meta-analysis and current contexts.

• 36

  This is based on Imdad et al. 2017, a meta analysis finding a 24% reduction in mortality using a random-effects estimate. We conducted our own meta analysis using only the trials intended to measure all-cause mortality to address concerns about potential publication bias, and found a 19% reduction in mortality. For more detail, see the cell note here.

• 37

  Note: as of the time of publication, we are in the process of updating our VAS intervention report to reflect our latest methodology, and the details in our public intervention report from April 2024 may no longer accurately describe the specific methodology used in this grant investigation. See this section of our cost-effectiveness analysis for how we currently calculate the effect of VAS on mortality.

• 38

  See the “Version compare (2021-2024)” sheet of GiveWell, “Helen Keller Intl lookback CEA comparison” for details on changes in our VAS cost effectiveness estimates.

• 39

  For example, the cost-effectiveness of Cameroon went from 7.3x (below the bar) to 8.8x (above the bar) after we updated our routine coverage estimates. Source: GiveWell, Internal changelog of VAS CEA. (unpublished)

• 40
  • Cameroon (8.1x): We see the qualitative case for Cameroon as ~neutral, in which case we defaulted to the cost-effectiveness estimate being above the bar to make our decision.
    • Pros of funding
      • We spoke with a government official in Cameroon this summer and received a positive impression of their view of the value of Helen Keller support. (Unpublished)
  • Cons of funding
    • This is a large investment ($5.1m, 22% of the grant total); the cost to GiveWell of "getting it wrong" is high.
    • Cameroon is one of the locations in which our best guess is that Helen Keller would prefer to switch to supporting routine delivery, and it's possible we should hold off on recommending a grant focused on supporting campaigns in Cameroon. Source: GiveWell in conversation with Helen Keller, November 14, 2024. (unpublished)
  • Niger State (9.4x): We think there's a weakly positive outside-the-model case for funding Niger State.
    • Pros of funding
      • This grant represents a relatively small investment ($900,000, 4% of the grant total) that allows us to learn about Helen Keller's ability to scale VAS in a new state.
      • We spoke with a government official in Niger State, who was enthusiastic about receiving additional support for their VAS program. Source: GiveWell's conversation with Pharm Jiya Daniel Tsado (B.Pharm, MPH), Director Food and Drugs, Ministry of Primary Health Care, Niger State, November 18, 2024 (unpublished)
    • Cons of funding
      • We're proposing to fund Niger State for three years, so our commitment is longer than other locations if we learn something new that negatively updates our view on the cost-effectiveness of the program.

  • 41
    • See here for a summary of the 2020 Burkina Faso national survey results
    • "Vitamin A deficiency measured by low serum retinol (<0.70 µmol/L) and corrected for inflammation was 50% among children 6-59 months. The modified relative dose response (MRDR) test is a qualitative assessment of vitamin A liver stores and was conducted in a randomly selected subsample. A total of 37% of children were vitamin A deficient with a MRDR ratio ≥0.060 (retinol adjusted for inflammation)." Burkina Faso National Micronutrient Survey 2020, p. 27

  • 42

    For example:

    • We currently apply a -10% adjustment to reported target populations in all locations to account for potential overestimation, but based on an early-stage internal analysis of different data sources we may reduce this to -5% in the future. This would increase our cost-effectiveness estimate.
    • We currently put 75% weight on past costs and 25% weight on projected future costs in our cost-per-supplement analysis (see here), although Helen Keller has told us past costs are more reliable than future projections. Reducing or eliminating the weight we place on projected future costs would increase our cost-effectiveness estimate for Burkina Faso.
      • "Historically, GiveWell has updated the cost per supplement for vitamin A supplementation (VAS) annually, using actual program data from the previous year. Recently, however, there has been a shift towards using forecasted budgets and population estimates rather than real program population and expenditure data in calculating this metric. The rationale behind this transition remains unclear, especially given that forecasted budgets often differ from actual costs incurred. Budget estimates, which serve as the numerator in this calculation, are inherently approximations and may not align with the actual program expenditures, which can vary significantly due to various factors (see Table 1)" Helen Keller Intl, Helen Keller Intl, Cost-per-Supplement Queries and Recommendations, p. 6
    • We currently put 75% weight on the most recent VAD survey in Burkina Faso. Because this survey is recent and the different measures of VAD in the survey are aligned, it's possible we should give it 100% weight. This would increase our cost-effectiveness estimate for Burkina Faso.
    • If we reduced our population adjustment to -5%, put full weight on past cost data, and put full weight on the most recent VAD survey, our cost-effectiveness estimate for Burkina Faso would be above our bar at 8.2x. Source: GiveWell, Internal analysis (unpublished).

  • 43

    See here in our CEA for our cost-effectiveness estimates for Kaduna, Sokoto, and Kebbi

  • 44

    Note: as of the time of publication, we are in the process of updating our VAS intervention report to reflect our latest methodology, and the details in our public intervention report from April 2024 may no longer accurately describe the specific methodology used in this grant investigation, although our overall impression of VAS and relative uncertainty remains unchanged.

  • 45

    See our summary of VAD surveys in this spreadsheet. See this section of our cost-effectiveness analysis for our most recent approach to estimating VAD.

  • 46

    See the “Changes in vitamin A deficiency” section of our intervention report on VAS. Note: as of the time of publication, we are in the process of updating our VAS intervention report to reflect our latest methodology, and the details in our public intervention report from April 2024 may no longer accurately describe the specific methodology used in this grant investigation. See this section of our cost-effectiveness analysis for our most recent approach to estimating VAD.

  • 47

    See this section of our CEA for more detail.

  • 48

    See GiveWell's analysis of recent vitamin A deficiency surveys, 2024 v2 for detail.

  • 49

    We give the Burkina Faso survey extra weight because it is significantly more recent (2020) than the other nationally representative surveys we rely on, and because there's a reasonable degree of concordance between its two measures of vitamin A deficiency (VAD) (serum retinol and MRDR, at 50.2% and 36.9%, respectively). (We don't use MRDR in our CEA estimate, but we see this agreement between the two measures as a signal that there is more likely to be true high VAD rates in Burkina Faso.) See table "Vitamin A Deficiency by Retinol and MRDR Among Children 6-59 Months," p. 8

  • 50

    2021 Nigeria National Food Consumption and Micronutrient Survey

  • 51

    We discuss these discrepancies in results in GiveWell, Update to Nigeria VAD rate estimate, October
    2024

  • 52

    See this row and the associated cell note for detail.

  • 53
    • IDinsight, "Estimating Populations, IDinsight’s summary & recommendations to GiveWell," 7 February 2023 (unpublished)
    • The GiveWell grantmaking team focused on malaria makes a -5% adjustment to the target populations reported by Malaria Consortium, which we use as a rough benchmark for comparing potential overreporting of target populations by Helen Keller. This is based on a rough, internal analysis (unpublished).

  • 54

    See this row and the associated cell note for calculations and detail.

  • 55

    "Forecasted Costs and Targeted vs. Actual Beneficiaries and Expenditures…Historically, GiveWell has updated the cost per supplement for vitamin A supplementation (VAS) annually, using actual program data from the previous year. Recently, however, there has been a shift towards using forecasted budgets and population estimates rather than real program population and expenditure data in calculating this metric. The rationale behind this transition remains unclear, especially given that forecasted budgets often differ from actual costs incurred. Budget estimates, which serve as the numerator in this calculation, are inherently approximations and may not align with the actual program expenditures, which can vary significantly due to various factors (see Table 1)... Significant fluctuations in currency exchange rates, rising fuel costs, and inflation have led to notable discrepancies between forecasted and real expenditures over the last 24 months. Additionally, program-specific factors, such as the need for Helen Keller to adjust regional support based on funding from other partners, play a role. In Cameroon, Côte d'Ivoire, and Guinea, for example, Helen Keller expanded support due to reduced contributions from other partners, increasing the direct campaign support costs while keeping certain fixed expenditures constant, thereby lowering the achieved cost per supplement… Integrating VAS with other health campaigns (e.g., polio vaccination) can reduce total program costs by sharing resources, logistics, and personnel costs, frequently resulting in actual expenditures that are lower than budget estimates." Helen Keller Intl, Cost-per-Supplement Queries and Recommendations, p. 7

  • 56

    GiveWell in conversation with Sherry Tanumihardjo, Professor of Nutritional Sciences, University of Wisconsin-Madison, June 26, 2024 (unpublished). Key points:

    • Risks include diarrhea, nausea, vomiting and in extreme cases, mortality.
    • The risks of acute vitamin A toxicity dissipate after 24 hours. Children receiving two doses over the course of a few days is less risky than kids getting two doses in one day.
    • VAD may be low in Nigeria, based on a study of one biomarker (MRDR). This would imply increased risk of toxicity from double dosing, since children are less vitamin A deficient. (It is possible VAD is low because of VAS and the expert wouldn't recommend stopping the program on the basis of the finding).
    • Mitigation measures to prevent double-dosing are important, such as marking children's hands with indelible ink.

  • 57

    See the “Factors we have excluded” section of our intervention report on VAS.

  • 58
    • Our guess is that the most likely places where Helen Keller would want to reorient its support toward routine delivery are Cameroon, Guinea, and Cote d'Ivoire, based on a conversation we had with Helen Keller. (In Côte d'Ivoire, Helen Keller is already supporting routine delivery in ~half the country. We're recommending exiting in Cote d'Ivoire, but may consider an additional exit grant to support this transition.)

    The conversation that informed our understanding was focused on the locations in our CEA that were closest to or below our bar, so we didn't discuss the prospects for routine delivery in DRC, Mali, and Niger, which are well above the bar. We're unsure how promising Helen Keller thinks transitioning to routine delivery in those locations would be. However, Helen Keller did note in its annual report that it is piloting routine delivery in DRC and analyzing approaches to switching in Guinea. Source: Givewell in conversation with Helen Keller Intl, November 14, 2024. (unpublished)

    • Helen Keller has told us that it does not necessarily support routine delivery models over mass campaigns, given the evidence on the cost-effectiveness of mass campaigns. But rather that they see a critical need to build evidence on routine approaches to VAS delivery to ensure VAS is sustained even if external support should end in the future. Source: Helen Keller Intl, comments on a draft of this page, August 2025 (unpublished)

  • 59

    Givewell in conversation with Helen Keller Intl, November 14, 2024. (unpublished)