Against Malaria Foundation – Support for LLIN Campaigns in Nigeria (2022), Uganda (2023), and Togo (2023)

Note: This page summarizes the rationale behind a GiveWell grant to the Against Malaria Foundation (AMF). AMF reviewed this page prior to publication.

Published: April 2022; Last Updated: November 2022

Planned activities and budget

The $52.8 million will enable AMF to purchase LLINs for campaigns in:

• A set of states in Nigeria, at a cost of $6.9 million.1 In Nigeria, most states are assigned to receive funding for malaria programs either from the Global Fund to Fight AIDS, Tuberculosis and Malaria ("the Global Fund"), from the President's Malaria Initiative (PMI), or through a loan from the World Bank and the Islamic Development Bank.2 We expect AMF to use this grant to support campaigns that are scheduled to occur in 2022 in states that are supported by the Global Fund.3 This is the first time that AMF will support LLIN campaigns in these states. We have heard from multiple sources that there is insufficient funding from the Global Fund for these campaigns because the portion of its current malaria allocation to Nigeria allocated to LLIN campaigns was based on an underestimate of the number of LLINs that would be needed to fully cover the population in these states.4 While in some cases the Global Fund will shift funding between countries or diseases in order to cover unfunded needs, it is not expected to be able to do so in time to fill the funding gap for 2022 campaigns.5 We expect AMF to use most of this grant to purchase a portion of the LLINs required for these campaigns6 and, consequently, for the Global Fund to purchase fewer LLINs and instead spend that funding on delivering the LLINs that AMF purchases. A smaller portion of the funding will be used to collect post-distribution monitoring data.
• Uganda, at a cost of $38.5 million.7 Uganda typically implements a nationwide LLIN campaign every three years; the next campaign is scheduled to occur in 2023. Historically, Uganda has not had enough funding from other funders (primarily the Global Fund) to fully cover the country's population. AMF therefore filled a similarly-sized funding gap for the country's 2017 and 2020 LLIN campaigns.8 We expect AMF to use most of this grant to purchase a portion of the LLINs required for the 2023 campaign9 and, consequently, for the Global Fund to purchase fewer LLINs and instead spend that funding on delivering the LLINs that AMF purchases. A smaller portion of the funding will be used to collect post-distribution monitoring data.
• Togo, at a cost of $7.4 million.10 Togo typically implements a nationwide LLIN campaign every three years; the next campaign is scheduled to occur in 2023. Historically, Togo has not had enough funding from other funders (primarily the Global Fund) to fully cover the country's population. AMF therefore filled a similarly-sized funding gap for the country's 2017 and 2020 LLIN campaigns.11 We expect AMF to use most of this grant to purchase a portion of the LLINs required for the 2023 campaign12 and, consequently, for the Global Fund to purchase fewer LLINs and instead spend that funding on delivering the LLINs that AMF purchases. A smaller portion of the funding will be used to collect post-distribution monitoring data.

We conduct "room for more funding" analysis to understand what portion of a grantee's ideal future budget it will be unable to support with the funding it has or should expect to have available. We may then choose to either make or recommend grants to support those unfunded activities. To determine the size of these grants, we updated our room for more funding analysis for AMF. In this section of our review of AMF's work, we describe how we typically analyze AMF's room for more funding; for this analysis, we deviated from our standard process in two ways:

• In our room for more funding analyses, we typically include reserved funding as funding available to support program activities.13 In this analysis, we chose to exclude the $7.5 million that AMF holds in reserves from our estimate of available funding.14 Our reason for doing so was that in the agreements AMF signs with national malaria programs, it commits to purchasing a certain number of nets, rather than allocating a certain amount of funding. This means that any fluctuations in net price will affect the total amount of funding it ultimately must allocate to a campaign. It therefore seems reasonable to hold some funding in reserve in order to respond to any increases in net price, especially given the substantial number of nets AMF is currently committed to purchase.15 We believe that this approach is consistent with the three-year funding runway we typically provide to other grantees, which enables them to use funding budgeted for future years to pay for any cost increases in an earlier year.
• In our room for more funding analyses, we typically project the amount of additional funding that grantees will receive over the next three years and then add this projected funding to the amount available to allocate to grantee's unfunded spending opportunities. In this analysis, we did not include any projected revenue in our estimate of the funding AMF will have available to support its work because we believe that AMF should commit funding for these campaigns as soon as possible (more below). (This means that one effect of the funding we have granted to AMF is to free up AMF's future revenue for support to other campaigns. See further discussion below.)

The case for the grants

• Cost-effectiveness. During these grant investigations, we used our existing cost-effectiveness model for LLIN campaigns and updated various parameters to match the specifics of each funding gap. We believe that each of these grants is cost-effective. (More)
• Funding landscape for LLINs. We investigated the funding landscape for LLINs in these countries and believe that these funding gaps are unlikely to be fully filled by other funders. We adjust our cost-effectiveness estimates to account for the extent to which we believe our funding may be crowding out funding that would otherwise have come from other sources. (More)
• AMF's track record in Uganda and Togo. AMF's track record in Uganda and Togo suggests that we can expect a reasonably high proportion of LLINs purchased for future campaigns in these locations to reach and be used by their intended recipients. (More)
• Time-sensitivity. Making these grants now will enable AMF to order LLINs as soon as possible, in order to avoid delaying campaigns. (More)

Cost-effectiveness

How we use cost-effectiveness estimates in our grantmaking

After assessing a potential grantee's room for more funding, we may then choose to investigate potential grants to support the spending opportunities that we do not expect to be funded with the grantee's available and expected funding, which we refer to as "funding gaps." The principles we follow in deciding whether or not to fill a funding gap are described on this page.

The first of those principles is to put significant weight on our cost-effectiveness estimates. We use GiveDirectly's unconditional cash transfers as a benchmark for comparing the cost-effectiveness of different funding gaps, which we describe in multiples of "cash." Thus, if we estimate that a funding gap is "10x cash," this means we estimate it to be ten times as cost-effective as unconditional cash transfers. As of this writing, we have typically funded opportunities that meet or exceed a relatively high bar: 8x cash, or eight (or more) times as cost-effective as GiveDirectly's unconditional cash transfers. We also consider funding opportunities that are between 5 and 8x cash.

Note that our cost-effectiveness analyses are simplified models that do not take into account a number of factors. There are limitations to this kind of cost-effectiveness analysis, and we believe that cost-effectiveness estimates such as these should not be taken literally due to the significant uncertainty around them. We provide these estimates (a) for comparative purposes to other grants we have made or considered making, and (b) because working on them helps us ensure that we are thinking through as many of the relevant issues as possible. Our process for estimating cost-effectiveness focuses on determining whether a program is cost-effective enough that it is above our bar to consider funding; it isn't primarily intended to differentiate between values that are above that threshold.

Nigeria

We estimate that the cost-effectiveness of this grant is 16x cash.16 To generate this estimate, we used our existing cost-effectiveness model for LLIN campaigns and updated various parameters to match the specifics of this funding gap. The key parameter updates we made (or considered making) during this grant investigation include:

• Efficacy reduction due to insecticide resistance. Our cost-effectiveness model for LLINs includes an adjustment to account for the efficacy lost due to resistance within mosquito populations to the insecticides used in LLINs (see our separate report on this topic for more details). We updated this adjustment to (a) incorporate more recent insecticide resistance data and (b) incorporate evidence on the efficacy of LLINs that have been treated with piperonyl butoxide (PBO) in addition to standard insecticide. See full details in changes 3 and 18 in version 1 of our 2022 cost-effectiveness model changelog.
• Malaria-attributable deaths averted per 1,000 children per year targeted with seasonal malaria chemoprevention. We added an adjustment to account for deaths averted by seasonal malaria chemoprevention (SMC) in areas reached by both SMC and LLIN campaigns, which reduced our estimate of the deaths averted by LLINs. We believe this adjustment is needed because we are using estimates of deaths as of 2019, and SMC has been scaled up in some of the states where AMF-funded nets will be used since 2019. See full details in change 15 in version 1 of our 2022 cost-effectiveness model changelog.
• Malaria burden (multiple parameters). The impact of LLINs is moderated by rates of malaria incidence and mortality where they are distributed; we expect campaigns that target populations affected by higher malaria rates to have higher impact. We use data on malaria incidence and mortality in the countries where AMF works in our cost-effectiveness model to estimate the impact of LLINs on malaria rates in these countries. We considered updating our existing cost-effectiveness model for LLINs in Nigeria to use state-specific malaria burden data from Global Fund-supported states, rather than national data. After determining that these two data sets showed similar malaria burden, we decided not to make this update.
• Likelihood that the Global Fund would replace philanthropic costs. See details below.

Uganda

We estimate that the cost-effectiveness of this grant is 15x cash.17 To generate this estimate, we used our existing cost-effectiveness model for LLIN campaigns and updated various parameters to match the specifics of this funding gap. The key parameter updates we made during this grant investigation include:

• Cost per LLIN. Based on information received from AMF about its net procurement plan for Uganda, we increased our estimate of the proportion of PBO nets (which are more expensive than standard nets) that it will purchase to 100%. See full details in change 6 in version 1 of our 2022 cost-effectiveness model changelog.
• Efficacy reduction due to insecticide resistance. See change described above.
• Likelihood that the Global Fund and/or PMI would replace philanthropic costs. See details below.

Togo

We estimate that the cost-effectiveness of this grant is 8x cash.18 To generate this estimate, we used our existing cost-effectiveness model for LLIN campaigns and updated various parameters to match the specifics of this funding gap. The key parameter updates we made during this grant investigation include:

• Efficacy reduction due to insecticide resistance. See change described above.
• Malaria-attributable deaths averted per 1,000 children per year targeted with seasonal malaria chemoprevention. We added an adjustment to account for deaths averted by seasonal malaria chemoprevention (SMC) in areas reached by both SMC and LLIN campaigns, which reduced our estimate of the deaths averted by LLINs. We believe this adjustment is needed because we are using estimates of deaths as of 2019, and SMC has been or will be scaled up in Togo since 2019. See full details in change 15 in version 1 of our 2022 cost-effectiveness model changelog.
• Likelihood that the Global Fund would replace philanthropic costs. See details below.

Funding landscape for LLINs

We investigated the funding landscape for LLINs in Nigeria, Togo, and Uganda by having conversations with national malaria programs and other major LLIN funders (see below), reviewing published data on malaria funding from other major LLIN funders, and reviewing spreadsheets (which AMF asked national malaria programs in Togo and Uganda to complete) that compare each country's malaria funding during the Global Fund's 2018-20 and 2021-23 grant cycles.19 As a result of these investigations, we believe that these funding gaps are unlikely to be fully filled by other funders, but that some of our funding may be crowding out funding that would otherwise have come from these funders.

We adjust our cost-effectiveness estimates to account for the extent to which we believe our funding may be crowding out funding that would otherwise have come from other sources (in the case of LLINs, this is typically the Global Fund and/or PMI). Specifically, these adjustments represent the proportion of a grantee's funding that we believe may crowd out funding from other sources (for example, if we use an adjustment of 25%, we believe that 25 cents of every 1 dollar spent by the grantee would otherwise have come from other sources). See more details in this blog post. We have incorporated considerations about the funding landscape for LLINs in Nigeria, Togo, and Uganda into location-specific adjustments, which are accounted for in the cost-effectiveness estimates listed in the previous section. We describe these considerations below.

Note that in the sections below, we describe reasons why for each grant we believe that some portion of our funding may be crowding out funding that would otherwise have come from other sources. The case for these grants rests not on a belief that they will result in minimal crowding out, but rather on the belief that after thorough investigation, we are sufficiently well-informed about the funding landscapes in these countries to not be missing important considerations from our location-specific cost-effectiveness adjustments and on the fact that after incorporating these adjustments, our estimates of the cost-effectiveness of these grants exceed our bar for grantmaking.

Nigeria

In Nigeria, most states are assigned to receive funding for malaria programs either from the Global Fund, PMI, or through a loan from the World Bank and the Islamic Development Bank.20 We have heard from multiple sources that there is insufficient funding for the campaigns scheduled to occur in 2022 and 2023 in Global Fund-supported states because the portion of its current malaria allocation to Nigeria allocated to LLIN campaigns was based on an underestimate of the number of LLINs that would be needed to fully cover the population in these states.

We think it is unlikely that another funder will fill the gap for 2022 campaigns. While in some cases the Global Fund will shift funding between countries or diseases in order to cover unfunded needs, it is not expected to be able to do so in time to fill the funding gap for 2022 campaigns. (A similar gap exists for campaigns that are scheduled to occur in 2023 in states that are supported by the Global Fund. We do not yet know whether the Global Fund will decide to reallocate funding to fill that gap.) We do not believe that this gap will be filled by either of the other major funding sources for LLINs in Nigeria: we are aware that PMI has insufficient funding for LLIN campaigns in the states it supports (GiveWell made a previous grant to AMF to support campaigns in two states that are supported by PMI), and the World Bank and Islamic Development Bank loan funding is restricted to other states.21 We have chosen to use a value of 20% for the proportion of AMF's funding that we believe may crowd out funding from the Global Fund.

Uganda

AMF filled a similarly-sized funding gap for Uganda's 2017 and 2020 LLIN campaigns.22 Prior to this grant investigation, we used a value of 40% for the proportion of AMF's funding that we believe may crowd out funding from the Global Fund and/or PMI. This document explains our reasoning for that value.

In investigating this grant, we learned that between the 2018-20 and 2021-23 grant cycles, the Global Fund's malaria allocation increased by a larger percentage than its allocation to the LLIN campaign,23 with much of the funding increase going to other malaria programs and health systems strengthening.24 We think it is possible that if AMF did not have a history of providing support to Uganda's LLIN campaign—and thus that support from a non-Global Fund partner for the 2023 campaign was less likely—the Ugandan government may have chosen to allocate less to these other programs and more to the LLIN campaign.

We also think it is possible that if AMF did not fill this funding gap, the Global Fund might do so by reallocating funding within its broader portfolio. Our understanding is that the first round of such reallocations will occur in early 2022 and that funding gaps for 2023 LLIN campaigns are likely to be a top priority. We could have waited to see whether the Global Fund would reallocate funding to fill this gap before making this grant. However, we decided that the benefit of AMF ordering nets as soon as possible (see below) outweighed the benefit of waiting for more information. We have chosen to retain the value of 40% for the proportion of AMF's funding that we believe may crowd out funding from the Global Fund.

Togo

AMF filled a similarly-sized funding gap for Togo's 2017 and 2020 LLIN campaigns.25 Prior to this grant investigation, we used a value of 60% for the proportion of AMF's funding that we believed may crowd out funding from the Global Fund (PMI does not fund LLIN campaigns in Togo26 ). This document explains our reasoning for that value. A key consideration informing that value was the fact that between the 2018-20 and 2021-23 grant cycles, the Global Fund's malaria allocation to Togo grew substantially.27 We therefore believed that Togo would have more funding available for its 2023 LLIN campaign and would not need AMF to fill a gap of the same size.

In investigating this grant, however, we learned that $26 million (roughly 45%) of the 2021-23 malaria allocation is being spent on health systems strengthening,28 rather than programs specific to malaria. This is up from $7.5 million of the Global Fund malaria allocation spent on health systems strengthening in the 2018-20 grant cycle.29 We think it is possible that if AMF did not have a history of providing support to Togo's LLIN campaign—and thus that support from a non-Global Fund partner for the 2023 campaign was less likely—the Togolese government may have chosen to allocate less to health systems strengthening and more to the LLIN campaign. Our understanding, however, is that this funding is being spent on projects that the Togolese government considers high-priority and that have been planned for some time, and is therefore highly unlikely to be reallocated at this point in the grant cycle.

We also think it is possible that if AMF did not fill this funding gap, the Global Fund might do so by reallocating funding within its broader portfolio. Our understanding is that the first round of such reallocations will occur in early 2022 and that funding gaps for 2023 LLIN campaigns are likely to be a top priority. We could have waited to see whether the Global Fund would reallocate funding to fill this gap before making this grant. However, we decided that the benefit of AMF ordering nets as soon as possible (see below) outweighed the benefit of waiting for more information. We have chosen to retain the value of 60% for the proportion of AMF's funding that we believe may crowd out funding from the Global Fund.

AMF's track record in Uganda and Togo

AMF's track record in Uganda and Togo suggests that we can expect a reasonably high proportion of LLINs purchased for future campaigns in these locations to reach their intended recipients. AMF purchased LLINs for the 2017 and 2020 campaigns in both Uganda and Togo.30 As such, we can review results from past campaigns in these locations to understand the outcomes we should expect from future campaigns.

• Post-distribution monitoring. AMF contracts with independent organizations to conduct post-distribution monitoring, during which households in the areas targeted by a campaign are surveyed to assess the presence, usage, and condition of LLINs in those households. We have spent considerable time understanding the methodology used in these surveys and assessing the possible sources of bias in their results; overall, we believe that these surveys provide suggestive but imprecise evidence on the outcomes of the relevant LLIN campaign. Prior to these grant investigations, we had reviewed results from post-distribution monitoring conducted for the 2017 campaigns in Uganda and Togo. Full details of those surveys are in this spreadsheet and in this section of our review of AMF's work. Overall, we believe that results from these surveys provide meaningful evidence that a reasonably high proportion of LLINs purchased for these campaigns reached and were used by their intended recipients, and we believe that these results are likely to be indicative of what we can expect from future distributions in Uganda and Togo. At the time of these grant investigations, results from the 2020 campaigns in Uganda and Togo were expected but were not yet available.
• Registration and backcheck data. AMF collects data on the number of LLINs households are registered to receive during campaigns and the number of people living within those households ("registration data"). It also requires that 5% of these households be revisited so that this data can be recollected ("backcheck data") and compared against the original registration data. Because post-distribution monitoring results were not yet available from the 2020 campaigns in Uganda and Togo at the time of these grant investigations, we decided to review registration31 and backcheck32 data from those campaigns (see footnotes for the key findings from our review). Overall, we did not find any major cause for concern in these data, which slightly increases our confidence that there were not major issues during the 2020 campaigns. We note, however, that we have spent considerably less time understanding the methodology used in collecting registration and backcheck data and assessing the possible sources of bias within these data sets. We therefore view this as weaker evidence on the outcomes of these campaigns than we would post-distribution monitoring results.

Time-sensitivity

Making these grants now will enable AMF to order LLINs as soon as possible, in order to avoid delaying campaigns. According to AMF, it can take over a year for LLINs to be ordered, manufactured, shipped to countries, and transported within countries. This means that AMF needs to commit funding to a campaign well in advance of when it is scheduled to occur. Recently, we have heard from multiple sources that LLIN manufacturing and shipping is experiencing longer-than-usual delays, which means that LLINs should ideally be ordered even further in advance of campaigns than they typically would. We decided to make these grants now to minimize the risk that these campaigns—all of which are scheduled to occur within the next two years—will be delayed.

Risks and reservations

• Risk of future crowding out. We think it is possible that by filling these funding gaps, we are setting the expectation that GiveWell funding will be available to fill future gaps for LLIN campaigns in these countries. This may lead national malaria programs and other funders to reduce the funding they direct to those campaigns. (More)
• Concerns about AMF as a partner. We have heard feedback that AMF can be difficult for national malaria programs and other partners to work with, relative to other implementing organizations. (More)
• Uncertainties in our cost-effectiveness models. We are uncertain about the values we use for several of the parameters included in our cost-effectiveness models. (More)
• Uncertain use of future revenue. We don't know how AMF will use the future revenue that is freed up by these grants. (More)
• Lack of post-distribution monitoring results from 2020 campaigns. Post-distribution monitoring results from the 2020 campaigns in Togo and Uganda were not yet available at the time of these grant investigations. (More)

Risk of future crowding out

The adjustments we describe above are primarily intended to account for the extent to which we believe our funding may be crowding out funding that would otherwise have come from other sources to fill these specific, immediate funding gaps. They may not fully account for the effect that our grantmaking may have on other funders' behavior over the long term.

We think it is possible that by filling these funding gaps, we are setting the expectation for national malaria programs and other funders that GiveWell funding will be available to fill future gaps (for example, for the 2026 campaigns in Uganda and Togo or for campaigns in Global Fund-supported Nigerian states in 2023 and beyond). This belief may, in turn, lead them to direct funding that they would have directed toward these future campaigns to other programs and services. In our conversations with national malaria programs and other funders, we have and will continue to communicate that to the extent possible, our goal is for the funding we direct to LLIN campaigns to add to the pool of funding available for those campaigns, rather than to replace funding that would otherwise have been in that pool.

Concerns about AMF as a partner

We have heard occasional feedback that AMF can be difficult for national malaria programs and other partners to work with, relative to other implementing organizations. We see this as a concern for two main reasons, both related to GiveWell's values. First, it could suggest that we are failing on our value of maximizing global wellbeing. If AMF is a lower-quality partner than other implementing organizations (for example, because it adds friction to the planning process that causes delays or otherwise lowers the quality of LLIN campaigns), this would reduce the impact that our grants to AMF can achieve. More broadly, if GiveWell's relationship with AMF as a grantee damages our relationships with other partners, this could reduce our organization-wide ability to make highly impactful grants (for example, by discouraging partners from sharing information with us about other funding gaps). Second, it could suggest that we are failing on our value of emphasizing considerateness, if we continue to support a grantee that creates challenges for other partners.

In investigating these grants, we had conversations with partners in each country, during which we explicitly asked for feedback on AMF as a partner. Broadly, we heard positive feedback about AMF that reduced our concern about the risks described above. We heard two main points of constructive feedback. First, the fact that AMF only pays for LLINs and does not fund any of the costs of delivering those nets can cause coordination challenges, because partners must determine how to fund those delivery costs, which can then lead to delays. Second, AMF's requirement that 5% of households registered to receive LLINs be resurveyed can cause challenges, because AMF does not cover the costs of this procedure and because national malaria programs may have existing procedures that then must be replaced with a new methodology. We plan to share this feedback with AMF and discuss potential ways to reduce the challenges caused by AMF's operating model and requirements.

Uncertainties in our cost-effectiveness models

We are uncertain about the values we use for several of the parameters included in our cost-effectiveness models. For these grants, the values we are particularly uncertain about include:

• Counterfactual mortality with no bednet distribution (per 1,000 child-years) in Nigeria. Our estimate for Nigeria of child mortality in the absence of LLINs is one of the highest of all the countries where AMF works. While the value we use (an estimated 18 deaths per 1000 child-years)33 is not a major outlier (the value is comparable in Guinea and is 10-12 deaths in DRC, Togo, and Uganda),34 we find it surprising because, according to the World Bank, Nigeria has higher per capita income than those other countries,35 and higher per capita income generally correlates with lower child mortality.36 This raises a concern that the data may not reflect reality.
• Ratio of 5+ malaria deaths to under-5 malaria deaths in Togo.37 Many of the parameters in our cost-effectiveness models use data we obtain from the Institute of Health Metrics and Evaluation (IHME)'s Global Burden of Disease (GBD) project. One such parameter is the ratio of malaria deaths in individuals over 5 and in individuals under 5. The ratio for Togo is one of the highest of all the countries for which we have reviewed this data and is substantially higher than the median for these countries.38 We aren't aware of a reason why Togo would be an outlier for this parameter and are therefore concerned that the data may not reflect reality. We plan to discuss this question with IHME.
• Likelihood that the Global Fund and/or PMI would replace philanthropic costs.39 In evaluating grants that we make or recommend, we routinely ask: what would happen to this program if it did not receive this grant? Would another funder step in to support it, or would it go unfunded? Through this line of questioning, we develop a best guess of what would happen in a scenario without GiveWell-directed funding, which we refer to as "the counterfactual." Our best guesses about what would happen without GiveWell-directed funding can substantially affect our estimate of a grant's cost-effectiveness, but they are necessarily based on subjective guesses because we don't have the opportunity to witness the counterfactual. This is particularly true when we try to make predictions about the counterfactual behavior of other funders, as we can only speculate about their future priorities and decisions. As such, the adjustments we make to account for our crowding out of other funders (described above) are highly uncertain.

Uncertain use of future revenue

We don't know how AMF will use the future revenue that is freed up by these grants. We project that AMF will raise about $30 million over the next year. As discussed above, we did not include this revenue in our estimate of the funding available to fill these funding gaps because we believe that AMF should commit the funding for these campaigns as soon as possible (see above). Consequently, one effect of these grants is to free up AMF's future revenue for support to other campaigns. While we don't know which campaigns AMF will choose to support, the opportunities we expect it to consider are likely to be cost-effective. These include campaigns in PMI-supported states in Nigeria in 2023 and in provinces in DRC in 2024. See this page for a discussion of the funding we granted or recommended to AMF in early 2021 to support similar opportunities in a different set of PMI-supported states in Nigeria and in DRC.

Not yet available post-distribution monitoring results

As discussed above, at the time of these grant investigations, post-distribution monitoring results from the 2020 campaigns in Uganda and Togo were expected but were not yet available. We have therefore reviewed only registration and backcheck data, which we view as weaker evidence on the outcomes of these campaigns than we would post-distribution monitoring results. We could have waited for these results to become available before making these grants. However, we decided that the benefit of AMF ordering nets as soon as possible (see above) outweighed the benefit of waiting for more information, given our expectation that this information is unlikely to substantially decrease our estimate of the cost-effectiveness of this grant.

Plans for followup

• We have monthly calls with AMF to discuss its work. We plan to share partner feedback with AMF and discuss potential ways to reduce the challenges caused by AMF's operating model and requirements.
• We will request information from AMF about the types of nets it procures for each campaign and the information it uses to make those decisions.
• We will track whether campaigns occur on schedule or are delayed.
• We will review the registration data and post-distribution monitoring data collected for these campaigns to understand what proportion of LLINs reach their intended recipients, are used by those recipients, and remain effective while they are in use.
• We will track how national malaria programs and other funders choose to allocate the funding available for future LLIN campaigns in these countries.
• We will continue to build relationships with the partners listed below to request feedback on AMF's work.

Internal forecasts

Grant	Confidence	Prediction	By time
Nigeria	50%	Conditional on completion of a retrospective analysis to update parameters in the model such as costs, total number of people reached, and durability of nets distributed, we will conclude that this grant was at least 15.8x cash.	End of 2026
Nigeria	60%	Campaigns in Global Fund-supported states that are scheduled for 2022 will start in 2022 in at least 75% of the states.	End of 2022
Uganda	50%	Conditional on completion of a retrospective analysis to update parameters in the model such as costs, total number of people reached, and durability of nets distributed, we will conclude that this grant was at least 15.4x cash.	End of 2027
Uganda	80%	The average time between when districts received nets in the 2020 campaign and the next campaign will be at least 34 months.	End of 2024
Togo	50%	Conditional on completion of a retrospective analysis to update parameters in the model such as costs, total number of people reached, and durability of nets distributed, we will conclude that this grant was at least 7.8x cash.	End of 2027

Our process

We learned about these funding gaps from AMF. Our grant investigations relied heavily on our prior work modeling the cost-effectiveness of LLIN campaigns supported by AMF and our relationship with AMF and knowledge of its work. To generate grant-specific cost-effectiveness estimates, we used our existing cost-effectiveness model for LLIN campaigns and updated various parameters to match the specifics of each funding gap. One Program Officer and two Senior Researchers who weren't otherwise involved in the grant investigations reviewed and gave feedback on the main uncertainties in our cost-effectiveness models.

We aim to get feedback on our grantmaking from stakeholders other than our top charities, such as government officials, other implementers involved in delivering the program, or other organizations working in the relevant context. The goals of these conversations are to learn more about the context in which a program will be delivered, to confirm the need for additional support of the program, and to seek feedback on the activities that a potential grant to support the program would enable. The external conversations we had about these grants include:

• Nigeria
  • Nigeria’s national malaria program (National Malaria Elimination Programme)
  • The Global Fund
  • PMI
  • Malaria Consortium
• Uganda
  • Uganda’s national malaria program (National Malaria Control Program)
  • The Global Fund
  • PMI
  • Malaria Consortium
  • PACE, an NGO that has conducted monitoring for AMF in Togo
  • AMF's former representative based in Uganda
• Togo
  • Togo’s national malaria program (Programme National de Lutte Contre Le Paludisme)
  • The Global Fund
  • GRASE, an NGO that has conducted monitoring for AMF in Togo

We value the insights we gained by speaking with these organizations and appreciate the time they spent answering our questions. We note that the views expressed on this page, and any errors, are our own.

Sources

Document	Source
Against Malaria Foundation, "4 of Togo's 5 Regions, 2017, Togo"	Source (archive)
Against Malaria Foundation, "4 of Togo's 5 Regions, 2020, Togo"	Source (archive)
Against Malaria Foundation, "All 4 Regions, 2020, Uganda"	Source (archive)
Against Malaria Foundation, "Western & Eastern, 2017, Uganda"	Source (archive)
Against Malaria Foundation, Committed (under contract) nets	Unpublished
Against Malaria Foundation, Funding status, November 20, 2021	Unpublished
Against Malaria Foundation, Uganda Country Funding	Unpublished
GiveWell, "Against Malaria Foundation," 2021	Source
GiveWell, "Changelog for 2022 CEA Version 1," 2022	Source
GiveWell, "GiveDirectly," 2020	Source
GiveWell, "GiveWell Values," 2021	Source
GiveWell, "GiveWell's "Excess Assets" Policy," 2017	Source
GiveWell, "Grants and Funding Recommendations in January and February 2021" 2021	Source
GiveWell, "Insecticide Resistance and Malaria Control," 2020	Source
GiveWell, "Malaria Consortium — Support for LLIN Distribution Campaigns in Ondo and Anambra States, Nigeria (March 2021)," 2021	Source
GiveWell, "Recommendation to Open Philanthropy for Grants to Top Charities," 2019	Source
GiveWell, "Revisiting leverage," 2018	Source
GiveWell, "Room for more funding analysis for AMF," 2021 (public)	Source
GiveWell, "Top Charities Fund," 2021	Source
GiveWell, "Why we can’t take expected value estimates literally (even when they’re unbiased)," 2011	Source
GiveWell, "Why we’re excited to fund charities’ work a few years in the future," 2020	Source
GiveWell, 2020 campaign registration data	Unpublished
GiveWell, Cost-effectiveness analysis - version 1, 2022	Source
GiveWell, GBD 2019 SSA under 5 to over 5 malaria mortality ratios, 2021	Source
GiveWell, Likelihood of crowding out GFATM/PMI (October 2020) [public]	Source
GiveWell, Likelihood of crowding out GFATM/PMI (updated October 2021) (public)	Source
GiveWell, Summary of AMF PDM results and methods [2020]	Source
GiveWell, Tracking of Global Fund spending in countries receiving funding from AMF [2021]	Source
IHME, "Global Burden of Disease"	Source (archive)
Malaria Consortium, Rapid Scoping to Delineate Priority Areas for ITN Distribution and Gap Analysis, 2020	Source
Roser, Ritchie and Dadonaite, "Child and Infant Mortality," 2013	Source (archive)
The World Bank, "GDP per capita (current US$) - Sub-Saharan Africa"	Source (archive)

• 1

  See our room for more funding analysis here, “RFMF projections sheet.”

• 2

  For more details, see this page. The two states not covered by one of these funding sources are Ondo and Anambra.

• 3

  This is not to be confused with a previous grant that GiveWell made to AMF to support campaigns in two states that are supported by PMI.

• 4

  The Global Fund supports the following states in Nigeria: Kwara, Niger, Adamawa, Gombe, Taraba, Yobe, Jigawa, Kaduna, Kano, Katsina, Delta, Ogun, and Osun. Malaria Consortium, Rapid Scoping to Delineate Priority Areas for ITN Distribution and Gap Analysis, 2020, Pg. 9.

• 5

  A similar gap exists for campaigns that are scheduled to occur in 2023 in states that are supported by the Global Fund. We do not yet know whether the Global Fund will decide to reallocate funding to fill that gap.

• 6

  AMF plans to buy 2.7 million nets at a cost of roughly $2.40 per net in Nigeria. AMF will also use a small portion of the grant to pay for post-distribution monitoring. Against Malaria Foundation, Funding status, November 20, 2021 (unpublished)

• 7

  See our room for more funding analysis here.

• 8
  • AMF plans to buy 15.7 million nets with this round of support. Against Malaria Foundation, Funding status, November 20, 2021 (unpublished)
  • In 2017 and 2020, AMF distributed about 13 million nets per each campaign. Details of the 2017 campaign can be found here and for the 2020 campaign here.

• 9

  AMF plans to buy 15.7 million nets at a cost of roughly $2.00 to $2.30 per net in Uganda, depending on the proportion of different net types—which differ in price—that it decides to buy. Number of nets from Against Malaria Foundation, Funding status, November 20, 2021 (unpublished), cost of nets from Against Malaria Foundation, Email, October 14, 2021 (unpublished).

• 10

  See our room for more funding analysis here, “RFMF projects” sheet, Togo row.

• 11
  • AMF plans to buy 4 million nets with this round of support. Against Malaria Foundation, Funding status, November 20, 2021 (unpublished)
  • In the 2017 campaign, AMF distributed about 2.4 million nets and in the 2020 campaign, AMF distributed about 4 million nets. Details of the 2017 campaign can be found here and for the 2020 campaign here.

• 12

  AMF plans to buy 4 million nets at a cost of roughly $2.00 per net in Togo. AMF will also use a small portion of the grant to pay for post-distribution monitoring. Against Malaria Foundation, Funding status, November 20, 2021 (unpublished)

• 13

  Some of our top charities have a policy of holding funding reserves. In our room for more funding analyses, we typically include reserved funding as funding available to support program activities. We do this both to ensure consistency across top charities (as not all top charities hold reserves) and to understand the true effect of granting additional funding (i.e. whether additional funding would support undertaking additional program activities versus building or maintaining reserves).

• 14 See our room for more funding analysis, “Funding Commitments” tab, Reserves row. We count reserved funding as a funding commitment, and thereby exclude it from available funding.

• 15

  Before receiving the grants described on this grant, AMF had committed to purchasing around 50 million nets for other campaigns. Against Malaria Foundation, Committed (under contract) nets (unpublished). These grants increase that figure to around 72 million nets (50 million nets + 2.7 million for Nigeria + 15.7 million for Uganda + 4 million for Togo).

• 16

  See our cost-effectiveness analysis here.

• 17

  See our cost-effectiveness analysis here.

• 18

  See our cost-effectiveness analysis here.

• 19

  The Global Fund provides funding to countries on a three-year cycle; the current cycle is running from 2021-23.

• 20

  For more details, see this page. The two states not covered by one of these funding sources are Ondo and Anambra.

• 21

  World Bank funding is restricted to Borno, Abia, Imo, Rivers, Ekiti, and Lagos. Islamic Development Bank funding is restricted to FCT, Kogi, Enugu, and Edo. Malaria Consortium, Rapid Scoping to Delineate Priority Areas for ITN Distribution and Gap Analysis, 2020, Pg. 9

• 22

  Details of the 2017 campaign can be found here and for the 2020 campaign here.

• 23

  The Global Fund’s malaria allocation for Uganda increased from $188,322,878 in the 2018-2020 cycle to $263,024,950 in the 2021-2023 cycle, an increase of 33%. Its allocation to Uganda’s LLIN campaign increased from $76,946,542 to $84,231,291 over the same period, an increase of 9%. GiveWell, Tracking of Global Fund spending in countries receiving funding from AMF, 2021.

• 24

  The malaria programs include indoor residual spraying, routine LLIN distribution systems, integrated community case management, and other smaller programs. Against Malaria Foundation, Uganda Country Funding

• 25

  Details of the 2017 campaign can be found here and for the 2020 campaign here.

• 26

  See this spreadsheet, PMI line under section ‘Total budget for 2020 universal coverage campaign (only).’

• 27

  In the 2018-2020 cycle, the Global Fund malaria’s allocation for Togo was $$31,939,623. In the 2021-2023 cycle, the allocation is $$34,172,674. GiveWell, Tracking of Global Fund spending in countries receiving funding from AMF, 2021.

• 28

  GiveWell, Tracking of Global Fund spending in countries receiving funding from AMF, Source: Togo 2021-2023, Re Global Fund 2021-2023 section, HSS line.

• 29

  GiveWell, Tracking of Global Fund spending in countries receiving funding from AMF, Source: Togo 2021-2023 Re Global Fund 2018-2020 section, HSS line

• 30

  See Against Malaria Foundation, "4 of Togo's 5 Regions, 2017, Togo", Against Malaria Foundation, "4 of Togo's 5 Regions, 2020, Togo", Against Malaria Foundation, "Western & Eastern, 2017, Uganda", and Against Malaria Foundation, "All 4 Regions, 2020, Uganda"

• 31
  • Our LLINs cost-effectiveness model includes a "Number of people covered with each net" parameter, for which we use a value of 1.8.
  • Registration data from Togo shows that the ratio of registered LLINs to people living in registered households was 1:1.73, which is similar to the value we use. In Uganda, this ratio was 1:2.05, which means that fewer LLINs per person were distributed, compared to the value we use. According to AMF, this is because underestimation of the target population led to an undersupply of nets. Paraphrased information from AMF responses to GiveWell, October 2021 (unpublished).

• 32
  • Backcheck data was collected from ~3.5% of registered households in both countries. The first step in analyzing this data is to pair backcheck records with original records. AMF paired almost 100% of backcheck records with original records in Togo, but only around 50% in Uganda. GiveWell, 2020 campaign registration data (unpublished).
  • According to AMF, this is because there were issues with the system for matching households, which required enumerators to identify ID numbers that had been written in chalk on households days earlier. Paraphrased information from AMF responses to GiveWell, October 2021 (unpublished).
  • The next step in analyzing this data is to check for correspondence between backcheck records and original records. Correspondence within +/- 1 (meaning the backcheck record for a household showed the same number of LLINs registered as the original record, one more, or one fewer) was 95% for LLINs in Uganda and 85% for LLINs in Togo. These rates are comparable with those we have seen in other AMF registration and backcheck data sets. GiveWell, 2020 campaign registration data (unpublished).

• 33

  See this cell of our cost-effectiveness analysis, ‘Counterfactual mortality with no bednet distribution (per 1,000 child-years)’

• 34

  See our cost-effectiveness analysis, ‘Counterfactual mortality with no bednet distribution (per 1,000 child-years)’

• 35

  In 2020, Nigeria had a per capita income of $2,097.10, DRC of $544.00, Togo of $915.00, and Uganda of $822.00. The World Bank, "GDP per capita (current US$) - Sub-Saharan Africa".

• 36

  “As one would expect, income level of the country is extremely correlated with child mortality rate. The poorest countries have the highest levels of child mortality, and the countries with the highest income have the lowest rates.” Roser, Ritchie and Dadonaite, "Child and Infant Mortality," 2013

• 37

  See the parameter for the ratio of 5+ malaria deaths to under-5 malaria deaths in our cost-effectiveness analysis here.

• 38

  The ratio of 5+ malaria deaths to under-5 malaria deaths in Togo was 189% while the median for the countries we surveyed was 84%. See GiveWell, GBD 2019 SSA under 5 to over 5 malaria mortality ratios, 2021

• 39

  See this parameter in our cost-effectiveness analysis.