Against Malaria Foundation — Support for ITN Campaigns in Chad, DRC, Nigeria, and Zambia (December 2024)

Note: This page summarizes the rationale behind a GiveWell grant to the Against Malaria Foundation (AMF). The information below reflects our views at the time we made the grant decision and does not reflect any information we have learned or work we have done since that point, including the 2025 funding cuts to the United States Agency for International Development (USAID) and President’s Malaria Initiative (PMI). AMF staff reviewed this page prior to publication.

Published: December 2025

1. Summary

1.1 Background

Malaria is a disease caused by Plasmodium parasites which are transmitted to people through the bites of infected mosquitoes. Symptomatic cases involve flu-like symptoms including fever, which can progress to severe illness or death. Insecticide-treated net (ITN) distribution is one of two main WHO-recommended strategies for malaria vector control.1 Our full research report on ITNs is available here.

In December 2024, GiveWell recommended a $96 million grant to the Against Malaria Foundation (AMF) to support the procurement and delivery of ITNs in Chad, the Democratic Republic of the Congo (DRC), Nigeria (Akwa Ibom, Bauchi, Benue, and Plateau states), and Zambia between 2025 and 2027. AMF is one of GiveWell's top charities.

1.2 What we think this grant will do

This grant will fund AMF to purchase and support the distribution of approximately 35 million nets for use in mass ITN campaigns in Chad, the DRC, Nigeria (Akwa Ibom, Bauchi, Benue, and Plateau states), and Zambia between 2025 and 2027. In all countries, AMF will lead the net selection and procurement process as well as campaign monitoring and will work with national and international partners on campaign logistics.

We think this grant will increase the number of people protected against malaria by ITNs and avert the deaths of approximately 17,000 children under the age of five. We also expect the grant to avert the deaths of approximately 10,000 older children and adults and increase the incomes of protected children later in life.

1.3 Why we made this grant

Our best estimate is that this grant will be approximately 9-19x as cost-effective as unconditional cash transfers (GiveWell’s benchmark for comparing different programs), varying by location.2 At the time we made this grant, GiveWell’s funding bar for Top Charities was to fund grants that we estimate to be 8x or more as cost-effective as cash transfers.

High cost-effectiveness

We think there is a strong intuitive case for cost-effectiveness because:

• Malaria is a significant cause of child deaths in Chad, the DRC, Nigeria, and Zambia: We estimate that the total malaria-attributable mortality rate among 1-59 month olds in these countries ranges from 0.26% (Zambia) to 1.11% (Ituri, DRC). We triangulate our baseline estimate of the malaria mortality rate based on inputs from the Global Burden of Disease (GBD) study 2021, the United Nations Inter-agency Group for Child Mortality Estimation (UN IGME), and select additional surveys. Given that our funding will fund ITN distributions in specific areas of Chad, the DRC, Nigeria, and Zambia, we added a subnational malaria burden adjustment to reflect our best guess of the burden in the specific areas of these countries that GiveWell funding is likely to target. In addition, we added adjustments for deaths indirectly caused by malaria, baseline net coverage, and overlapping programs (seasonal malaria chemoprevention and malaria vaccines) that are not accounted for in our sources of malaria burden data. (more)
• ITNs provide significant protection against malaria. We think there is strong evidence that nets are effective at reducing child mortality related to malaria (68% reduction in malaria mortality for people under five with full effective ITN coverage). For more information about the impact of ITNs on malaria, see our ITNs intervention report.
• It is relatively cheap to reach people with ITNs: ITNs are relatively inexpensive to purchase and deliver. We estimate that it costs less than $5 to purchase and deliver an ITN in the countries we are supporting with this grant. Taking into account the cost per net delivered, net usage rates, the number of people who sleep under nets, and higher net usage among young children, we think that it costs roughly $17 to $21 per additional child under five to sleep under a net, depending on geography. (more)
• We think most people who access nets through campaigns would not otherwise have access to them: To estimate the counterfactual impact of our funding, we estimate the proportion of children under 5 who would use ITNs without these campaigns. This involves estimating the proportion of children under 5 who would access ITNs via routine channels - antenatal care (ANC) clinic visits or through immunization appointments (the Essential Programme on Immunization, or EPI) - or non-routine channels. We use our prior estimate of counterfactual ITN coverage in the DRC as a baseline and adjust it for other countries using survey and monitoring data. (more)
• We think that each ITN provides effective protection for one to two years: We estimate that ITNs provided via mass campaigns in the countries we are targeting with this grant provide between 1.2 and 1.6 years of effective coverage. Our approach for estimating the effective duration of protection from ITNs takes into account attrition of ITNs, the physical integrity of nets, and the residual insecticide content of nets. For more information about the duration of effective coverage of ITNs, see our ITNs intervention report.
• ITNs probably provide significant benefits beyond averting child mortality: In addition to the primary benefit of averting the deaths of young children, we expect this grant to avert the deaths of approximately 10,000 older children and adults. We also think that by averting malaria during a sensitive period of childhood development, ITNs could lead to income increases later in life. We estimate that around 18-39% of the total benefits of this grant come from increased income, depending on geography. (more)

A simple sketch of our cost-effectiveness analysis is below, using Nigeria as an example:

What we are estimating	Best guess (rounded)	Confidence intervals (25th-75th percentile)	Implied cost-effectiveness
Grant size (arbitrary value)	$1,000,000
Cost per person under age five reached	$20.20	$15.15 - $25.25	20x - 12x
Number of people under age five reached	~49,500
Proportion of reached children who would have slept under ITNs in the absence of the program	22%	14% - 30%	16x - 14x
Additional children sleeping under ITNs as a result of the program	~38,500
Years of coverage provided by ITNs	1.2	0.8 - 1.5	10x - 19x
Malaria-attributable mortality rate among people under age five	0.68%	0.34% - 1.02%	7x - 22x
Effect of ITN distributions on deaths related to malaria	55%	44% - 66%	12x - 18x
Number of deaths averted among people under age five	~170
Initial cost-effective analysis
Cost per under-five death averted (before adjustments)	~$5,875
Moral value of averting the death of a person under age five	116
Initial cost-effectiveness in terms of multiples of GiveDirectly's unconditional cash transfer program	6x
Summary of primary benefits (% of modeled benefits)
Mortalities averted for people under age five	50%
Mortalities averted for people age five and older	20%
Developmental benefits (long-term income increases)	30%
Additional adjustments
Adjustment for additional program benefits and downsides	53%	40% - 66%	14x - 16x
Adjustment for grantee-level factors	-4%	(-8%) - 0%	14x - 15x
Adjustment for leverage	0%
Adjustment for funging	-15%	(-26%) - (-3%)	13x - 17x
Overall cost-effectiveness (in multiples of cash transfers)	15x
Cost per life saved	~$4,300

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

Funding landscape for ITN campaigns

Our cost-effectiveness estimates are adjusted to consider the possibility that funding from GiveWell may displace funding that would have come from other sources–a concept we refer to as funging. In the context of ITN campaigns, these other funding sources are typically the Global Fund to Fight AIDS, Tuberculosis and Malaria (the Global Fund) and the President's Malaria Initiative (PMI). Funging could occur at the country level, where the Global Fund or PMI reduces their support for a particular country due to GiveWell’s support, or within-country, where Global Fund or PMI country teams shift funding away from mass ITN campaigns to other interventions within that country.

We investigated the funding landscape for mass ITN campaigns in these countries by speaking with national malaria programs in each country, state malaria programs in the case of Nigeria, officials from the Global Fund and PMI at both the international and country levels, and other experts, as well as by reviewing published data on malaria funding from the Global Fund and PMI. Based on this investigation, we believe (with a moderate degree of confidence) that the funding gaps targeted by this grant are unlikely to be completely filled by other funders. Our cost-effectiveness estimates are discounted by 10% (Kasaï-Oriental) to 16% (Chad) to account for the relatively small possibility of funging other funders. (more)

AMF's track record in the nets space and as a grantee

We believe that AMF has a strong track record of procuring ITNs for mass campaigns in a cost-effective manner and forming strong partnerships with other stakeholders involved in mass ITN campaigns. We have heard from stakeholders that AMF's involvement in mass ITN campaigns (notably their push for the digitization of campaigns) adds value above and beyond their direct procurement of nets. We also think that AMF's monitoring systems help us understand the functioning of ITN campaigns, and we are excited by additions that AMF is making to its monitoring regime that will help us understand the immediate impact of net distributions. (more)

1.4 Our main reservations

Our main cross-cutting reservations are as follows:

• We are particularly uncertain about the level of malaria burden we’re estimating in these geographies. We incorporate malaria burden estimates from different sources, but we don’t deeply understand why these sources differ and we don't know how confident we should be in each source. While we do have a cursory understanding of these models and the data sources they draw on, we don't have a strong sense of which estimation methods/sources are most trustworthy and we lack the data to interrogate the differences in these models. Furthermore, the subnational adjustments we apply to our burden estimates are based on noisy, uncertain inputs in which we have limited confidence. To partially address our uncertainties regarding subnational malaria burden, we regress our subnational estimates towards the national average, but we may further refine our approach for assessing subnational malaria mortality in the future.
• We are moderately uncertain about the longer-term effects of our funding on other funders’ plans and behaviors. Our assessment of funging risk focuses on the probability that the campaigns we are directly funding with this grant would have otherwise been funded by other actors, but we have not explicitly considered how our grantmaking may alter other funders' longer-term allocations. We think it's plausible that other funders could develop expectations (or could have already developed expectations) regarding GiveWell's funding of ITN campaigns and may adjust their funding allocations accordingly.
• We have a large outstanding research agenda in our ITN portfolio. We have several unanswered research questions that could impact how we model the cost-effectiveness of ITN campaigns in the future or alter our approach to considering malaria vector control funding opportunities. Advancing on our research agenda is a priority for the coming years.
• We have not conducted extensive follow-up on most of our past grants with AMF. This increases our uncertainty about future campaigns. We conducted a shallow review of some past campaigns during this investigation, but aim to conduct significantly more comprehensive reviews of AMF grantmaking starting in 2025.

We outline our country-specific risks and reservations below.

2. Basics

2.1 What is the problem?

Malaria is caused by Plasmodium parasites that are transmitted to people through the bites of infected mosquitoes.3 Malaria can result in life-threatening symptoms including fever and can pose a significant health risk, particularly to children under five years old and others with low immunity.4

2.2 What is the intervention?

An insecticide-treated net (ITN) is a net which has been treated with insecticide to kill and repel the mosquitoes that carry malaria.5 Long-lasting insecticidal nets (LLINs) are factory-treated ITNs made of material into which insecticide is incorporated or bound around the fibers.6 These nets are typically hung over beds to provide protection during sleep.

The World Health Organization (WHO) recommends ITNs, particularly LLINs, as a main strategy for malaria vector control.7 ITNs are primarily distributed via mass distribution campaigns, in which nets are delivered door-to-door to households or through central distribution sites in a community. Based on current WHO guidance,8 most countries aim to deliver these campaigns every 36 months. For more information about malaria and ITN campaigns, see our intervention report.

The specifics of mass ITN distribution campaigns vary by country. Generally, National Malaria Control Programmes (NMCPs) are responsible for national malaria control strategies and for organising mass ITN distribution campaigns. We provide more country-level detail on the campaigns we expect this grant to support below.

3. The organization

The Against Malaria Foundation (AMF) funds the procurement of insecticide-treated nets (ITNs).9 AMF engages with NMCPs in malaria-endemic countries and other parties to assess net needs, identify distribution partners, and establish accountability measures for ITN distribution campaigns.10 AMF does not generally carry out campaigns itself or employ staff in countries where it funds ITNs; distribution is arranged by NMCPs and/or other distribution partners. As of October 2025, AMF employs 15 staff members.11

AMF is a GiveWell Top Charity.12 We have directed over $230 million in grants to AMF to support the procurement of ITNs since 2014.13

4. The grant

This grant of $96.3 million will fund AMF to purchase and support the distribution of ~35 million nets14 for use in mass campaigns in Chad, the Democratic Republic of the Congo (DRC), Nigeria, and Zambia between 2025 and 2027.

In all countries, AMF will lead the net selection and procurement process as well as campaign monitoring (including a series of checks and surveys prior to, during, and after net distribution). AMF will also work with local and international partners on logistics, staff training, campaign planning, and community outreach.15

We think this grant will increase the number of people protected against malaria by ITNs and in turn avert a number of child deaths and increase the incomes of protected children later in life.

4.1 Grant budget and activities by country

Chad

We are allocating $25.9 million to fund the procurement and distribution of approximately 8.2 million ITNs for mass distribution campaigns in Chad beginning in January 2026.16 Of Chad’s 23 provinces, we think roughly 16 have sufficient malaria burden for ITN campaigns to be cost-effective.17 The Global Fund had previously committed to funding net procurement and distribution costs for eight provinces in Chad. We expect that our funding will enable AMF to procure ITNs for mass campaigns in Chad and, consequently, for the Global Fund to purchase fewer ITNs than they initially planned and instead spend some of their allocated funding on delivering the ITNs that AMF purchases. We expect that the combined Global Fund and AMF contributions will allow for mass campaigns to occur in the 16 high-priority provinces identified by AMF. The government of Chad has committed to funding 2 additional low-burden areas, but AMF was skeptical that funding will materialize.18 AMF aims to procure only CFP nets for the campaign in Chad, but may have to procure some PBO nets due to CFP manufacturing constraints.

DRC

We are allocating $14.9 million19 to fund the procurement of approximately 6.6 million ITNs20 for mass distribution campaigns in two provinces in the DRC: Ituri (beginning June 2026) and Kasaï-Oriental (beginning December 2025).

The ~$10 million in funding for Ituri will enable AMF to procure approximately 4.4 million PBO nets. This funding is meant to cover the cost of procuring ITNs for the entire province. No other funders have committed to funding the procurement or delivery of ITNs in Ituri for the mass campaign in June 2026, so further funding will be needed to cover the costs of net distribution in Ituri. AMF told us they believe that it is possible that the Global Fund will commit funds for net distribution through their portfolio optimization process,21 and that their commitment to funding net procurement costs increases the odds of funds coming through from the Global Fund.22 The Global Fund DRC team also told us that a commitment to fund net procurement would increase the chances of portfolio optimization funds being approved, but they also highlighted that there are very limited resources for portfolio optimization for the current grantmaking cycle and that they are doubtful that funding will emerge to cover net distribution costs.23 To manage the uncertainty surrounding net delivery funding, we made the portion of this grant earmarked for Ituri conditional on funding for net delivery materializing.24

The ~$4.9 million in funding for Kasaï-Oriental will enable AMF to procure approximately 2.2 million PBO nets for lower priority health zones in the province.25 AMF had previously committed $2.3 million to fund the purchase of 1.2 million ITNs to cover the highest priority health zones in Kasaï-Oriental.26 AMF told us that their prior funding enabled them to purchase nets for people living in health zones ranked 1 to 3b (~37% of the province).27 We expect that the current grant will enable AMF to purchase ITNs for the next highest priority health zones (zones 4 and 5) in Kasaï-Oriental. As Kasaï-Oriental has already been fully funded by TGF for non-net costs, the gap is for net costs only.28

We believe that the allocation of nets funding across DRC provinces has been historically fluid and subject to change. This belief is partially informed by the difficulty we have faced when looking back on previous grants we have made for ITN campaigns in the DRC; the precise counterfactual impact of our funding has been difficult to identify. While our best guess is that this grant will be funding nets campaigns in Ituri and Kasaï-Oriental, there is a chance that it could ultimately be counterfactually funding other provinces. That said, we have some degree of confidence that the funds won't go to lower-burden (and less cost effective) urban areas like Kinshasa and Lubumbashi.29

Nigeria

This $46.1 million grant will fund net campaigns in four PMI-supported states in Nigeria: Akwa Ibom (distribution starting Oct 2025), Bauchi (distribution starting May 2026), Benue (distribution starting Dec 2026), and Plateau (distribution starting Dec 2026).30 Based on state-specific insecticide resistance data, AMF plans to use Dual AI (IG2) nets in all states except for Akwa Ibom, which will use PBO nets.31

Though this grant will directly fund nets for campaigns in these four states, we believe that, counterfactually, we are ultimately contributing to PMI’s overall portfolio of ten states in Nigeria.32 Without our support, we believe that there will be a longer length of time between mass campaigns in the PMI states in Nigeria, whereas for other countries, we assume that regions that don’t receive a campaign will not receive a campaign within that cycle. This is based on our understanding of PMI’s decision-making process, as well as our review of past campaigns. PMI Nigeria does not have sufficient funding to run mass net campaigns in all 10 states every 36 months,33 and we believe that if they do not receive any additional funds, they will have a longer interval between campaigns, rather than dropping states entirely.34 We roughly estimate that our grant will shorten the interval between campaigns for the ten PMI-supported states, from roughly 48-55 months to 40 months. We are uncertain about the exact duration that we are shortening the interval between campaigns, as well as how to model the impact of this effect, but we are confident that we are shortening the interval between campaigns by at least six months.35

Zambia

We are allocating $9.8 million to enable AMF to procure 4.5 million PBO nets for the mass campaign in Zambia in June 2025.36 This funding will partially fill the funding gap for high burden provinces for the upcoming mass ITN campaign and enable campaigns in provinces that otherwise would likely not receive one. Based on all information we have received from AMF and the Zambia National Malaria Elimination Centre (NMEC), our understanding (with moderate confidence) is that this grant will counterfactually enable two additional provinces with malaria prevalence estimated at 20-25% to receive a mass ITN campaign–likely some combination of Eastern, Central, and Copperbelt.37 Combined with the existing funding from the Global Fund and PMI, we expect that mass ITN campaigns will occur in eight out of ten Zambian provinces; the two excluded provinces (Southern and Lusaka) have relatively low malaria burden.38

5. The case for the grant

For a general overview of why we think the mass distribution of ITNs are a cost-effective intervention, see our nets intervention report.

We are recommending this specific grant to enable AMF to procure ITNs for mass distribution campaigns in Chad, the DRC, Nigeria, and Zambia for the following reasons:

• High cost-effectiveness: Our best guess is that this grant is 9-19x as cost-effective as unconditional cash transfers,39 depending on geography, which is above our current bar for directing funding (GiveWell’s benchmark for comparing different programs).40 Our estimate of the dollar-weighted average of cost-effectiveness across all of these campaigns is ~14x cash. The main benefit that we expect from this grant is reduced child mortality. Our best guess is that the grant will avert ~17,000 child deaths from malaria. (more)
• Without our funding, campaigns in Chad, DRC and Zambia would cover fewer people, while campaigns in Nigeria would likely occur later, with a longer gap between campaigns: We think it is unlikely that another funder would have stepped in to fill these funding gaps if we did not make this grant. (more)
• AMF's track record in the nets space: We believe that AMF is a respected actor in the nets space and they increase the rigor and quality of net distribution and monitoring for the campaigns in which they are involved. AMF has also been a productive working partner to GiveWell over the course of this investigation. (more)

5.1 High cost-effectiveness

We are recommending this grant because we expect it to be a cost-effective use of funding. Our best guess is that the dollar-weighted cost-effectiveness of this grant across the geographies we are funding will be approximately 9-19 times as cost-effective as unconditional cash transfers (GiveWell’s benchmark for comparing different programs),41 which is above our current bar for directing funding.42 The main benefit that we expect from this grant is reduced child mortality. Our best guess is that the grant will avert ~17,000 child deaths from malaria.43

In our separate report on nets, we explain why we think ITNs are generally a cost-effective intervention. The specific factors driving high cost-effectiveness in this particular grant are discussed below.

Malaria is a major cause of child deaths in Chad, DRC, Nigeria, and Zambia

Our estimate of the under-five mortality rate attributable to malaria in these countries ranges from 0.26% (Zambia) to 1.11% (Ituri, DRC). We rely primarily on malaria-specific mortality estimates from the Global Burden of Disease (GBD) and United Nations Inter-agency Group for Child Mortality Estimation (IGME) models and make adjustments to these estimates to account for additional considerations. The key aspects informing our malaria mortality rate estimates in the countries we investigated for this grant are described below.

Incorporating various sources of burden data

First, we calculated the malaria mortality rate among 1-59 month olds based on GBD 2021 data and adjusted it based on complementary inputs from IGME.44 Historically, we relied almost exclusively on GBD as our source of burden data, but our red-teaming exercise found that there are sometimes substantial differences between GBD estimates and other sources of burden data. For this grant investigation, we incorporated burden estimates from non-GBD sources because we think they improve our estimates compared to relying exclusively on GBD. We adjust our estimates for the following outcomes with non-GBD data sources:

• Under-5 all-cause mortality: For all countries, we compared estimates from GBD 2021 with UN IGME 2021, and also did a quick search for additional recent surveys that report under-5 all-cause mortality rates. We calculated a weighted averaged under-5 all-cause mortality rate based on the available data and divided this by the GBD 2021 under-5 all-cause mortality rate in order to calculate an adjustment for differences between GBD 2021 and other sources.45
• Share of deaths attributable to malaria among 1-59 month olds: For all countries, we compared estimates from GBD 2021 with UN IGME 2021. We divided the weighted average by the mortality share estimate from the GBD 2021 data to calculate an adjustment for differences between GBD 2021 and other sources.46

Adding in these two adjustments caused a large increase in our burden estimate for Chad (+72%), a moderate increase for the DRC and Nigeria (+10% and +14%, respectively), and no significant change for Zambia.47

Subnational malaria burden

Next, we added an adjustment for subnational malaria burden because we want our estimates of malaria burden to reflect our best guess of the burden in the specific areas of Chad, the DRC, Nigeria, and Zambia that GiveWell funding is likely to target.48 We calculated the subnational adjustment by comparing proxies for malaria burden at the national and subnational levels. Where subnational proxies indicate a higher malaria burden, we applied an upward adjustment to our estimate of malaria mortality among 1-59 month olds in the CEA (and vice versa when subnational indicators point to lower burden).

Adding in this adjustment changed our burden estimates from -34% in Kasaï-Oriental (DRC) to +26% in Ituri (DRC), relative to national averages.

For Chad and Zambia, we calculated the subnational targeting adjustment by comparing two measures at the national and subnational levels:

1. The under-5 malaria mortality rate49
2. The under-5 all-cause mortality rate50

For the DRC, we followed the same logic, but only used the under-5 malaria mortality rate given that IGME does not report the under-5 all-cause mortality estimates at the subnational level for the DRC. For Kasaï-Oriental, we also added a second level of adjustments to reflect funding targeting lower priority health zones within the province.51

We added a 50% 'haircut' to our subnational adjustments for Chad and Zambia to reflect our uncertainty in subnational burden data.52 This haircut regresses the subnational burden estimate halfway between the national average and the naïve results based purely on subnational data.

• We apply a 75% haircut (i.e. an even greater regression towards the national mean) for our province-level adjustments for the DRC given that: a) we have fewer data points informing our subnational adjustment, and b) net allocations in the DRC seem especially dynamic, which makes us slightly less confident that our funding will go exactly where we expect it to.

In addition to incorporating various sources of burden data and adjusting for subnational targeting, we also add adjustments for deaths indirectly caused by malaria, baseline net coverage, and overlapping programs (seasonal malaria chemoprevention and malaria vaccines) that are not accounted for in our sources of burden data. We do not elaborate on these adjustments in this page because they had relatively less impact on our bottom-line understanding of malaria burden in the countries we investigated for this grant, and we have written about these adjustments in our ITN intervention report. We have several uncertainties about our burden estimates that we elaborate on below.

It is relatively cheap to reach people with ITNs

ITNs are generally a cheap commodity to purchase and deliver via mass campaigns (see our ITN intervention report for a detailed overview of how we calculate net costs). Based on information received from AMF, we believe that the cost per net delivered in the campaigns we are funding with this grant ranges from $3.85 (in Zambia) to $4.89 (in Chad).53

Taking into account net use, the number of people sleeping under each net, the proportion of the population under 5, and higher net use among children under 5, we estimate that it ranges between $17.21 (in Chad) and $20.20 (in Nigeria) for one additional child under 5 to sleep under a bed net.

We think most people who access nets through campaigns would not otherwise have access to them

In order to understand the counterfactual impact of our funding, we attempted to estimate the proportion of children under 5 in the countries we investigated for this grant who would use ITNs in the absence of these campaigns. At a high level, this involved estimating the proportion of children who would have accessed ITNs via a) routine channels (antenatal care [ANC] or infant immunization [EPI] visits) or b) non-routine channels (e.g. school-based or community-based distribution or through the private market) . We did this by building upon our analysis of counterfactual net coverage from a prior AMF grant to fund ITNs in the DRC.

We used our prior estimate of counterfactual ITN coverage in the DRC as a baseline and arrived at estimates for the other countries by incorporating adjustments based on survey and monitoring data as well as insights from conversations with external stakeholders. In the DRC, we estimate that the effective ITN coverage rate from routine channels among children under age 5 is 12% and that the effective ITN coverage rate from sources other than routine distribution channels is 6%.

After applying adjustments based on The Demographic and Health Surveys (DHS) program and Malaria Indicator Surveys (MIS), modelled estimates from the WHO/UNICEF, and academic papers, we ultimately arrived at the following estimated counterfactual ITN coverage rates in the absence of a distribution:54

• 18% for Chad
• 18% for DRC (both Ituri and Kasaï-Oriental)
• 15% for Nigeria
• 19% for Zambia

Our approach to estimating counterfactual ITN coverage in the absence of campaigns is moderately uncertain. We think that the data we have on net coverage from routine and non-routine sources in all countries is sparse and of low quality. We also have limited understanding regarding how nets actually move through routine distribution channels. Our approach to modelling routine net coverage assumes that coverage rates increase steadily over time, but this assumption may not be valid across the board. (more)

ITNs probably provide significant benefits beyond averting child mortality

In addition to the primary benefit of averting the deaths of young children, we expect this grant to avert the deaths of nearly 10,000 older children and adults.55 These benefits account for approximately 20% of the value of this grant. We estimate the number of deaths averted among older children and adults by estimating the ratio of over-five to under-five malaria deaths based on data from GBD 2021 and the WHO.56 We multiply this ratio by our estimate of the total deaths averted among people under age 5 to arrive at our estimate of under-five deaths averted. See our intervention report for more information on how we think about over-five mortality.

We also think that by reducing the incidence of malaria among children under the age of 15 (a sensitive period of childhood development), ITNs could lead to small income increases later in life. We estimate that around 18-39% of the total benefits of this grant come from increased long-term income, depending on geography. To arrive at our best guess on the effect of reductions in malaria on later-life income, we rely on evidence from two historical quasi-experiments, Bleakley 2010 and Cutler et al. 2010, that measure the effect of eradicating malaria on later-life income in the Americas and India. Based on this evidence (and further adjustments we apply to account for replicability and external validity concerns), we estimate that each case of childhood malaria averted increases adult income by 0.6%, and that these benefits persist over 40 years of an individual’s working life. See our intervention report for a detailed overview of how we think about long-term income increases.

5.2 Funding landscape for ITNs

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–a concept we refer to as fungibility. In the case of ITN campaigns, the other actors who may fund these campaigns are typically the Global Fund to Fight AIDS, Tuberculosis, and Malaria (“the Global Fund”) or the President’s Malaria Initiative (PMI). “Funging” could occur at the country level – (the Global Fund or PMI reducing their support for a particular country as a result of GiveWell’s support) or within-country (Global Fund or PMI country teams directing funding away from mass ITN campaigns to other interventions within that country).

We investigated the funding landscape for mass campaigns in these countries by having conversations with national malaria programs and other major ITN funders and reviewing published data on malaria funding from the Global Fund and PMI. As a result of this investigation, we believe that the funding gaps we are targeting with this grant would likely not be fully filled by other funders in the absence of this grant, but that some of our funding may be crowding out funding that could otherwise have come from other sources. We discount our cost-effectiveness estimates by 11% (Kasaï-Oriental) to 16% (Chad) to account for the possibility that we are funging other funders.

Below, we outline our general approach to assessing the risk of funging the Global Fund and PMI and then present our funging assessment for each of the four countries we investigated for this grant. We also briefly discuss the possibility of funging within AMF (i.e. our grant crowds out funds that AMF would have otherwise spent on these campaigns).

Assessing Global Fund and PMI funging risk

To estimate funging risk, we:

1. Reviewed data on how the Global Fund and PMI’s funding allocations have changed since AMF entered the countries we are investigating: Our goal was to identify changes in funding trends at the country level that may have resulted from AMF stepping in to fund mass ITN campaigns in given countries. We looked at several metrics that we believe could give us insight into this question, including the proportion of Global Fund spending allocated to malaria for a given country, the proportion of spending allocated to malaria vector control for both funders, and share of ITN funding going towards mass campaigns versus routine distribution channels.57 There are significant limitations to this analysis in terms of identifying funging. Primarily, funding allocation decisions are complex and attributing changes in trends to specific causes (such as AMF's funding) is not possible because we can't observe the counterfactual world in which AMF did not fund campaigns. Despite this significant limitation, we believe that this form of data analysis can be one useful method (among others) to gain context on the funding landscape.
2. Spoke with stakeholders, including staff from the Global Fund (TGF) and PMI (both international and country-level teams), national malaria programs, and a handful of implementers: These conversations helped us to better understand the decision-making process behind the Global Fund and PMI’s funding allocations and calibrate our judgements regarding funging risk. Our current understanding of these funders' allocations processes, and the ways in which funging risk could materialize, is roughly as follows:
   1. PMI: We do not fully understand how PMI’s cross-country funding allocations are determined. PMI staff told us that the U.S. Global Malaria Coordinator decides the funding allocation across countries by incorporating guidance from PMI leadership. We do not have concrete information on how these decisions are made.58
   2. Global Fund: We find the Global Fund's allocation process to be relatively transparent, as they provide public documentation on their process.59 However, we are uncertain about the extent to which AMF/GiveWell funding commitments are taken into account in these decisions. The Global Fund uses several factors when allocating funds across countries, one of which is expected funding from other donors.60 This suggests that our funding of AMF could reasonably crowd out funding from the Global Fund. Since GiveWell's funding of ITN campaigns is typically allocated late in the Global Fund grantmaking cycle (after initial country allocations are made), if our funding is crowding out the Global Fund, this would likely apply to cycles subsequent to the cycle in which we contributed funding.

We also take into account the counterfactual value of the Global Fund and PMI's spending when assessing the impact that funging has on our cost-effectiveness estimates. Because money is fungible, our grants can have the practical effect of freeing up the Global Fund or PMI to spend their funding on another program they would not otherwise have been able to fund. Our rough analysis of Global Fund spending suggests that their average program is roughly five times as cost-effective as unconditional cash transfers – below our bar for Top Charity funding, but still relatively cost-effective compared to many potential giving opportunities.61 Our perception of the relatively high cost-effectiveness of Global Fund and PMI funding means a high funging risk does not necessarily translate into a substantial impact on our cost-effectiveness estimates.

Chad

We roughly estimate that there is a 25% likelihood of funging in Chad for this campaign. This estimate is highly uncertain, but is informed by the following considerations:

• PMI does not currently support ITN distributions in Chad, so we would not expect it to fund this campaign in the absence of our funding.62
• GiveWell has not previously supported ITN campaigns in Chad, and AMF supported only one previous campaign. Therefore, we think it’s unlikely that the Global Fund allocated a lesser amount to Chad based on anticipating funding from AMF.63
• At the time we made the grant, there was an active $22,462,428 portfolio optimization funding request in the Global Fund's Register of Unfunded Quality Demand (UQD) for a universal ITN campaign for roughly the amount that we were planning to fund in Chad.64 We think it’s possible but unlikely that the Global Fund will fill this request, based on a) conversations with AMF and Global Fund staff65 and b) our assessment of previous Global Fund allocations to ITN coverage campaigns.66

DRC

We estimate that there is a 15% likelihood of funging in the DRC for these campaigns. This estimate is moderately uncertain and is informed by the following considerations:

• AMF’s previous support of DRC campaigns could present a moderate funging risk; however, we don’t think that the Global Fund or PMI are reducing their allocations to DRC due to AMF’s support. This is because there are still significant funding gaps for net campaigns in the DRC67 and because stakeholders have expressed skepticism regarding additional funds materializing from global funders.68
• It seems very unlikely that national or provincial-level governments in DRC would provide funding for net campaigns in Kasaï-Oriental or Ituri.69
• It's plausible that Global Fund's decision not to fund Ituri's ITN campaign stemmed from an expectation that AMF would provide funding.70 However, there are also reasons to believe that this decision didn't have to do with expectations of AMF providing funding.71

Nigeria

We estimate that there is a 20% likelihood of funging in the Nigerian states supported by this campaign. This estimate is moderately uncertain and informed by the following considerations:

• Our review of the data and conversations with stakeholders suggest that there is a genuine gap for ITN funding in Nigeria and a low probability of additional funders stepping in to fill this gap.72
  • Nigeria represents a significant proportion of global malaria mortality,73 but it has a lower level of net coverage than many other countries with similar levels of malaria burden,74 in part due to country caps imposed by the Global Fund and PMI.75
• Based on our analysis of PMI funding, we speculate that GiveWell's past funding may have led to a slight decrease in overall PMI Nigeria funding.76

Zambia

We roughly estimate that there is a 25% likelihood of funging in Zambia for this campaign. This estimate is highly uncertain (we think it could plausibly be anywhere between 0% and 50%), and is informed by the following considerations:

• PMI and the Global Fund are aware that GiveWell is considering funding this Zambia campaign which suggests that other funders may anticipate funding from GiveWell/AMF and reduce their funding correspondingly. However, our data analysis suggests that the Global Fund and PMI funding for malaria has remained constant over time despite AMF having supported past campaigns in Zambia.77
• A $15,894,963 funding request for supporting the 2025 net campaign is included in the Global Fund UQD register (unpublished). We believe it is unlikely, but possible, that the Global Fund could otherwise fulfill some or all of this request through portfolio optimization.78

Funging within AMF

We also considered the counterfactual value of AMF's funding when assessing this grant. During the course of this investigation, we worked with AMF to allocate approximately $39 million of their funds to what we believe were highly cost-effective opportunities (mass ITN campaigns in South Sudan and Kebbi, Nigeria). At the time we made this grant, we estimated that AMF had approximately $20 million in remaining unallocated funds.79 By choosing to fully fund the campaigns we investigated (Chad, DRC, Nigeria, and Zambia), we expected to funge this amount into other campaigns that AMF will fund in the future.

We are uncertain about how exactly AMF will allocate these remaining funds, and about how they generally allocate their own unrestricted funds to different campaigns. We intend to work with AMF going forward to better understand how they use their non-GiveWell funds.

5.3 AMF's track record

We believe that AMF has a strong track record of procuring ITNs for mass campaigns efficiently and acting as a partner to other stakeholders involved in mass ITN campaigns. Since 2014, GiveWell has provided 34 grants totalling over $230 million to support AMF's net procurement activities.80

In recent years, government officials, funders, and technical experts have told us that AMF makes valuable contributions to digitizing campaigns and works effectively with other stakeholders. We have heard that AMF's requirements for digitization and campaign monitoring are initially a hurdle for countries to comply with, but these requirements can enhance the accountability, transparency, and traceability of campaigns, so countries ultimately come to appreciate these efforts.81 Stakeholders and technical experts have also told us that AMF is an engaged partner and maintains constant communication with other stakeholders to ensure the success of net campaigns.82

AMF also has monitoring systems in place to track the transportation and distribution of ITNs and monitor net retention and use post-distribution. We believe these systems improve our understanding of the impact of net distributions. We have historically had concerns about AMF's distribution monitoring, but we have been working with AMF and IDinsight to improve certain aspects of AMF’s monitoring.83

6. Risks and reservations

6.1 Cross-cutting risks and reservations

Our main cross-cutting reservations about this grant are as follows:

We are particularly uncertain about the level of malaria burden we’re estimating in these geographies.

• We incorporate malaria burden estimates from different sources but we don’t deeply understand why these sources differ: GBD, IGME, Malaria Atlas Project (MAP), and (to a lesser extent) DHS surveys present estimates for the same outcome that sometimes differ substantially (>100%), and we don't deeply understand what drives these differences. For example, IGME's estimate for the share of deaths due to malaria among 1-59 month olds in Chad is more than 2x the GBD estimate for this outcome,84 and we can't readily understand why this discrepancy exists. For our subnational adjustments, we don't have a solid understanding of why MAP malaria mortality estimates vary across states/provinces. While we do have a cursory understanding of these models and the data sources they draw on, we don't have a strong sense of which estimation methods/sources are most trustworthy and we lack the data to interrogate the differences in these models.
  • Relatedly, we are also uncertain about the weights we place on these sources of estimates: We incorporated IGME and DHS survey estimates (alongside GBD estimates) of under-5 all-cause mortality and malaria mortality in the countries targeted by this grant. We weigh these data sources equally, but we feel uncertain about how much to trust each source.
• Our subnational adjustments are based on noisy, uncertain inputs: The estimates from MAP, GBD, and IGME that we draw on stem from complicated models that we don't fully understand. We have also heard several stakeholders express concerns about the accuracy of subnational malaria mortality estimates.85 Furthermore, we don't have direct estimates of the under-five malaria mortality rate at the subnational level, so we apply a conversion ratio between all-ages and under-five mortality rates, but we haven't verified the validity of this conversion ratio.86 Finally, as described above, we have some residual uncertainty regarding precisely which subnational areas GiveWell funding will target (especially in the DRC), which makes us hesitant to take the subnational burden adjustments at face value. To partially address our uncertainties regarding subnational malaria burden, we regress our subnational estimates towards the national average. We may further refine our approach to assessing subnational malaria mortality in the future.

We are uncertain about the longer-term effects of our funding on other funders’ plans and behaviors. Our assessment of funging risk focuses on the probability that the campaigns we are directly funding with this grant would have otherwise been funded by other actors in the absence of GiveWell funding, but we have not incorporated an explicit assessment for how our ITN grantmaking may alter other funders' longer-term allocations. We think it's plausible that other funders will develop expectations (or may have already developed expectations) regarding GiveWell's funding of ITN campaigns and could adjust their funding allocations accordingly. The specific risk we are concerned about is that other funders would leave funding gaps for ITN campaigns unfilled (when they would have otherwise filled them in a world without GiveWell's grantmaking) because they expect that GiveWell will step in and fill these gaps. We suspect that our ITN campaign funding decisions will impact the future actions of other funders, though we are very uncertain about what this could look like concretely and how to weigh the shaping of future expectations against present impact.

We have a large research agenda and significant research debt in our ITN portfolio. We have several unanswered research questions that could impact how we model the cost-effectiveness of ITN campaigns in the future or alter our approach to considering malaria vector control funding opportunities. An overview of some of the questions we have about ITN campaigns can be found here. Advancing on our research agenda is a priority for the coming years.

We have not conducted extensive follow-up on many of our past grants with AMF. This increases our uncertainty about future campaigns. We conducted a shallow review of our past campaigns during this investigation, but aim to conduct significantly more comprehensive lookbacks on AMF grantmaking starting in 2025.

6.2 Country-specific risks and reservations

• Chad
  • We are highly uncertain about the probability that the Global Fund would fund this campaign through its portfolio optimization process in the absence of GiveWell funding. We estimate this probability at 25%, though we think it could plausibly be higher or lower. (more)
• DRC
  • We have significant uncertainty about the counterfactual impact of our funding in the DRC. We have had difficulty understanding the impact of previous grants for ITN campaigns in the DRC, and are also unsure about the prioritization process that led to the current funding gaps in the DRC (specifically for Ituri). These challenges in understanding the dynamics of ITN campaigns in the DRC make us uncertain about the areas our grant will counterfactually fund, though we do feel reasonably confident that we won't be funding lower-burden urban areas (more).
    • Related to our uncertainty about the subnational counterfactual impact of our funding, we are unsure about using subnational burden estimates for Kasaï-Oriental and Ituri. Not only do we think that subnational burden estimates are noisy, but we are also unsure that these are the correct subnational areas to consider as our counterfactual.
• Nigeria
  • We have not incorporated data on subnational variation in net use across Nigeria. Though data collected by Malaria Indicator Surveys (MIS) conducted by the DHS program suggest that there is substantial variation in ITN use across regions of Nigeria, our cost-effectiveness analysis does not incorporate subnational data; instead, we use a national average for net use. We do not know whether the average in PMI states is higher or lower than the national average, so we do not know what the impact on cost-effectiveness would be. However even in a maximally pessimistic scenario, where we shift our estimate for “proportion of distributed nets that are used” in Nigeria from 73% to 50%, cost-effectiveness would shift from 14.8x to 11.1x, remaining above our Top Charity funding bar of 8x.87
  • We are uncertain about the best way to model our belief that funding campaigns in these states will shorten the interval between campaigns as opposed to causing campaigns to happen that wouldn’t otherwise. The current model uses a 33% downward adjustment to cost-effectiveness based on a previous analysis that assumed a nine-month reduction in campaign intervals. However, our best guess for this grant is that we are speeding up campaigns by more than nine months (from 48-55 months to 40 months). This makes us think that we are more likely to be underestimating rather than overestimating cost-effectiveness.
• Zambia
  • We are highly uncertain about the probability that the Global Fund would fund this campaign through its portfolio optimization process in the absence of GiveWell funding. We estimate this probability at 25%, but believe it could plausibly be significantly higher or lower. If this probability was significantly higher, this would bring our cost-effectiveness estimate for Zambia below our Top Charities funding bar of 8x (more).

6.3 Less important uncertainties

• We are unsure how long nets will last and be effective in these contexts: We didn’t spend much time examining net durability for this grant investigation, but improving our understanding of this topic is a priority on our research agenda. The durability parameters in our CEA are based on historical durability studies, which are likely noisy and not available for all countries or subnational regions. We guess that net durability could be -30% to +30% of what we are currently estimating. This implies a cost-effectiveness range of 9x to 17x for the whole grant; the most pessimistic scenario would push Kasaï-Oriental (DRC) and Zambia below our cost-effectiveness bar.88
• We are uncertain about the effects of insecticide resistance on our estimates: We made a few minor updates to our model of insecticide resistance during this investigation (updating our estimate of pyrethroid resistance in Chad and accounting for the possibility of chlorfenapyr resistance developing in the future). However, we still have several open questions about our approach to insecticide resistance, including PBO net efficacy by geography, whether we’ve properly accounted for chlorfenapyr net effects, and the relationship between mosquito mortality in lab tests and reductions in net efficacy. Insecticide resistance is also on our research agenda and we plan to do more work on this topic in the future.
• We are uncertain about our adjustment for counterfactual coverage: The data we have on net coverage from routine and non-routine sources is generally limited and of low quality across the board. In all countries, we have at least some data on ANC and EPI attendance or routine net funding. However, since this survey data is old in many cases and funding situations can change quickly, it is difficult to confidently estimate how many infants actually receive nets from these sources. Furthermore, we assume that routine net coverage gets a bit better over time, but this may not be a valid assumption across the board. For instance, we believe that ANC attendance in Zambia is already very high, so there may not be room for much improvement in routine coverage.

• Accounting for internally displaced persons (IDPs): Stakeholders mentioned that the presence of IDPs presents some challenges with respect to net distribution, coverage, and durability.89 IDPs are a relatively small fraction of the national population of the countries covered by these campaigns (~1-6%), though a larger fraction of Ituri (28%).90 Consequently, we only conducted a shallow review of the potential impact of IDPs on campaign efficacy and did not make any changes to our cost-effectiveness analysis on this basis.91

7. Plans for follow up

• We will have monthly calls with AMF to discuss the progress of net procurement and campaign implementation, including which types of nets are ultimately procured for each campaign and who will be implementing them.
• We will review the full set of monitoring information AMF collects during and after campaigns, including indicators related to coverage and net use. We will also track progress on AMF's research projects that are not explicitly part of this grant agreement, namely a pilot implementation of the use of individual QR codes for net distribution monitoring.
• We will compare the number of nets distributed to population estimates for all campaigns, with a particular focus on the DRC.
• We will track the timing of campaigns in all countries to understand whether they are occurring on schedule or are delayed.
• We will follow AMF's decision-making about how it uses its revenue from sources other than GiveWell grants, and we may estimate, in some form, the cost-effectiveness of the campaigns they fund with this revenue.
• We will try to identify ways to assess the claim that AMF’s participation in campaigns improves the quality of campaigns beyond the direct value of the additional nets they provide
• We will prioritize and execute some part of our malaria research agenda to resolve some of our unanswered questions.
• See our forecasts below for specific predictions we aim to follow up on.

8. Internal forecasts

For this grant, we are recording the following forecasts:

Confidence	Prediction	By time
50%	Less than 400 million dollars will be allocated through the next Global Fund portfolio optimization process	April 2025
50%	Non-net costs for the Ituri campaign become available through portfolio optimization	Mid-2025
25%	Zambia receives any portfolio optimization funds for the 2026 mass campaign92	End of 2026
50%	On average, the 4 Nigeria campaigns happen with a gap of less than or equal to 40 months	February 2027
60%	The campaign in Zambia happens within 3 months of the current projected start date	September 2026
45%	The campaign in Kasaï-Oriental happens within 3 months of the current projected start date	March 2026
40%	The campaign in Ituri happens within 3 months of the current projected start date	September 2026
55%	The campaign in Chad happens within 3 months of the current projected start date	April 2026
70%	Weighted-average cost-effectiveness of this grant will be >10x when assessed retrospectively (using AMF's monitoring data and updated burden estimates for 2021), by December 2027.	December 2027
50%	GiveWell's estimate of the malaria-attributable mortality rate among people under age five for these countries in 2021 will be >25% different in December 2026.	December 2026

9. Our process

• AMF presented us with information on this funding gap, after which we decided to launch a grant investigation.
• In investigating this grant, we relied on review of program documents and data from AMF, the Global Fund, PMI, and national malaria programs, and external literature.
• We had 24 conversations with third-party (non-AMF) experts and stakeholders involved in ITN campaigns in the countries and regions of interest.
• In addition, we had 10 calls with AMF’s leadership and sent 7 rounds of questions that they responded to via email.
• We updated our CEA for ITN campaigns with updated adjustments and parameters for this investigation.
• The grant team leading this investigation had check-ins with GiveWell's CEO and Director of Research throughout the investigation.

10. Sources

Document	Source
AMF, All Countries - Total Costs, 2024	Unpublished
AMF, Distributions	Source (archive)
AMF, DRC Funding Allocation Summary, 2024	Unpublished
AMF, Gap Estimate for Nigeria PMI States, 2024	Unpublished
AMF, Opportunity Assessment Document for Nigeria, 2024	Unpublished
AMF, Our process	Source (archive)
AMF, People	Source (archive)
AMF, Runthrough of AMF activities, 2024	Unpublished
AMF, What we do	Source (archive)
AMF, Zambia Prospective Country Allocation	Unpublished
Bleakley 2010	Source (archive)
Brooks et al. 2017	Source (archive)
Conversation with PMI, November 16, 2024	Unpublished
Conversation with The Global Fund DRC team, Nov 6 2024	Unpublished
Correspondence with Zambia NMEC, Nov 2024	Unpublished
Cutler et al. 2010	Source (archive)
GiveWell, Against Malaria Foundation	Source
GiveWell, Against Malaria Foundation — Support for ITN Campaigns in DRC (June 2024)	Source
GiveWell, Analysis of PMI and TGF funding tables, 2024	Unpublished
GiveWell, Analysis of share of deaths caused by malaria - 2024 AMF investigation	Source
GiveWell, IDinsight — Review of AMF's Monitoring (March 2024)	Source
GiveWell, Malaria Consortium — Support for LLIN Distribution Campaigns in Ondo and Anambra States, Nigeria (March 2021)	Source
GiveWell, Mass Distribution of Insecticide-Treated Nets (ITNs)	Source
GiveWell, Our Approach to Foreign Aid Cuts	Source
GiveWell, Subnational analysis of malaria burden, 2024	Source
GiveWell, What We Learned From Red Teaming Our Top Charities, Nov 2024	Source
GiveWell's analysis of ITN usage following mass campaigns, December 2024	Source
GiveWell's analysis of the counterfactual value of Global Fund spending, July 2023	Source
GiveWell's CEA of insecticide-treated net (ITN) distributions, December 2024	Source
ITN Access and Use Report, Nigeria	Source (archive)
PMI, Where We Work	Source (archive)
The Global Fund, Description of the 2023-2025 Allocation Methodology, 2022	Source (archive)
The Global Fund, Filling Critical Gaps	Source (archive)
The Global Fund, Operational Policy Note - Portfolio Optimization, 2024	Source (archive)
The Global Fund, Portfolio Optimization Operational Procedures, 2024	Source (archive)
The Global Fund, Register of Unfunded Quality Demand	Source (archive)
The Institute for Health Metrics and Evaluation	Source
The Malaria Atlas Project	Source
The UN Inter-agency Group for Child Mortality Estimation (IGME)	Source
WHO Malaria Terminology, 2021 Update	Source (archive)
WHO, Essential Programme on Immunization	Source
WHO, Guidelines for Malaria Vector Control, 2019	Source (archive)
WHO, Guidelines for malaria, 2023	Source (archive)
WHO, Malaria Fact Sheet, 2024	Source (archive)
WHO, World Malaria Report, 2023	Source (archive)

• 1

  “Interventions that are recommended for large-scale deployment in terms of malaria vector control are those that have proven protective efficacy to reduce or prevent infection and/or disease in humans and are broadly applicable for populations at risk of malaria in most epidemiological and ecological settings. Vector control interventions applicable for all populations at risk of malaria in most epidemiological and ecological settings are: i) deployment of insecticide-treated nets (ITNs) that are prequalified by WHO, and ii) indoor residual spraying (IRS) with a product prequalified by WHO.” WHO, WHO guidelines for malaria, 16 October 2023, p. 41.

• 2
  • 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.

• 3

  World Health Organization, Malaria Factsheet, December 8th 2022. Available here.

• 4

  World Health Organization, Malaria Factsheet, December 8th 2022. Available here.

• 5

  The WHO defines an insecticide-treated net as: “A mosquito net that repels, disables or kills mosquitoes that come into contact with the insecticide on the netting material. Insecticide treated nets (ITNs) include those that require treatment and retreatment (often referred to as conventional nets) and those are “long-lasting” (see definition of long-lasting insecticidal net).” WHO, WHO Malaria Terminology 2021 Update, p. 19.

• 6

  WHO definition: “A factory-treated mosquito net made of material into which insecticide is incorporated or bound around the fibres. The net must retain its effective biological activity for at least 20 WHO standard washes under laboratory conditions and 3 years of recommended use under field conditions.” WHO, WHO Malaria Terminology 2021 Update, p. 17.

• 7

  “Core interventions for malaria vector control are applicable for all populations at risk of malaria in most epidemiological and ecological settings, namely: i) deployment of insecticide-treated nets (ITNs) that are prequalified by WHO, which in many settings are long-lasting insecticidal nets (LLINs); and ii) indoor residual spraying (IRS) with a product prequalified by WHO. Once high coverage with one core intervention has been achieved, supplementary interventions – namely the deployment of chemical or biological larvicides – can be used in addition to the core interventions in specific settings and circumstances.” World Health Organization, Guidelines for Malaria Vector Control, 2019, xiv.

• 8

  “Campaigns should also normally be repeated every three years, unless available empirical evidence justifies the use of a longer or shorter interval between campaigns.” World Health Organization, Guidelines for Malaria Vector Control, 2019, pp. 39-40

• 9

  "We fund anti-malaria nets, specifically long-lasting insecticidal nets (LLINs), and work with distribution partners to ensure they are used. We track and report on net use and impact." AMF, What we do

• 10

  AMF, Our process

  • The accountability measures for ITN distribution campaigns include (but are not limited to) photo and video documentation of distributions, post-distributions report sent to AMF, and post-distribution check-ups to assess net usage.

• 11

  AMF, People

• 12

  GiveWell, Our top charities

• 13

  GiveWell has made 34 grants to AMF since 2014. The total value of these grants is $234,875,033. See GiveWell, Against Malaria Foundation - Previous AMF grants

• 14

  AMF, All Countries - Total Costs, 2024 (unpublished)

• 15

  AMF, Runthrough of AMF activities, 2024 (unpublished)

• 16

  AMF, All Countries - Total Costs, 2024 (unpublished)

• 17

  This reflects AMF calculation of the provinces that merit ITN campaigns. Their calculation of need (unpublished) is primarily based on surveys that measure 1) the prevalence of malaria among children aged 5-59 months and 2) the incidence of malaria in the general population. We have not vetted AMF's calculations or the data they use.

• 18

  Email from AMF, Nov 21, 2024 (unpublished)

• 19

  $9,980,000 for Ituri + $4,880,000 for Kasaï-Oriental = $14,860,000. AMF, All Countries - Total Costs, 2024 (unpublished)

• 20

  AMF, All Countries - Total Costs, 2024 (unpublished)

• 21

  The Global Fund uses portfolio optimization to reallocate "unused grant funds to areas where they can be invested to achieve greater impact." Global Fund, Operational Policy Note - Portfolio Optimization, 2024, p. 2

  The process involves determining how much funds the Global Fund has to reallocate and identifying high-priority programs from the Register of Unfunded Quality Demand that can absorb this funding.

• 22

  Email from AMF, Oct 31, 2024 (unpublished)

• 23

  Conversation with The Global Fund DRC team, Nov 6 2024 (unpublished)

• 24

  We will grant AMF funds for net costs in Ituri conditional on funds for non-net costs becoming available through TGF portfolio optimization. If these funds do not materialize, we will either not disburse those funds, or - if already disbursed - require that GiveWell has the right to sign off on alternative use of these funds. We may also investigate in 2025 whether funding non-net costs directly through TGF is a suitable solution. As the campaign in Ituri is not scheduled to begin until June 2026, this decision does not need to be made immediately - we believe can wait for the outcome of TGF portfolio optimization before deciding how we want to proceed (i.e., funding non-net costs in Ituri or requiring the funds to be used elsewhere).

• 25

  The health system in the DRC organizes the country into health zones (sub-provincial units). Health zones are stratified into eight priority levels (1a, 1b, 2a, 2b, 3a, 3b, 4, and 5) based on inputs including malaria burden, access to health care, receptivity (prevalence levels in the year 2000), and presence of internally-displaced people (IDPs). Email from AMF, March 12, 2024 (unpublished).

• 26

  AMF, DRC Funding Allocation Summary, 2024 (unpublished)

• 27

  We’re not sure of this exact figure because there is a discrepancy in the information provided by AMF: the DRC Funding Allocation Summary (unpublished) states 34% while the Oct 31, 2024 email from AMF (unpublished) states 39%. We noticed this discrepancy late in the investigation and did not prioritize confirming the true value; instead, we use a rough average of the two in this write-up.

• 28

  Written responses from AMF (unpublished)

• 29

  We have heard from both AMF and the Global Fund that they do not intend to fund campaigns in these areas given funding constraints. Source: written responses from AMF (unpublished) and conversation with The Global Fund DRC team (unpublished).

• 30

  These are the current projected start dates, based on a campaign interval of ~36 months. As discussed below, we anticipate that the campaigns will likely be delayed several months past the currently projected start dates.

• 31

  AMF email, Nov 11 2024 (unpublished)

• 32

  The ten states supported by PMI are: Oyo, Nasarawa, Kebbi, Akwa Ibom, Ebonyi, Bauchi, Cross River, Benue, Plateau, and Zamfara.

  • This is why we model all PMI states in aggregation in the CEA as opposed to only the subgroup of these four states or each of the four states individually.

• 33

  AMF, Gap Estimate for Nigeria PMI States, 2024 (unpublished)

• 34

  AMF, Opportunity Assessment Document for Nigeria, 2024 (unpublished)

• 35

  This is based on the fact that we’re adding significant funding that would otherwise likely not be available. On priors, it would make sense that we’re speeding up campaigns; 6 months is a fairly arbitrary guess.

• 36

  AMF, All Countries - Total Costs, 2024 (unpublished)

• 37

  We were told by AMF that, without further funding, campaigns would be funded in the six highest-prevalence provinces only (Northwestern, Luapula, Western, Muchinga, Northern, and Copperbelt) and that AMF funding would enable campaigns in Eastern and Central provinces. The Zambia NMEC told us that without further funding, they could potentially cover the provinces of Northwestern, Northern, Muchinga, Western, Luapula, and Eastern as well as high-burden districts in Central province; with the full amount of funding requested, they could cover all provinces except Lusaka. Source: Email from Zambia NMEC, Nov 2024 (unpublished). We don’t think the Zambia NMEC would fund Southern province with AMF funds.

• 38

  AMF, Zambia Prospective Country Allocation (unpublished)

• 39
  • Our estimate of the value per dollar donated to cash transfers is out of date as of 2024. We are continuing to use this outdated estimate for now to preserve our ability to compare across programs, while we reevaluate the benchmark we want to use to measure and communicate cost-effectiveness.
    • For more on our updated assessment of the impact of unconditional cash transfers, see this page.
  • We model cost-effectiveness for each country separately in our CEA. Within the DRC, we also model cost-effectiveness separately for Ituri and Kasaï-Oriental.

• 40

  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. As of early 2025, our bar for directing funding is about 10 times as cost-effective as unconditional cash transfers for non-top charity programs and about 8 times as cost-effective as unconditional cash transfers for top charity programs. See GiveWell’s Cost-Effectiveness Analyses webpage for more information about how we use cost-effectiveness estimates in our grantmaking.

• 41

  Our estimate of the value per dollar donated to cash transfers is out of date as of 2024. We are continuing to use this outdated estimate for now to preserve our ability to compare across programs, while we reevaluate the benchmark we want to use to measure and communicate cost-effectiveness.
  For more on our updated assessment of the impact of unconditional cash transfers, see this page.

• 42

  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. As of early 2025, our bar for directing funding is about 10 times as cost-effective as unconditional cash transfers for non-top charity programs and about 8 times as cost-effective as unconditional cash transfers for top charity programs. See GiveWell’s Cost-Effectiveness Analyses webpage for more information about how we use cost-effectiveness estimates in our grantmaking.

• 43

  See calculations here.

• 44

  In past investigations, we have relied exclusively on GBD data as the foundation of our disease burden estimates. However, we don't fully understand the methods GBD uses to estimate disease burden and have noticed inconsistencies that lower our confidence in their estimates. Further, experts we have spoken with expressed concerns about the accuracy of GBD's estimates and suggested balancing GBD's estimates against other sources (unpublished conversations with external experts, 2024).

• 45

  Where we only have data from GBD and IGME, we gave 50% weight to GBD and 50% weight to IGME. This is based on our belief that both these data sources are imperfect but that there don't seem to be compelling reasons to trust one over the other.

  • For DRC (for which we also incorporate a recent DHS survey), we put ⅓ weight on each source.

• 46

  UN IGME directly reports the share of 1-59 months old mortality attributable to malaria. For GBD we divided the number of deaths due to malaria among 1-59 month olds by total 1-59-month old mortality in each geography. We didn’t look for other data sources for this outcome.

  • We calculated a weighted average share of mortality due to malaria by putting 50% weight on GBD and 50% weight on IGME.

• 47

  Percent changes are reported relative to GBD's malaria mortality rate estimate.

• 48
  • Best guess subnational areas for Chad: Mayo Kebbi Est, Tandjile, Chari Bagurimi, Hadjer Lamis, Guera, Batha, Sila, Ouaddai
  • Best guess subnational areas for the DRC: Ituri, Kasaï-Oriental (health zones in priority levels 4 and 5)
  • Best guess subnational areas in Nigeria: Our best guess is that the correct counterfactual would be all PMI states except Sokoto (i.e. Oyo, Nasarawa, Kebbi, Akwa Ibom, Ebonyi, Bauchi, Cross River, Benue, Plateau, and Zamfara), because it sounds like Sokoto orphaned now. Our best guess is that the impact of keeping Sokoto in is small, because there are around 10 other PMI states.
  • Best guess subnational areas for Zambia: Central, Copperbelt (in our adjustment calculation we use estimates for Central and Eastern, but the data for Copperbelt and Eastern are very similar so updating the precise provinces wouldn't make a material difference)

• 49

  Data for malaria mortality came from MAP and IHME.

• 50

  Data for all-cause mortality came from UN IGME.

• 51

  Brief explanation of methods for the Kasaï-Oriental subnational adjustment: This adjustment involved two stages:

  Stage 1: We compared the provincial burden to national burden using the under-5 malaria mortality rate to arrive at a province–level adjustment. We applied a 75% haircut to reflect our uncertainty and keep ourselves somewhat anchored to national burden estimates, which we see as being much more reliable than the provincial estimates.

  Stage 2: Stage 2 refines this adjustment to reflect that marginal funding for K-O will target lower priority health zones.

  We calculated part 2 of this adjustment using data AMF provided us during the previous DRC grant investigation. We compared incidence, prevalence, and mortality estimates at the provincial level with estimates for priority 4 and 5 health zones to arrive at implied adjustment factors based on each outcome. We took an equally weighted average of these outcome-specific adjustment factors to arrive at the 'Within province health zone adjustment factor'.

  We multiplied the 'Within province health zone adjustment factor' by the 75% adjustment from part 1 to arrive at our final adjustment factor for Kasaï-Oriental.

  The second stage of this adjustment was a quick and rough attempt to account for within-province targeting. We suspect that the data we used to inform this adjustment isn't great, and actual burden could be higher or lower than we estimate.

• 52

  We heard experts express significant concerns about the quality and reliability of subnational malaria burden data (unpublished). We have also noticed some surprisingly large differences in burden estimates for neighboring subnational regions, which don't seem plausible and decrease our confidence in the subnational data.

• 53

  We present the cost per net excluding in-kind government contributions because we account for these in-kind government contributions separately in our adjustment for other actors’ spending. Because we think AMF's spending causes these government costs to be incurred, we adjust the impact of the program downward to account for benefits that are lost as a result of these funds not being spent elsewhere. We also present AMF as covering all costs other than the government’s contributions in our summary of the proportion of costs covered by different actors, even though we think the Global Fund actually covers some costs. This is because we do not have permission to publish the proportion of costs covered by the Global Fund and because these proportions do not make a difference to our bottom line. See more in our ITN intervention report.

• 54

  These estimates take into account attrition in net use and the durability of nets.

• 55

  The country-by-country estimate of over-5 mortalities averted can be found here.

• 56We rely on three data sources to estimate the ratio of over-five malaria deaths to under-five malaria deaths. These data sources and the weight we place on each source are as follows:
  • GBD 2021 country-level estimate (50% weight): We divide the number of malaria deaths among people age 5 and older by the number of malaria deaths among people age under 5. We use country-level estimates from GBD 2021. These estimates can be found here.
  • GBD 2021 global estimate (25% weight): We divide the number of malaria deaths among people age 5 and older by the number of malaria deaths among people age under 5. We use global estimates from GBD 2021. The global ratio for this value is 0.76.
  • WHO's 2023 World Malaria Report (25% weight): This report states that "The percentage of total malaria deaths in children aged under 5 years decreased from 87% in 2000 to 76% in 2015." (p. xix). This implies that, in 2015, 24% of global malaria deaths occurred amongst people aged over 5 years old. We arrive at the ratio of over-five malaria deaths per under-five death by calculating 0.24/(1-0.24) = 0.32.

  Previously, we relied only on GBD country/state-level estimates to calculate the over-5:under-5 death ratio. These country/state-level ratios from GBD showed unexpected variation, which made us worried about introducing noise to our model. Furthermore, WHO has an alternative methodology for estimating the over-5:under-5 malaria mortality ratio that we think is worth giving some weight to in our model. We think our current approach of using multiple data sources reduces the amount of country/state-level noise we introduce into our model by balancing out country-level GBD estimates with global estimates from GBD and WHO.

• 57

  GiveWell, Analysis of PMI and TGF funding tables, 2024 (unpublished)

• 58

  Conversation with PMI, November 16, 2024 (unpublished)

• 59

  See for example Global Fund, Description of the 2023-2025 Allocation Methodology

• 60

  "The country’s raw allocation for the disease is then adjusted to account for: • Minimum shares (US$500,000 per disease component) to ensure that the funding is viable; • Maximum shares (disease allocations are limited to a maximum of 10% of total available disease funding, and country allocations are limited to 7.5% of total funding) to ensure that the funding is not overly concentrated in a few countries; • Projections of other external financing, to help align the global distribution of total external resources for the disease with the distribution of the raw allocation." Global Fund, Description of the 2023-2025 Allocation Methodology (p. 3)

• 61

  We have not conducted a comparable cost-effectiveness analysis of PMI spending but apply the same counterfactual value of spending for PMI as we do for the Global Fund.

• 62

  PMI works in 27 countries in Africa; we do not know why they work in some countries rather than others.

• 63

  AMF supported the 2023 mass campaign in Chad.

• 64

  The UQD register includes funding requests from countries that cannot be funded by their initial Global Fund allocation. Some requests on this register end up being funded when resources become available.

  • "When countries submit funding applications to the Global Fund, they are encouraged to include more health investments than can be funded through the Global Fund alone. This prioritized request for funding above their allocation is evaluated by the Technical Review Panel, and the strategically important programs for which there is no immediate funding are registered as “unfunded quality demand”.

  Should funding become available in the future, such as from government, private sector or nongovernment donors, the unfunded quality demand can be easily and efficiently identified and funded." Global Fund, Filling Critical Gaps

  • "These investments are often funded through savings or efficiencies during grant-making, but can also be funded through additional resources that may become available during the cycle, such as Portfolio Optimization, debt to health swaps or private sector contributions." Global Fund, Unfunded Quality Demand

• 65

  Conversations with external stakeholders, 2024 (unpublished)

• 66

  For example:

  • Our analysis of funding data from the Global Fund's 2021-23 and 2024-26 grantmaking cycles (unpublished) indicates that the Global Fund contribution to Chad has increased on a proportional basis: from the 2021-23 cycle to the 2024-26 cycle, the allocation to malaria funding in Chad as a percentage of total TGF malaria funding has increased slightly, from 1.5% to 1.76%.
    • These percentages are calculated as follows. Data from AMF and data.globalfund.org.
    • $60,906,351 (21-23 Chad malaria funding) / $4,061,486,741 (21-23 Total TGF malaria funds allocated) = 1.5%
    • $73,459,637 (24-26 Chad malaria funding) / $4,176,981,877 (24-26 Total TGF malaria funds allocated) = 1.76%
  • It seems reasonable (though still very uncertain) to assume that Global Fund contributions will remain consistent cycle to cycle; the fact that they instead increased suggests both that the Global Fund is not reducing contributions in expectation of AMF or GiveWell support, as well as that further funding through portfolio optimization may be somewhat less likely.

• 67

  The Global Fund caps disease allocations for each country at 10%. Given the magnitude of DRC’s funding need, this cap is binding for DRC and there are consistently funding gaps for the country’s malaria prevention efforts.

  "Maximum shares: components are limited to a maximum of 10% of total disease funding." Annex 3 of the Allocation Methodology for the 2023-2025 Allocation Period (p. 21 of system pagination)

• 68
  • AMF told us that they believe there is a very low risk of funging because the Global Fund has exhausted their funding for the DRC, and there is still a significant funding gap in high-burden areas. Conversation with AMF, Oct 24 2024 (unpublished)
  • Stakeholders from the Global Fund have also told us that portfolio optimization funds will be very limited this cycle. Assuming this is true (which we are uncertain about), it is unlikely that any provinces with funding gaps for net campaigns will receive additional funding from TGF through portfolio optimization. Conversation with external stakeholders, 2024 (unpublished)
  • The PMI DRC team told us that they have very limited funds available for mass campaigns and are only planning to support operational costs for one province (Kasaï Central). Conversation with PMI, Nov 7 2024 (unpublished)

• 69

  Neither PMI nor the Global Fund DRC team could confirm that domestic resource mobilization would result in significant funds for mass ITN campaigns. Conversation with PMI DRC team, Nov 7 2024, and comments on a draft of this page by The Global Fund DRC team, Oct 23 2025 (unpublished)

• 70

  We heard from both AMF and TGF that one of the reasons that TGF has not funded an ITN campaign in Ituri in this cycle to date is because it is the second to last province in the current campaign cycle, meaning there is more time for fundraising. This might mean that TGF decided not to fund Ituri under the hope that AMF would provide funding for net costs, while funding for non-net costs could be secured through portfolio optimization. Though TGF indicated that AMF committing funding for net costs could increase the likelihood of securing PAAR funding, they did not explicitly confirm that AMF’s potential contribution was a factor in Ituri being left out of the TGF budget. Email from AMF, Oct 9 and conversation with The Global Fund DRC team, Nov 6 2024 (unpublished)

• 71
  • The Global Fund DRC team emphasized to us that they are only stratifying the country because of budgetary constraints - if budget was not a concern, they would want to conduct net campaigns across the country because malaria burden is generally high everywhere. As Ituri is a large province and The Global Fund is stratifying by province, it could simply be that TGF does not have the funds to cover the entire province. Conversation with The Global Fund DRC team, Nov 6 (unpublished)
  • There could also be non-malaria factors at play; we heard from AMF that The Global Fund’s province-based stratification is not solely based on malaria considerations, but also takes TB and HIV-related factors into account. Written responses from AMF (unpublished). Note: we did not confirm this with The Global Fund.

• 72
  • Conversations with stakeholders in Nigeria indicated that there has consistently been minimal to no funding available for routine net distribution; the only source of nets for these channels is the leftovers from mass campaigns. Conversations with external stakeholders, 2024 (unpublished)
  • Several states have been “orphaned” from Global Fund or PMI malaria funding and received no mass net campaigns in many years; around 13 states were previously orphaned as of 2021, and PMI recently decided to drop Sokoto due to lack of funding. Some (we are not sure whether all) “orphaned” states have some funding for malaria programs via loans from development banks, including the World Bank and the Islamic Development Bank, but our understanding is that these loans will not be used for mass net campaigns. This understanding is from our conversation with external stakeholders, 2024 (unpublished).

• 73

  Per GBD 2021, Nigeria accounts for approximately 30% of the global share of under-5 and all-ages malaria mortality.

• 74

  Per MAP, the estimated rate of net use in Nigeria is slightly lower than average across Africa and is slightly lower than in the next highest burden countries (DRC, Uganda, Mozambique), though confidence intervals for net use all overlap.

• 75
  • The Global Fund caps the amount of funding it will allocate to any given country at 7.5%. "Country allocations are limited to 7.5% of the total funding." Annex 3 of the Allocation Methodology for the 2023-2025 Allocation Period (p. 21 of system pagination)
  • PMI does not have a formal cap, but has historically followed TGF’s lead in limiting its investment in Nigeria to roughly 9-10% of its total budget (unpublished)

• 76

  We reviewed data on PMI funding from 2020-2024 (unpublished). GiveWell first funded mass ITN campaigns in Nigeria in 2021 so data from 2020 represents PMI spending prior to any influence by GW. From 2020 to 2024, PMI’s Nigeria allocation dropped 5%, from $77 million to $68 million, despite the fact that the overall PMI budget increased by 7% during that same period. Taken together, the portion of the PMI budget allocated to Nigeria from 2020 to 2024 dropped 17.5%, from 10.4% to 8.6%.

  • It is possible, though highly speculative, that by allowing PMI to conduct more frequent ITN campaigns in Nigeria, GiveWell’s support has reduced the pressure on PMI to reconsider their low level of support (on a per-capita basis) for Nigeria.
  • If we roughly estimate that GiveWell’s ITN funding has caused approximately 50% of the reduction in PMI funding for Nigeria, that suggests it has led to a $4.5 million annual reduction; the most pessimistic assumption (e.g., assuming we were 100% responsible for the decrease) would be approximately $9 million less annually. Attributing 50% of the decrease to GiveWell is a rough estimate based on an intuitive sense that it’s likely one factor out of several, and the actual value could plausibly range from 0-100%.
  • We are fairly confident that our funding has not led to within-country funging away from ITNs in Nigeria. From 2020 to 2024, support for ITN campaigns has remained roughly around 35% of PMI Nigeria’s budget (unpublished). This suggests that PMI has continued to allocate similar funding to mass net campaigns since 2020 despite GiveWell’s support for mass campaigns in this period.

• 77
  • While GiveWell has not directly supported campaigns in Zambia, AMF has supported previous ITN campaigns in 2018 and 2023.
  • Our analysis of TGF and PMI funding data (unpublished) from 2014-2024 shows that total funding to Zambia has increased over this time horizon. Over this time horizon both total funding to Zambia and total universal ITN coverage funding have increased. As a percentage of the total international malaria budgets for each institution, funding to Zambia has stayed roughly similar for PMI (around 3.5% from 2014 to 2024) and increased for Global Fund. From 2017 onwards, TGF funding to mass campaigns in Zambia as a percentage of overall TGF malaria funding has increased, from 0.35% in 2018-2020 to 0.61% in 2024-2026. The fact that the funding to mass ITN campaigns in Zambia is increasing rather than decreasing may suggest that TGF is not reducing its funding to Zambia as a result of GiveWell/AMF funding.

• 78
  • We were told by a number of stakeholders that countries have been instructed to anticipate no or minimal portfolio optimization funding due to the lower amount of total funds available in this TGF cycle. We are uncertain how accurate this is and will follow up to see how this went.
  • Our analysis of funding data (discussed in the above footnote) suggests that PMI funding to Zambia has stayed consistent since prior to when AMF began supporting campaigns in Zambia, and TGF funding has increased. The fact that the current proposed level of TGF funding for Zambia (prior to any potential rounds of portfolio optimization) represents an increase over past years suggests that they are less likely to provide further funding through portfolio optimization, as we assume that TGF typically keeps the proportion of funds going to each country roughly similar year over year.

• 79

  In written responses (unpublished), AMF confirmed that as of December 2, 2024, they had $62.6m to allocate in total.

• 80

  GiveWell has made 34 grants to AMF since 2014. The total value of these grants is $234,875,033. See GiveWell, Against Malaria Foundation - Previous AMF grants

• 81

  Several conversations with external stakeholders, 2024 (unpublished)

• 82

  Several conversations with external stakeholders, 2024 (unpublished)

• 83

  See the grant page for this engagement.

• 84

  See here.

• 85

  Conversations with external stakeholders, 2024 (unpublished)

• 86

  This conversion ratio gets applied in our subnational adjustment model here.

• 87

  This is almost certainly more pessimistic than the reality; however, we erred on the more negative side as we are uncertain as to how the net usage: access ratio translates into the proportion of distributed nets used in our CEA.

• 88

  See this row in the CEA sensitivity analysis.

• 89

  Conversations with external stakeholders, 2024 (unpublished)

• 90

  See our estimates of IDP populations:

  	Estimated IDP population	Source	Estimated Population	Source	% IDPs
  Nigeria	3,300,000	International Displacement Monitoring Centre (2023)	230,271,000	UN (2024)	1.4%
  DRC	7,100,000	UNHCR (2024)	120,000,000	DRC PNLP (see p.16)	5.9%
  Ituri	1,246,044	UN (2024)	4,392,200	City Population, citing Institut National de la Statistique, DRC	28.4%
  Kasai Oriental	13,261	UN (2024)	3,864,300	As above	0.3%
  Chad	220,610	UN (2024)	19,319,064	WHO (2023)	1.1%
  Zambia	21,000	Internal Displacement Monitoring Centre (2023)	20,723,965	WHO (2023)	0.10%

• 91
  • The WHO strongly recommends that nets are distributed in humanitarian emergency settings based on 4 RCTs indicating high effectiveness. A shallow review of this evidence suggests comparable levels of effectiveness in emergency settings to the rates that we model in our CEA for non-emergency settings; however, we did not review these studies or their generalizability in depth.
  • A qualitative review, however, suggests concerns related to net usage and durability. For example, that review suggests that nets are more difficult to use in the tents typically used in IDP camps, which might suggest lower usage rates, and that some nets may not last as long due to the need to more frequently clean nets that fall on the floor.
  • In this investigation, we placed the strongest weight on the fact that IDPs constitute a relatively small proportion of the population covered by these campaigns; hence, even if effectiveness were substantially lower in these settings (which we do not believe is the case given the RCT evidence), they would still be above our bar, all else held equal.

• 92We are filling the whole gap that we believe is cost effective, but there will still be small regions that will be uncovered, and they could receive funds through UQD to cover those gaps