Note: This page summarizes the rationale behind a GiveWell-recommended grant to Sightsavers. Sightsavers staff reviewed this page prior to publication.
Summary
In January 2023, GiveWell recommended that Open Philanthropy grant $11.2 million to Sightsavers. Most of this grant, $8.3 million, is to continue support for deworming programs that GiveWell grants have previously supported, in Nigeria, Chad, and Cameroon; $2.9 million is to support expansion of the program in Chad from 6 to 16 regions. Combined with prior funding, this grant will fund all covered geographies for three years (through Sightsavers 2025 program year, ending in March 2026).1
We recommended this grant because we believe that the work that the grant will support will be cost-effective. Deworming is among the most cost-effective programs we know of, in certain locations. The need for deworming appears to be high in the areas where Sightsavers expects to use this grant. We have followed Sightsavers' work on deworming since 2015 and, more recently, in these specific locations, and we believe that Sightsavers is well-positioned to support this work.
Published: April 2023
Table of Contents
Planned activities and budget
This grant will extend support for ongoing programs through March 2026. Sightsavers expects to use $1.9 million of this grant to continue supporting deworming in six regions in Chad, $4.9 million to continue supporting deworming in 12 states in Nigeria, and $1.5 million to continue supporting deworming in six regions in Cameroon. Sightsavers also expects to use $2.9 million of the grant to expand its support for deworming in Chad to ten additional regions for three years.2
See here for a breakdown of how Sightsavers has spent past funding for deworming programs across spending categories.
The case for the grant
- We estimate that this grant will meet our bar for cost-effectiveness. More below.
- We believe that it is unlikely that another funder will cover these costs. More below.
- Sightsavers has a track record of supporting successful deworming programs. More in our review of Sightsavers' deworming program.
Cost-effectiveness
Based on our cost-effectiveness analysis of the program, we believe it is in the range of cost-effectiveness of programs we expect to direct funding to, as of January 2023 (about 10 times as cost-effective as unconditional cash transfers,3 or "10x" for short). Our overall cost-effectiveness estimate for this grant (based on weighting each geography's cost-effectiveness by the amount of funding allotted) is approximately 19x.4 Our estimates of cost-effectiveness for Sightsavers' program in each geography are below.
| Program | Cost-effectiveness (times as cost-effective as unconditional cash transfers) |
|---|---|
| Chad (existing areas) | 23x |
| Chad (expansion areas) | 14x |
| West, Cameroon | 6x |
| North, Cameroon | 21x |
| Far-North, Cameroon | 16x |
| East, Cameroon | 8x |
| Adamoua, Cameroon | 10x |
| South, Cameroon | 11x |
| Benue, Nigeria | 15x |
| Kebbi, Nigeria | 21x |
| Kogi, Nigeria | 17x |
| Kwara, Nigeria | 26x |
| Sokoto, Nigeria | 26x |
| Yobe, Nigeria | 5x |
| Kaduna, Nigeria | 31x |
| Kano, Nigeria | 15x |
| Katsina, Nigeria | 8x |
| Niger, Nigeria | 33x |
| Adamawa, Nigeria | 21x |
| Zamfara, Nigeria | 33x |
We chose to include an additional year of funding in the three geographies listed below, despite estimating cost-effectiveness to be less than 10x.5 All three geographies previously were estimated to have cost-effectiveness of at least 10x.6 We have discussed with Sightsavers that we may not continue to fund these geographies in future grants, barring any changes to our cost-effectiveness estimates or our cost-effectiveness bar for funding decisions. It is likely our cost-effectiveness estimates may change as we address some of the issues outlined below, but we do not yet know by how much:
- East, Cameroon. Our worm burden estimate (a key driver of geographic variation in our deworming cost-effectiveness estimates) may be underestimated for two reasons: (1) the data we have was collected after some deworming had been conducted in the region,7 and (2) the data we have is at the regional level whereas treatment is provided to a subset of districts for which treatment is indicated based on WHO guidelines.8 As a result, we may be underestimating cost-effectiveness.
- Yobe, Nigeria. We may be underestimating worm burden and overestimating the cost per child. Due to data limitations, worm burden is estimated at the state level rather than based on the subset of districts being treated.9 Cost per child may be overestimated due to a higher than usual cost per child in the most recent year of the program. Costs were incurred but some treatments were delayed to the following program year due to issues with drug availability.10 However, our best guess is that we will continue to estimate that Yobe is less than 10x in the future.
- Katsina, Nigeria. We may be underestimating worm burden and overestimating the cost per child for the same reasons as in Yobe, but the cost-effectiveness of deworming in Katsina is more likely to be greater than 10x in future analyses given the current 8x estimate.11
This grant also includes funding for West, Cameroon, despite estimated cost-effectiveness of approximately 6x. At the time of making the grant, we had estimated cost-effectiveness to be approximately 10x. However, we have since identified and corrected an error in our worm burden calculations. Due to a mistake in data entry, we were substantially overestimating the prevalence of moderate infections for whipworm (one of the relevant worm species). We believe this estimate has been mistakenly informing our decision making since 2020. We have informed Sightsavers of this error and have requested that they redirect funding from GiveWell donors after a wind-down period of two years. This may allow the funding designated in this grant for West, Cameroon in 2025 ($0.3 million) to be deployed to another, more cost-effective use (to be discussed between Sightsavers and GiveWell).
Our cost-effectiveness analysis for this grant is based on the same structure as our model for other deworming grants. While investigating this grant, we updated parameters within that model to use inputs specific to this funding gap. Below, we highlight parameters that vary for different funding gaps and have a substantial impact on our headline cost-effectiveness figures:
- Worm burden: For every deworming grant, we ask the potential grantee to provide data on the prevalence and/or intensity of infections with each species of schistosomiasis and soil-transmitted helminths in the locations where they would support deworming. We prefer to use data on the prevalence of moderate-intensity infections and of heavy-intensity infections for each species. When this isn't available, we use average intensity of infection or prevalence of any infection. We then apply an adjustment to our cost-effectiveness estimates of deworming programs to account for differences between the prevalence and intensity of worm infections in the geographies targeted by our deworming grantees and the prevalence and intensity of worm infections among the population studied in Miguel and Kremer 2004, the randomized controlled trial (RCT) on which we base our estimate of deworming's impact on consumption. We revisited all worm burden estimates for these geographies ahead of making this grant. Due to limitations of the available data, we have used region-level estimates for Cameroon, state-level estimates for Nigeria, and district-level estimates for Chad.12 It is worth noting that worm burden may be underestimated in all geographies due to treatment only occurring in a subset of districts or sub-districts within the area for which worm burden has been estimated. Updates to worm burden adjustments made for this grant resulted in an average increase in cost-effectiveness of 12% across geographies included in the grant.13
- Cost per child dewormed: Our cost per child estimates for each geography can be seen here. In most cases, our updates to cost per child decreased cost-effectiveness: in the regions of Cameroon being funded, by an average of 13%, and in the states in Nigeria being funded, by an average of 34%.14
Two additional updates to our cost-effectiveness analysis also affected our final cost-effectiveness estimates:
- We decreased our adjustment for "misappropriation without monitoring results" from 10% to 3% here in the cost effectiveness analysis as we now have coverage surveys for a greater proportion of Sightsavers' spending (80% of all-time spending).15 This increases our confidence that coverage estimates used in our cost-effectiveness analysis are representative of GiveWell-funded programs. This increased cost-effectiveness by an average 4% for all regions being funded by this grant.16
- We incorporated an adjustment to our cost-effectiveness analysis to account for the possible decay of deworming benefits over time, which decreased cost-effectiveness by approximately 10%.17 See here to read more about this update.
Funding landscape
For this grant, we considered what we know about the funding landscape in each of the countries, combining historical and new knowledge:
- Chad: Our prior understanding—that other funders are unlikely to support mass drug administration (MDA) in Chad, as described in our 2021 grant page—still appears accurate. The Bill and Melinda Gates Foundation (BMGF) recently participated in a 2023 funding gap analysis for neglected tropical diseases (NTDs) that compiled funding commitments from other major NTD funders and indicated that the analysis found no other sources of funding for MDA for schistosomiasis and soil-transmitted helminths in Chad.18
- Nigeria and Cameroon: We wrote about the impact of the UK government ending its support for these programs in 2021 in our February 2022 grant page. At that time, we were aware that the Bill and Melinda Gates Foundation (BMGF), the Children's Investment Fund Foundation (CIFF), and ELMA Philanthropies were planning to provide additional NTD funding to help address the gap left by the UK government. However, we were not aware where or for what activities this funding would be deployed. Since then, we've learned that this funding consortium (ARISE) is focusing their $65 million Phase 2 NTD funding commitment on Senegal, Ethiopia, South Sudan, Burkina Faso, and Kenya.19 Further, our current understanding based on conversations with other funders is that funding gaps remain in Nigeria and Cameroon.
As a result, we hypothesize that the probability of crowding out other philanthropic funding in Chad, Nigeria, and Cameroon is relatively low.
Risks and reservations
- We expect to receive updated worm burden data on Chad in the next one to two years, which could reduce our estimate of the cost-effectiveness of the program. This could potentially lead to our deciding not to make additional grants to the program. Instead of expanding the program in Chad now, we could have decided to wait to expand until after reviewing the new worm burden data. We communicated this risk to Sightsavers prior to making the grant.
- We have some concerns with the quality of Sightsavers' coverage evaluation surveys (CES), which may bias the results, potentially leading us to underestimate or overestimate cost-effectiveness:
- Length of time between deworming and the CES: Surveys in 2021 and 2022 were largely conducted five to six months after deworming, raising concerns of recall bias.20 Sightsavers has agreed to target conducting all coverage surveys within two months of MDA by the end of 2023.21
- Unclear data quality control and verification processes: Coverage survey reports do not mention any re-surveying of households to check the accuracy of the data collected or use of verification questions to assess the accuracy of responses. Some of the surveys note that supervisors monitored the data collection but do not specify what these processes were, how common errors were, or how they were corrected.22
- As mentioned above, we have not incorporated an adjustment for crowding out of other philanthropic funding in our cost-effectiveness analysis. This may lead us to overestimate cost-effectiveness. We hypothesize that the risk is highest in Cameroon and the Nigerian states, given that other NTD funders are also supporting work in these countries. As GiveWell has been providing support for several years in these countries, it is also more difficult to assess the likelihood of another funder stepping in if GiveWell funding were not available. On the other hand, we can be more confident that there is a low likelihood of crowding out other funders in Chad given we have not heard of any other funders considering supporting work in Chad, some of the country has not received MDA at all in recent years, and the UK government, which was the prior funder of NTD programs in Chad, has exited from funding NTD programs more broadly.23
- There are multiple ways in which our worm burden estimates could be overestimated or underestimated. For example, they may be underestimated—as discussed above—due to using data at a state or regional level rather than the district level where treatment decisions are made. On the other hand, they may be overestimated if we are wrong in assuming that worm burden would increase all the way to baseline, pre-treatment levels if deworming were to cease.24
- WHO recently updated its treatment guidelines for schistosomiasis,25 leading many countries to reevaluate their deworming strategies and gradually move toward these updated guidelines. This includes moving to more targeted treatment (e.g., at the community or sub-district level, rather than the district level) and expanding target populations beyond school-age children.26 It's not clear to us how that will affect funding needs and cost-effectiveness of deworming programs in these locations over the next few years. Changes could result in us having overfunded or underfunded these programs, and could also make it more difficult to evaluate cost-effectiveness in the future.
- Lastly, our cost-effectiveness model for deworming relies primarily on a single study and associated follow-up studies that suggest deworming increases future income. Please see our deworming intervention report for more discussion of this.
Plans for follow up
Sightsavers provides annual updates for all GiveWell-supported deworming programs. These updates include narrative progress reports, coverage survey reports, and annual spending reports.27 Sightsavers also shares informal updates by email and in conversation.
Internal forecasts
| Confidence | Prediction | By time |
|---|---|---|
| 30% | Katsina, Nigeria is > 10x at the end of our next grant investigation. | January 2024 |
| 10% | Yobe, Nigeria is > 10x at the end of our next grant investigation. | January 2024 |
| 40% | East, Cameroon is > 10x at the end of our next grant investigation. | January 2024 |
| 60% | Cost per child decreases in the April 2022-March 2023 program year for geographies that saw an increase in the prior year due to MDA delays (i.e., substantially fewer MDAs are delayed than in the prior year). | January 2024 |
| 75% | We will receive coverage evaluation survey (CES) reports for > 80% of Sightsavers' spending in the April 2022-March 2023 program year. | January 2024 |
| 65% | Surveyed coverage of Sightsavers' MDAs will remain > 80%, based on an average of surveyed coverage across programs weighted by the total spending associated with each program. | January 2024 |
Our process
Our process for this grant relied heavily on (a) our prior work on modeling the cost-effectiveness of Sightsavers' deworming program, (b) our monthly discussions with Sightsavers, and (c) our following Sightsavers' work on deworming programs we have funded since 2016.28 For this particular grant, we:
- Updated our cost-effectiveness analyses as described above.
- Reviewed the latest coverage evaluation survey reports and latest spending and treatment data (through the last program year ending in March 2022).
- Analyzed Sightsavers' room for more funding request.
- Spoke with other funders supporting NTD programs about the broader deworming funding landscape.
For internal review, a Senior Researcher and a Senior Program Associate who were not otherwise involved in developing the grant investigation gave feedback on the plan for investigating the grant. A Senior Program Associate reviewed the case for making the grant and gave feedback prior to final grant approval by a Senior Program Officer.
Sources
- 1
See our room for more funding analysis here for the specific geographies supported and grant amounts.
- 2
See our room for more funding analysis here for the specific geographies supported and grant amounts. Existing geographies in Chad are Logone Occidental, Logone Oriental, Mayo Kebbi Est, Mayo Kebbi Ouest, Moyen Chari, Tandjile. Expansion geographies in Chad are Bahr El Gazel, Batha, Chari Baguirmi, Guéra, Hadjer Lamis, Mandoul, Ouaddaï, Salamat, Sila, Wadi Fira. The states in Nigeria are Kaduna, Kano, Niger, Zamfara, Adamawa, Katsina, Kebbi, Kogi, Kwara, Sokoto, Benue, Yobe. The geographies in Cameroon are West, North, Far North, East, Adamaoua, and South. The final grant amounts are based on the budgets projected by Sightsavers, after allocating additional funds that Sightsavers expects to have available (i.e., due to rollover of excess GiveWell funding from prior grants).
- 3
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 2023, our bar for directing funding is about 10x as cost-effective as cash transfers. See GiveWell’s Cost-Effectiveness Analyses web page for more information about how we use cost-effectiveness estimates in our grantmaking.
- 4
See calculation here.
- 5
Note that we did not renew funding for two geographies which were estimated at less than 10x and had previously been less than 10x: Taraba, Nigeria, and Littoral, Cameroon. See here for our current cost-effectiveness estimates and see here for our prior estimates.
- 6
See here.
- 7
- Deworming began in Cameroon in 2007, and our worm burden model uses data for the East Region from a 2018 mapping survey. "Since 2007 and 2008 a nationwide distribution of mebendazole and praziquantel (in relevant schistosomiasis endemic areas) is implemented annually, respectively." National Programme for the Control of Schistosomiasis and Intestinal Helminthiasis, Republic of Cameroon, Report of the Impact Study for Schistosomiasis and Soil-Transmitted Helminthiasis in the Littoral, East and Centre Regions of Cameroon, 2018.
- Our worm burden model estimates the cost-effectiveness of treatment based on the baseline prevalence and intensity of worm infections, assuming that if deworming efforts ceased, worm infections would return to that baseline level. However, the data we use for the East Region of Cameroon was collected after some deworming had already taken place, and thus is not a true baseline and is likely to report a lower level of infection. As a result, it likely underestimates the prevalence and intensity of worm infections were deworming efforts to cease in that region.
- 8
- In Cameroon, we use region-level data for worm prevalence, as indicated here.
- "The deworming programs that Sightsavers supports conduct MDAs which aim to treat the entire population of children within districts or implementation units that meet the World Health Organization (WHO)-defined minimum prevalence thresholds for MDA with deworming pills." GiveWell, "Sightsavers' Deworming Program – November 2021 version"
- 9
- In Nigeria, we use state-level data of worm prevalence, as indicated here.
- "The deworming programs that Sightsavers supports conduct MDAs which aim to treat the entire population of children within districts or implementation units that meet the World Health Organization (WHO)-defined minimum prevalence thresholds for MDA with deworming pills." GiveWell, "Sightsavers' Deworming Program – November 2021 version"
- 10
See cost per child calculation here. As noted here, MDA only occurred in 8 of 17 LGAs (local government areas) in Yobe.
- 11
See cost per child calculation here. We use the same cost per child estimate for six of the states in Nigeria, including Katsina. As noted here, MDA was delayed resulting in no reported treatments.
- 12
Worm burden adjustments are here in our cost-effectiveness analysis. This is the supplemental sheet from which these estimates are pulled.
- 13
See here. 12% is the unweighted average increase in cost-effectiveness for geographies included in the grant.
- 14
See here. These are unweighted averages for only the geographies included in the grant.
- 15
See here for a summary of what Sightsavers' programs we have coverage surveys for.
- 16
See here. This is an unweighted average for only the geographies included in the grant.
- 17
See here.
- 18
Bill and Melinda Gates Foundation, conversation on January 25, 2023, and associated email communication (unpublished).
- 19
Bill and Melinda Gates Foundation, conversation on January 25, 2023 and associated email communication (unpublished).
- 20
See our review of Sightsavers' coverage surveys here
- 21
Call with Sightsavers, October 5, 2022 (unpublished).
- 22
See our charity review of Sightsavers here for more. Sightsavers also shared the following additional information on supervision practices after reviewing a draft of this page: "In our [coverage evaluation surveys], we focus on quality training of the teams, use of electronic data capture to minimize errors, and support supervision (on the ground and remote). Supervisors on the ground will move with the survey teams, to check they are adhering to survey methodology/protocol and review survey teams' interactions with various households. This includes how they are collecting and entering the information. If there are problems identified, teams will be provided additional support and, if still a problem, the team member may be replaced. There is normally about 1 supervisor to about 5 teams, so each team will get on the ground support and the supervisor can adequately address any issues. This is supported by remote supervision through the dashboard, focusing on data quality, team performance, and, through location tracking, ensuring sampling of households is as expected. Supervisors at all levels support trouble-shooting."
- 23
See here for more on Chad's funding landscape.
- 24
To the extent possible, we base our worm burden estimates on baseline, pre-treatment worm burden data.
- 25
See the WHO guidelines published in February 2022 here.
- 26
- "In endemic communities with prevalence of Schistosoma spp. infection ≥ 10%, WHO recommends annual preventive chemotherapy with a single dose of praziquantel at ≥ 75% treatment coverage in all age groups from 2 years old, including adults, pregnant women after the first trimester and lactating women, to control schistosomiasis morbidity and advance towards eliminating the disease as a public health problem." World Health Organization, World Health Organization, WHO guideline on control and elimination of human schistosomiasis, 2022, p. xvi.
- "Community mapping of the epidemiology of schistosomiasis can reduce the need for praziquantel, as treatment can be better targeted to communities and at-risk regions." World Health Organization, WHO guideline on control and elimination of human schistosomiasis, 2022, p. xvii.
- 27
See past narrative reports, coverage surveys, and spending reports in the Sources section of our Sightsavers page. The most recent coverage surveys are reviewed and linked in our monitoring review here. Sightsavers' most recent spending and treatment data are linked from our cost per child analysis here.
- 28
GiveWell first recommended funding to Sightsavers in 2016: GiveWell, "Our updated top charities for giving season 2016".