Evidence Action's Deworm the World — Renewal Grant for Nigeria, Pakistan, Kenya, and India (March 2024)

Note: This page summarizes the rationale behind a GiveWell grant to Deworm the World. Deworm the World staff reviewed this page prior to publication.

Published: June 2024

The organization

Deworm the World is a program led by Evidence Action, an organization that focuses on scaling up interventions that it believes are cost-effective and evidence-based. Deworm the World provides financial and technical assistance to governments implementing school-based and community-based mass drug administration for soil-transmitted helminths (STH) and schistosomiasis (SCH). We have been supporting Deworm the World since 2014 (see our previous grants in this table), and we continue to believe it shares our focus on cost-effectiveness and is well-positioned to be a government partner. For more information on the type of support Deworm the World provides, see our review of the organization here.

The intervention

Mass deworming involves treating large numbers of people with parasite-killing drugs: praziquantel kills the parasites that cause schistosomiasis (SCH), while albendazole or mebendazole kills soil-transmitted helminths (STH).1 Treatment is cheap and the side effects of the drugs are believed to be minor; thus, everyone in a given population (sometimes schoolchildren; sometimes the community at large) is treated, without being individually tested for the presence of infections.2

The potential consequences of SCH and STH infections include short-term declines in general health (such as weight and hemoglobin levels); in rare cases, severe symptoms such as organ damage or intestinal obstruction; and development impairment that can lower quality of life over the long-term.3 We believe the evidence for deworming's impact on short-term general health is thin (we believe deworming averts intestinal worm infections, but we don't think there is strong evidence for the impact that averting those infections has on general health), and that severe symptoms resulting from SCH and STH infections are highly infrequent, so the primary benefits we model from mass deworming are developmental benefits. We think it is highly plausible that children who receive deworming treatment will earn higher incomes over their lifetime, due to the protection against developmental impairment they receive in childhood. For more information on how we assess the benefits of mass deworming, see our separate report on deworming.

The grant

This $4.4 million4 grant provides an additional year of support for deworming programs in select locations in Nigeria, Pakistan, Kenya, and India. Our previous grants supported these programs through 2025.5

This grant breaks down as follows (see budget):

• Nigeria: $0.7 million to support deworming in Ogun, Oyo, Rivers, and Cross River in 2026. Note that we are not renewing support for Lagos, where we estimate deworming is less cost-effective.6
• Pakistan: $1 million to support deworming in Khyber Pakhtunkhwa in 2026.
• Kenya: $1 million to support deworming in 2026 and an impact assessment7 in 2026 that is expected to inform treatment strategies and exit timelines.
• India: $0.8 million to support deworming in Bihar and Uttar Pradesh in 2026 (including an impact assessment in Uttar Pradesh8 ), and $150,000 to support surveillance approaches in 2025 in three states where Deworm the World expects to have transitioned to remote technical assistance.9
• Other: $0.7 million to cover an increase in Deworm the World's global indirect costs rate from 18% to 20% for the 2024-2026 period.10

The case for the grant

We are recommending this grant because:

• We think mass deworming is likely cost-effective in locations with limited prior deworming (Pakistan, Nigeria).
• We are unsure how cost-effective deworming is in Kenya and India due to declining worm prevalence, but we think Deworm the World is pursuing a reasonable exit strategy to maintain those gains without supporting those programs indefinitely.

Cost-effectiveness

Cost-effectiveness is not the basis for our funding decision in all geographies because we do not believe our cost-effectiveness model accurately estimates the benefits of deworming in locations that have had extensive prior deworming. Prior deworming matters because our cost-effectiveness model currently assumes that worm burden would return to pre-treatment ("baseline") levels if mass deworming was stopped11 , but the latest guidance from WHO suggests that mass deworming can be halted when worm prevalence is sufficiently low. We take this to imply that mass deworming can durably drive down worm prevalence12 , which means that our model is likely overestimating the benefits of mass deworming by not accounting for worm burden declining over time. We think this flaw in our cost-effectiveness model makes it unsuited for estimating the benefits of deworming in locations like Kenya and India, where impact assessments have shown sharp declines in worm prevalence relative to baseline levels.13 We previously estimated deworming to be highly cost-effective in these locations.14

Since Nigeria and Pakistan have had fewer years of prior deworming treatment compared to the other geographies we support, we expect that worm burden is sufficiently high that it would likely rebound back to baseline levels if mass deworming were to stop. In other words, we think our estimate of the benefits of mass deworming in these locations could still be fairly accurate. We currently estimate deworming to be 11-15 times as cost-effective as unconditional cash transfers ("11-15x cash") in Nigeria and 10x cash in Khyber Pakhtunkhwa, Pakistan.15

See the relative maturity of the deworming programs supported by this grant below16 :

	First year of MDA	Most recent MDA	Year of impact assessment
Kenya	2012	2023	2021; [2026]
Bihar, India	2014	2023	[2025]
Uttar Pradesh, India	2014	2023	2022, [2026]
Cross River, Nigeria	2016	2023	[2024]
Ogun, Oyo, Rivers Nigeria	2017	2023	2022: Rivers [2024/25]: Ogun, Oyo
Khyber Pakhtunkhwa, Pakistan	2019	2022	[2025] tentative

Exit strategies

We are extending support for Deworm the World's work in Kenya and India, despite not having updated cost-effectiveness estimates for these locations, because 1) previous impact assessments suggest that mass deworming has successfully reduced worm prevalence and Deworm the World expects additional deworming will further reduce prevalence,17 2) Deworm the World has worked with governments to reduce treatment frequency (and thus reduce costs) in areas with lower worm prevalence,18 and 3) Deworm the World has drafted internal guidance that we believe lays out reasonable exit plans for geographies that reach very low levels of worm prevalence. We are renewing support for these locations this year, rather than providing exit funding now, because we are unsure how soon Deworm the World will be able to exit. Exiting too soon could run the risk of worm prevalence rebounding in the absence of Deworm the World's support, and we expect the results of future impact assessments (see table above) to inform exit timelines in these countries. At the time of recommending this grant, Deworm the World was expecting to be able to exit from Bihar and Uttar Pradesh in 2027, and to determine exit timelines for Kenya based on the results of the 2026 impact assessment.19

This grant also provides $150,000 to support surveillance approaches to monitor the risk of disease resurgence in three Indian states where Deworm the World expects to exit from in-person technical assistance soon.20 We expect the learnings from those pilots will be applicable to future exits from Bihar and Uttar Pradesh, and may inform exit strategies in other geographies as well.

Risks and reservations

Our main reservations about this grant are:

• We're highly uncertain about how deworming treatment affects worm burden over time, and how many years of deworming treatment is cost-effective. We're recommending funding an additional year of deworming mass drug administration (MDA) in all locations except a few states in India, but it's possible that additional MDA would not meet our cost-effectiveness bar if we had a more accurate understanding of counterfactual worm burden. We estimate deworming programs in both Pakistan and Nigeria are right around our funding bar, so even modest changes to our model could push them below it.
• Our take on deworming is unusual compared to other funders, and it's been several years since we have prioritized making connections in this space that might change our mind. As described above, we fund deworming programs based on evidence of income increases in adulthood. Most other deworming funders we have encountered consider deworming a health intervention and support it as part of a larger focus on neglected tropical diseases (NTDs) as a whole. However we have generally found other NTD funders place higher priority on combating diseases like onchocerciasis or lymphatic filariasis,21 leaving some funding gaps for deworming programs. This difference may result from our particular approach to modeling expected value. It's possible that we are missing important considerations in the case for or against deworming.
• We may be overfunding Deworm the World in Kenya and India. As noted above, we are unsure how soon Deworm the World will be able to exit from Kenya and India, and expect its exit timelines to be informed by future impact assessments. This grant provides funding for 2026 based on Deworm the World's current best guess about the population that will be targeted for deworming treatment in 2026 and the level of technical assistance that will be required. However, it's possible that Deworm the World will be able to wind down its technical assistance sooner than projected, or that future changes in treatment strategies will result in smaller treatment populations (and thus smaller budgets) for 2026 than currently projected. In the event that this grant provides more funding than Deworm the World requires, we expect any surplus funding would be re-allocated to other GiveWell-supported geographies.
• We're recommending funding for fairly expensive impact assessments in Kenya and Uttar Pradesh, but we're unsure how much treatment strategies might change as a result of those assessments. Both locations had impact assessments fairly recently (in 2021 or 2022), and we don't have a strong understanding of how quickly worm prevalence changes. Deworm the World will also need government buy-in to adjust treatment strategies following these impact assessments. This matters because continuing to treat locations at the same frequency despite declines in worm prevalence may not be a very cost-effective use of funds, if worm prevalence is low enough that children are at low risk of worm infections despite infrequent (or even suspended) treatment. However, Deworm the World successfully advocated for reducing treatment frequency in Kenya following the 2021 impact assessment, and has informed us that it expects these assessments to show further reductions in worm prevalence that will inform resource allocation for these programs.22 Additionally, we have reviewed Deworm the World's internal guidance for adjusting, and eventually ceasing, its technical assistance as worm prevalence declines, and believe it is broadly aligned with our values and our focus on cost-effectiveness.23

Plans for follow up

Deworm the World provides annual updates on program monitoring, results of coverage validation surveys, and costs for each program it supports. We will also continue to have quarterly calls with the Deworm the World team. We plan to use these calls to check in on:

• Are planned MDAs on track in each geography?
• Have there been any additional changes to treatment strategies (particularly in Nigeria and India given recent/upcoming impact assessments)?
• How are the pilots of the surveillance approaches in India going?
• What are the latest forecasted exit timelines for each geography?
• Has Deworm the World been able to fundraise for geographies not covered by this grant (e.g., other provinces in Pakistan; Andhra Pradesh, India; etc.)?
• How is the funding landscape for deworming evolving in each country, and globally? Have any funders entered/exited the space, or otherwise changed the scope of their funding?

Internal forecasts

For this grant, we are recording the following forecasts:

Confidence	Prediction	By time
50%	We will not renew support for Kenya in our next grant cycle24	January 2025
70%	We will renew support for Khyber Pakhtunkhwa, Pakistan in our next grant cycle	January 2025
70%	We will renew support for Cross River, Ogun, Oyo, and Rivers in our next grant cycle	January 2025
40%	We will provide exit funding for Bihar and Uttar Pradesh in our next grant cycle25	January 2025
65%	Deworm the World will recommend reducing the frequency of deworming in Oyo and Ogun following the results of the 2024 impact assessments	December 2025
75%	Deworm the World will no longer provide financial assistance (such as paying for teacher training) in Kenya by the end of 2028	December 2028

Our process

This was a relatively light-touch grant investigation as we think our main uncertainties around deworming are fairly intractable, and we thought additional staff time would be better spent on other investigations.

As part of this grant investigation we:

• Had several conversations with Deworm the World
• Reviewed Deworm the World's latest spending and treatment data, and updated our cost per child estimates for geographies close to our funding bar26
• Reviewed Deworm the World's room for more funding request27
• Conducted a light-touch review of the latest coverage evaluation surveys to check that coverage levels were roughly in line with previous years and to check for data quality concerns28
• Spoke with government partners in Nigeria and Bihar to get their qualitative perspectives on Deworm the World's support29

Sources

Document	Source
Evidence Action, Deworm the World	Source (archive)
GiveWell estimates of Deworm the World's cost per child dewormed [2023]	Source
GiveWell, "Combination Deworming (Mass Drug Administration Targeting Both Schistosomiasis and Soil-Transmitted Helminths)"	Source
GiveWell, Evidence Action's Deworm the World Initiative – August 2022 version	Source
GiveWell, Evidence Action's Deworm the World Initiative — India (January 2023)	Source
GiveWell, Evidence Action's Deworm the World Initiative — Nigeria, Pakistan, and Kenya (January 2023)	Source
GiveWell's cost-effectiveness analysis for Deworm the World [2024 renewal] (public)	Source
GiveWell's room for more funding analysis for Deworm the World [2023]	Source
Miguel & Kremer 2004	Source
Summary of GiveWell's 2020 Worm Burden Adjustment Model	Source
WHO, Guideline on control and elimination of human schistosomiasis, February 2022	Source
World Health Organization, Preventive chemotherapy in human helminthiasis	Source

• 1

  See World Health Organization, Preventive chemotherapy in human helminthiasis, p. 10.

• 2

  See World Health Organization. Preventive chemotherapy in human helminthiasis, p. 41.

• 3

  See this section of our separate report on deworming.

• 4

  The exact grant amount is $4,371,604.

• 5

  In January 2023, we made a $2.5 million grant to support Deworm the World's work in India and a $5.2 million grant to support deworming programs in Nigeria, Pakistan, and Kenya.

• 6

  We estimate lower "worm burden" in Lagos compared to other Nigerian states, and relatively high costs, and confirmed with Deworm the World that this makes sense from a qualitative standpoint.

  • Source for lower cost-effectiveness in Lagos: this cell of our most recent cost-effectiveness analysis of deworming.
  • "The worm burden adjustment is an important component of GiveWell's deworming cost-effectiveness analysis that compares the aggregate baseline worm prevalence for each charity deworming program to worm prevalence in the Miguel & Kremer 2004 deworming RCT study population, weighted by both worm species and infection intensity. . . . The worm burden adjustment, and a 13% replicability adjustment that's common to all programs, are the largest deductions to total benefits in GiveWell's deworming CEAs, and the worm burden adjustment (along with differences in cost per treatment) drives a large share of the difference in cost-effectiveness across different country programs." Summary of GiveWell's 2020 worm burden adjustment model
  • Deworm the World informed us that Lagos is more urban than other states we support and has higher rates of unprogrammed deworming (i.e. deworming happening outside of Deworm the World-supported mass deworming campaigns) and better access to water, sanitation, and hygiene (WASH) infrastructure. Deworm the World, conversation with GiveWell, October 16, 2023 (unpublished).

• 7

  An impact assessment measures the prevalence of worms in areas where mass deworming has been taking place, to inform the frequency and targeting of future rounds of deworming. WHO recommends conducting an impact assessment after 5 rounds of deworming and every three years thereafter. The budget for the impact assessment in Kenya is ~$363,000.

  • "Routine monitoring for effective coverage and evaluation of the impact of the intervention (using repeat population-based surveys conducted after five rounds of preventive chemotherapy, or more frequently with a mid-term evaluation after three rounds) should be integral parts of preventive chemotherapy programmes, to help inform the decision on changing the strategy and continuing or stopping the programme."
  • "Close epidemiological monitoring (sentinel sites surveys or mid-term evaluation every 3 years) should be established to monitor Schistosoma spp. prevalence, especially in settings previously endemic with ≥ 10% prevalence or with a history of preventive chemotherapy with praziquantel if reducing the frequency of preventive chemotherapy." WHO, Guideline on control and elimination of human schistosomiasis, February 2022, pp. 13-14.
  • "Additionally funding requested in 2025 is $363k for a follow-up impact assessment to inform treatment strategy decisions for Evidence Action exit expected potentially in 2028 pending the results of the impact assessment." GiveWell's room for more funding analysis for Deworm the World [2023], 'DtW RFMF' tab, "Rationale" column. During review of this page, Deworm the World informed us that the impact assessment in Kenya has been pushed back to early 2026.

• 8

  The full budget for the Uttar Pradesh impact assessment is $599,694, but some unspent funds from previous years will be used to cover part of the cost. Deworm the World, comments on a draft of this page, May 2024 (unpublished)

• 9

  The three states are Karnataka, Uttarakhand, and Jharkhand. See here.

  • "We are tentatively assuming that we will be exiting in-person TA and initiating remote TA focused on M&E support in 2025. We are seeking $50k to implement a cost-effective surveillance approach, to monitor disease resurgence following suspension of MDA [mass drug administration] in the state." GiveWell's room for more funding analysis for Deworm the World [2023], 'DtW RFMF' tab, "Rationale" column

• 10

  In its room for more funding submission, Deworm the World shared the following note about the increase in global indirect costs: "Evidence Action has been making strategic investments in organizational strengthening and capacities to meet the goal of doubling our impact by 2024. As a result of these critical investments, Evidence Action's global indirect cost rate has increased from 18% to 20% of direct program costs. We are seeking funds for full cost recovery of the organization's global management and administrative functions that support the effective implementation of all of our programs."

  We decided to cover the increase in global indirect costs for geographies that received GiveWell funding in 2024-2026. We adjusted the original funding requested for global indirect costs to exclude the costs attributable to geographies not supported by GiveWell funding. See our calculations here.

• 11

  We estimate worm burden by looking at the prevalence of "moderate" and "heavy" intensity worm infections in a given area. You can read more about our model here.

  "We try to consider baseline pre-treatment worm burden rather than current program worm burden, implicitly assuming that worm burden would return to baseline in the counterfactual where deworming ceased." Summary of GiveWell's 2020 Worm Burden Adjustment Model

• 12

  "In endemic communities with prevalence of Schistosoma spp. infection < 10%, WHO suggests one of two approaches based on programmatic objectives and resources: (i) where there has been a programme of regular preventive chemotherapy, to continue the intervention at the same or reduced frequency towards interruption of transmission; or (ii) where there has not been a programme of regular preventive chemotherapy, to use a clinical approach of test-and-treat, instead of preventive chemotherapy targeting a population." WHO, Guideline on control and elimination of human schistosomiasis, February 2022, p. 14.

  "The available evidence does not identify the optimal prevalence threshold for stopping preventive chemotherapy programmes, but there may be a point at which test, treat, track, test and treat may become more acceptable and efficient." WHO, Guideline on control and elimination of human schistosomiasis, February 2022, p. 19.

• 13

  The 2021 impact assessment survey in Kenya showed an 82% reduction in STH prevalence from the 2012 baseline level. Results from the 2022 impact assessment surveys in Karnataka, Uttarakhand, Jharkhand, and Uttar Pradesh are not yet public, but Deworm the World has informed us that there were large declines in worm prevalence in those geographies as well. Deworm the World, Summary of Impact Assessment Results (unpublished).

• 14
  • We have not updated our cost-effectiveness analysis for Kenya since April 2023, when we estimated deworming was approximately 35x as cost-effective as unconditional cash transfers. We believe that this is an overestimate, as it does not account for declines in worm burden over time, but we are unsure by how much.
  • We have not tried to quantify the cost-effectiveness of deworming in Bihar or Uttar Pradesh since 2020, when we estimated deworming was more than 150x as cost-effective as unconditional cash transfers. Since then, we have renewed support for these states in India on a largely qualitative basis, as we found it hard to assess the cost-effectiveness of Deworm the World's work in India since we expected the government would continue deworming regardless, and we were unsure how to quantify the impact of Deworm the World's support. However, we believed mass deworming was likely cost-effective in Bihar and Uttar Pradesh due to moderately high worm burden and relatively low cost per child. See our 2023 grant page for more information.

• 15

  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 2024, our bar for directing funding to non-Top Charity programs is 10 times as cost-effective as unconditional cash transfers. See GiveWell’s Cost-Effectiveness Analyses webpage for more information about how we use cost-effectiveness estimates in our grantmaking.

• 16

  Deworm the World confirmed the accuracy of the years in this table during its review of a draft of this page.

• 17

  The 2021 impact assessment survey in Kenya showed an 82% reduction in STH prevalence from the 2012 baseline level. Results from the 2022 impact assessment surveys in Karnataka, Uttarakhand, Jharkhand, and Uttar Pradesh are not yet public, but Deworm the World has informed us that there were large declines in worm prevalence in those geographies as well. Deworm the World, Summary of Impact Assessment Results (unpublished)

  Deworm the World informed us that it anticipates further reductions in prevalence in both Kenya and Uttar Pradesh following the next impact assessments, which would reduce the resources needed for those geographies. Deworm the World told us it only plans and budgets for impact assessments when it has strong confidence that they will result in changes to treatment strategies. Deworm the World, conversation with GiveWell, September 18, 2023 (unpublished)

• 18

  Deworm the World informed us that some areas in Kenya will now receive deworming treatment every other year rather than annually and some counties have suspended STH treatment, following analysis of the 2021-2022 impact assessment survey results which showed that some areas have reached <2% STH prevalence. Deworm the World, conversation with GiveWell, August 14, 2023 (unpublished)

  We roughly estimate this change in treatment strategy reduces Kenya's deworming budget by approximately $600,000 every other year. GiveWell's unpublished analysis of Deworm the World's room for more funding submission.

• 19
  • Deworm the World, conversation with GiveWell, August 14, 2024 (unpublished).
  • "Additionally funding requested in 2025 is $363k for a follow-up impact assessment to inform treatment strategy decisions for Evidence Action exit expected potentially in 2028 pending the results of the impact assessment." Deworm the World, 2024 RFMF & 2023 commitments, "DtW RFMF" tab. Note that during review of this page, Deworm the World informed us that the Kenya impact assessment has been pushed back to early 2026.

• 20

  The three states are Karnataka, Uttarakhand, and Jharkhand. See here.

  • "We are tentatively assuming that we will be exiting in-person TA and initiating remote TA focused on M&E support in 2025. We are seeking $50k to implement a cost-effective surveillance approach, to monitor disease resurgence following suspension of MDA [mass drug administration] in the state." GiveWell's room for more funding analysis for Deworm the World [2023], 'DtW RFMF' tab, "Rationale" column

• 21

  Deworm the World has noted that onchocerciasis and lymphatic filariasis may be higher priority for other funders due to having a clearer pathway to elimination than STH or SCH. Deworm the World, comments on a draft of this page, May 2024 (unpublished)

• 22
  • Deworm the World informed us that some areas in Kenya will now receive deworming treatment every other year rather than annually and some counties have suspended STH treatment, following analysis of the 2021-2022 impact assessment survey results which showed that some areas have reached <2% STH prevalence. Deworm the World, conversation with GiveWell, August 14, 2023 (unpublished)
  • Deworm the World informed us that it anticipates further reductions in prevalence in both Kenya and Uttar Pradesh following the next impact assessments, which would reduce the resources needed for those geographies. Deworm the World told us it only plans and budgets for impact assessments when it has strong confidence that they will result in changes to treatment strategies. Deworm the World, conversation with GiveWell, September 18, 2023 (unpublished)

• 23

  In particular, Deworm the World has informed us that although the decision to reduce treatment frequency is ultimately up to governments, it will eventually stop providing technical assistance for deworming in locations that have reached extremely low levels of worm prevalence, even if the government plans to continue mass deworming. Deworm the World, comments on a draft of this page, May 2024 (unpublished)

• 24

  This grant extends our support for deworming in Kenya through 2026. This forecast predicts the likelihood that we will not extend support beyond 2026 in our next grant cycle.

• 25

  Providing exit funding means we would provide funding for Deworm the World's work in Bihar and Uttar Pradesh beyond 2026 (likely just through 2027), but would consider that to be our last round of funding for those locations.

• 26

  We updated our cost per child estimates for Cross River, Lagos, Rivers, and Khyber Pakhtunkhwa. See GiveWell estimates of Deworm the World's cost per child dewormed [2023].

• 27

  See GiveWell's room for more funding analysis for Deworm the World [2023].

• 28

  At a high level, we did not identify any major concerns with the coverage achieved in recent deworming campaigns, and did not identify any changes in the methodology of the coverage surveys themselves that might raise concerns around data quality.

• 29

  At a high level, these government partners spoke highly of their engagement with Deworm the World. They noted that deworming would still be conducted absent its support, but that the quality of deworming campaigns would likely suffer.