Grantmaking in Response to COVID-19

Published: March 2021

Note: This page summarizes the rationale for GiveWell grants in response to the COVID-19 pandemic. Grantees reviewed this page prior to publication.

Note: Most of the text about each grant described below, including footnote text, was written in present tense shortly after the grants were made. Descriptions of interventions, funding landscape and needs, our reasons for supporting these projects, and our levels of knowledge and uncertainty were current as of the dates the grants were made. Some of the future-looking language used to describe future funding needs or milestones by which we would make decisions about further funding recommendations refers to dates that have already passed. In particular, most of this section was written in June 2020, and references to possible future funding were written before our follow-up grant to IDinsight was recommended in August 2020. We plan to follow needs that continue to emerge due to the pandemic, which may lead us to make additional recommendations in the future.

A brief note on sources for this page: Due to our short timeline for review and the limited information available around the COVID-19 outbreak and preventive strategies, our grant investigation largely consisted of phone conversations and emails that we do not expect to publish. We thus have more limited public documentation of the rationale behind these grants than we typically would share.

Grants

Innovations for Poverty Action — Randomized Controlled Trial on the Effect of Face Masks on COVID-19 (July 2020)

Our full write-up on this grant and two additional grants in support of the project is here.

IDinsight — Support for Governments Responding to COVID-19 (June 2020)

Summary

In June 2020, IDinsight received a grant of $656,000 from the Effective Altruism Global Health and Development Fund based on our recommendation. This funding was used to support IDinsight's work supporting governments responding to COVID-19. In August 2020, we recommended an additional $550,000, which IDinsight received in September, to continue this work.

These grants were part of GiveWell's work to identify COVID-19-related giving opportunities that might be comparable to our top charities in cost-effectiveness.

IDinsight has previously received several GiveWell Incubation Grants (more information here).

The organization

IDinsight provides analytical support to governments, funders, and other clients with the goal of generating rigorous evidence to improve social impact.1 This support may involve short-term engagements to support specific decisions, or longer-term engagements that require close work with clients to provide more flexible support on decisions as they arise.2

We've worked with IDinsight since 2014. Additional details on our relationship are on this page.

Relationship disclosures: We hired Dr. Neil Buddy Shah, IDinsight's CEO and Founding Partner, to serve as GiveWell's Managing Director in December 2019. Dr. Shah began working at GiveWell in July 2020. Dr. Shah was still employed at IDinsight at the time the June grant decision was made, though not when the follow-up grant decision was made in August. Dr. Shah was present in the first conversation about the grant but did not participate in subsequent conversations. He remains a member of IDinsight's Board of Directors. Additionally, Oliver Sabot, an independent consultant who is also on IDinsight's Board of Directors, briefly served as a management coach for James Snowden, who investigated this grant for GiveWell.

Planned activities

We expected that the June 2020 grant would primarily be used to support the short-term funding needs of three projects aimed at supporting governments' decisions during the COVID-19 pandemic (more below). We expected that a small portion of the grant would be used to respond to other inbound requests for support from governments.

At the time of the June 2020 grant, we also expected that we might recommend an additional grant to IDinsight to support the longer-term funding needs of these projects. That decision would depend on whether we saw indications that these projects were successfully contributing to government decision-making, and whether IDinsight had been able to raise funding from other sources. In August 2020, evidence shared with us by IDinsight led us to conclude that one of the three projects was meeting its goals; we then recommended a follow-up grant (more below).

Case for the grant

We recommended the June 2020 grant because:

1. The projects we expected this funding to support seemed promising to us.
2. We have a strong positive opinion of the quality of IDinsight's work and believe IDinsight is well suited to support governments on key decisions related to the COVID-19 pandemic.
3. Our impression is that governments are making difficult and important decisions in response to COVID-19.
4. We expected that without our support IDinsight would not have been able to pursue these projects, or would have had to significantly scale them down.

Promising nature of the projects

We expected that the majority of the June grant would be used to fill the short-term funding needs of three projects:

1. An embedded learning partnership with the government of Delhi.
2. Developing guidance on COVID-19 testing strategies in partnership with the Institute for Global Health Sciences at the University of California, San Francisco and the Clinton Health Access Initiative (CHAI).
3. Assisting the government of the Indian state of Uttarakhand to prioritize and adapt essential health services during the pandemic.

An embedded learning partnership with the government of Delhi

IDinsight is working with teams advising the Chief Minister and Health Minister in the Indian national capital territory of Delhi to establish data systems that would streamline the transfer of patient data between government and medical facilities.3

IDinsight had begun work on this project at the time of the June grant, but told us that it might need to be scaled back if it did not receive additional funding.4 IDinsight hoped this project would develop into a longer-term partnership with the Delhi government, which would require additional funding.5

We believed this project would have significant impact because:

• There appeared to be strong demand for this work from relevant stakeholders.6
• IDinsight has experience working on similar projects in India.7
• IDinsight told us that this project was its highest priority and that it was receiving substantial attention from IDinsight senior management and staff with relevant expertise.8

On the other hand:

• The project is focused on the needs of the government of the Indian national capital territory of Delhi, with a population of roughly 19 million,9 and so has limited scope for global impact.
• Longer-term, IDinsight plans to provide flexible support to the government of Delhi's COVID-19 response, and it is unclear what form this support will take.10 While we see the benefits of providing responsive analytical support, this makes ultimate impact difficult to assess.

Developing COVID-19 testing strategies in partnership with the Institute for Global Health Sciences at the University of California, San Francisco and Clinton Health Access Initiative (CHAI)

IDinsight and the Institute for Global Health Sciences at the University of California, San Francisco are working with CHAI to help governments decide how to prioritize their limited supply of COVID-19 diagnostic tests. At the time of our initial discussions with IDinsight, CHAI country teams in India and Zimbabwe had expressed interest in supporting country testing strategies further.11 CHAI told us it did not have sufficient capacity to develop a testing strategy alongside procurement, so it asked to partner with IDinsight on this project.12

The goal of this project is to develop a framework for prioritizing testing between different populations and different purposes, and assigning different kinds of tests to their most appropriate uses. IDinsight then plans to work with governments to implement that testing strategy.13

At the time of the grant, this project was ongoing but funded from IDinsight's limited supply of unrestricted funds rather than project-specific funding. Longer-term, IDinsight hopes that this project will scale to additional countries.14

We believe this project may have high impact because:

• There is widespread consensus that testing is one of the most important public health measures to reduce the spread of COVID-19.15
• Because CHAI manages some procurement of diagnostic tests, it appears well positioned to participate in decision-making on how those tests are used.16
• CHAI has told us that it has been impressed with IDinsight's work so far.17 We have also reviewed some of IDinsight's interim output and have found it high-quality.18

On the other hand:

• IDinsight told us that testing strategies in India, Zimbabwe, and other countries were in early stages, but our understanding of the current situation is limited, and we are unsure of how likely it is that this work will ultimately affect these governments' decisions.19
• IDinsight's path to impact would be mediated through CHAI, rather than directly connected to government decision-making.20 This increases the chance of diminished impact for reasons outside IDinsight's control.

Assisting the government of Uttarakhand to prioritize and adapt essential health services during the pandemic

IDinsight is planning to work with Evidence Action to redesign health programs in Uttarakhand, a state in India. IDinsight has reported a large decline in access to basic health care in Uttarakhand since the COVID-19 pandemic began,21 and Evidence Action has received a letter indicating government demand for this support.22 A description of potential activities involved in this project is available in the following footnote.23 IDinsight later told us that the scope of this project changed once work started.24

At the time of the grant, work on this project had not yet started. Longer-term, IDinsight hopes to support broader national guidelines on essential health services.25

We have less information about this project than the other two projects described above. We are uncertain about its potential for impact, but believe it addresses a real need and have seen evidence of demand from the government.26

Positive opinion of IDinsight

• We have worked with IDinsight before and have a positive impression of its work. IDinsight successfully conducted a randomized controlled trial on New Incentives, a potential future GiveWell top charity. IDinsight also completed a survey of people demographically similar to participants in programs supported by GiveWell to understand how they would make complex moral trade-offs. We were impressed with the quality of IDinsight's work in both of these engagements. Our impression is that IDinsight is well suited to the COVID-19-related projects described above, which require a combination of fast, decision-relevant analysis and academic knowledge.
• We have spoken with other funders and partners of IDinsight, who were positive about the quality of IDinsight's work27 and these opportunities.28 IDinsight has also shared positive feedback from government partners on its recent work.29

On the other hand:

• We have reviewed a narrative description of IDinsight's past work,30 but do not have a detailed understanding of its track record supporting governments.

Significant need for decision-making assistance among governments

Our impression from news reports and from conversations we have had with IDinsight and others is that the COVID-19 pandemic has required governments to make a number of difficult and important decisions on the best way to contain the spread of COVID-19 and maintain essential government functions. This has stretched their decision-making capacity. We therefore think there's a strong intuitive case that supporting government decision-making will be particularly impactful now.

Low likelihood of funding from other sources

• IDinsight told us it had a short-term funding gap for the three projects described above, and limited opportunities to raise funds for those projects in the short term.31 At the time of the grant, IDinsight was funding two of these projects with unrestricted funds, but told us it would likely have to wind them down without project-specific funding.32
• This is consistent with our understanding of IDinsight's funding model, in which the majority of funding is allocated to specific projects by funders.33
• IDinsight told us it is in conversations with other funders, but believes it is unlikely that funding will come quickly enough to support these projects' short-term needs.34
• Our impression is that there is more funding available for COVID-19-related commodities and research and development than for analytical support. This impression comes from conversations we have had with other charities and funders,35 and reviewing the objectives of large pooled funding mechanisms.36

On the other hand:

• IDinsight is in conversations with several other funders, and we are uncertain how likely IDinsight is to attract substantial additional funding in the medium term.37
• We expect that CHAI will contribute some funding toward the diagnostics project.38

Plan for follow-up

We planned to check in with IDinsight in mid-August 2020 for an update on its progress. At that point, we would consider whether to recommend additional funding to support the medium-term needs of these projects.

We did not expect to see strong evidence of impact from these projects by mid-August, but noted that we might get new information indicating that these projects were unlikely to be as impactful as we then expected. For example, governments might have indicated that IDinsight's assistance was no longer required.

We planned to make the decision to recommend additional funding based on the extent to which these projects appeared to be on the path to changing government decisions, and whether we believed IDinsight could still use additional funding.

Internal forecasts

All of the below forecasts were made at the time of the initial grant (June 2020) and have an end date of December 31, 2020, unless otherwise noted.

Confidence	Prediction	By time	Resolved?
75%	We recommend another grant to IDinsight		Yes, follow-up grant recommended in August 2020 (see below)
60%	We understand that the diagnostics project has changed the testing strategy of at least one country
20%	We see evidence that the percentage of COVID-19 patients lost to follow-up in Delhi has not declined from 35% to under 30%
40%	We see evidence that the percentage of COVID-19 patients lost to follow-up in Delhi has declined to 20-30%
10%	We see evidence that the percentage of COVID-19 patients lost to follow-up in Delhi has declined to under 20%
30%	We do not see evidence related to the proportion of COVID-19 patients being lost to follow-up in Delhi
15%	We believe IDinsight no longer has a medium-term funding gap to support these projects when we check in in August	Mid-August 2020	Yes, funding gap remained and follow-up grant made (see below)

Our process

In April 2020, we made a grant to IDinsight of $150,000 (discussed below) to provide flexible support to governments early in the pandemic. In May 2020, IDinsight sent us a proposal describing several COVID-19-related projects for which it was seeking funding.39 We had several conversations with IDinsight, including with the teams on each of the three projects described above, and reviewed documents IDinsight sent us. We also had several conversations with IDinsight's partners and other funders of IDinsight.

Additional follow-up grant (August 2020)

In August 2020, we recommended an additional grant of $550,000 to continue supporting IDinsight's work on COVID-19. IDinsight received this grant in September 2020.

IDinsight told us that $550,000 would be sufficient to cover an additional six months of support to the Delhi government, and an additional four months of the diagnostics work described above. We were most uncertain about the potential impact of the India health services delivery project, and so did not recommend additional funding to support this work.40 It is possible that IDinsight will use these funds on other activities related to its COVID-19 response if it believes those activities to be higher-impact. We consider this acceptable, given how quickly the situation is evolving.

We recommended this additional grant based on a brief update from IDinsight about how the projects were progressing. We felt most confident in the impacts of IDinsight's work with the Delhi government because we had seen indications that it was meeting its goals.41

Research ethics inquiry into IDinsight paper on pilot study

On June 29, 2020, IDinsight informed us about concerns raised on social media regarding its research ethics in a paper published using data from a small pilot study funded via GiveWell Incubation Grants and unrelated to IDinsight's work on COVID-19. The next day, IDinsight received a set of questions from the National Health Research Ethics Committee, Nigeria (NHREC).

The inquiry centers on questions about whether IDinsight had proper ethical approvals prior to starting the pilot study and whether IDinsight appropriately acknowledged its local collaborators. The NHREC has not yet issued a formal response.

We continue to have confidence in IDinsight as a research partner and believe that it is responding appropriately and in cooperation with the Nigerian public health research community. We recommended this grant to IDinsight in support of its COVID-19 work after we learned of this inquiry. We did so because the COVID-19 work this funding would support was: a) analytical, and did not involve gathering new experimental evidence; b) outside of Nigeria; and c) time-sensitive. We felt the costs of recommending funding before the NHREC had issued a formal response were outweighed by the benefits of work to reduce the effects of the pandemic.

We outline additional details on what happened and the steps we and IDinsight are taking on this page.

Development Media International — Public Health Broadcasting (April 2020)

Summary

We granted $200,000 to Development Media International (developmentmedia.net), one of GiveWell's standout charities, to support public health broadcasting in nine African countries (Burkina Faso, Côte d’Ivoire, Ethiopia, Madagascar, Malawi, Mozambique, Tanzania, Uganda, and Zambia) in response to the COVID-19 pandemic.42

The intervention

Development Media International (DMI) plans to implement or support mass media campaigns to promote behaviors that may reduce transmission of COVID-19. The behaviors targeted are:43

• Hand and respiratory hygiene: for example, promoting handwashing and coughing into the crook of the arm.
• Social distancing: encouraging people to stay two arm-lengths away from others.
• Protection of the vulnerable: encouraging at-risk individuals, such as those over 60 or with a health condition, to stay two arm-lengths away from others in their households, if possible.
• Government-specific social distancing: supplementing government advice or rules related to social distancing, if appropriate.

DMI has conducted two randomized controlled trials of the impact of mass media campaigns on behavior change.44 We discuss one of those trials in greater detail in our review of DMI, which found some evidence that DMI’s programs may lead to increases in self-reported health behaviors.45 However, limited rigorous information is available on the impact of the interventions DMI plans to promote in response to the pandemic.

The organization

DMI has been a GiveWell standout charity since 2014. More information about its work is available in our review.

Planned activities and budget

Planned activities

With this grant, DMI plans to support government COVID-19 responses, primarily those delivered over the radio,46 by:

• Supporting governments to produce high-quality messaging.47
• Directly producing media content to support government responses.48
• Buying airtime to increase coverage.49

In some countries, like Burkina Faso and Tanzania, DMI expects to take a leading role in running mass media campaigns. In others, such as Ethiopia, DMI expects to play a supporting role.50

Budget

DMI requested $250,000 on average per country for work in the nine countries listed above.51

We are providing $200,000 of flexible funding in support of DMI's COVID-19-response work in these nine countries. We are unsure specifically how it will be spent.

Case for the grant

• Common-sense case for promoting behavior change. We consulted a global development program officer at a large foundation, the CEO of a global health NGO, and health policy researchers as part of our grant review.52 They generally agreed that low-cost behavior change like respiratory hygiene and keeping distance between people is an important part of the pandemic response in countries where it is difficult to restrict people's movement.
• Potentially high cost-effectiveness. We performed a "quantitative gut check" (rougher than our typical cost-effectiveness estimates) that suggested this intervention may be comparable to or more cost-effective than where we would otherwise direct marginal funds.53 We were more uncertain about our gut check than our usual estimates for a number of reasons. For example:
  • There are no rigorous studies on whether mass media campaigns reduce transmission of COVID-19.
  • We relied on the predictions of the Imperial College epidemiological model linked in the following footnote to estimate how many lives would be saved by reducing transmission.54 That model doesn't factor in a number of considerations, such as higher rates of comorbidity or lower rates of oxygen availability in the areas in which DMI operates.55 We tried to subjectively adjust for these considerations.

  These uncertainties could lead to higher or lower cost-effectiveness. We don't know whether our gut-check figure is likely to be an underestimate or overestimate.

• DMI is an organization we know well and have a positive opinion of. DMI has been a GiveWell standout charity since 2014 and we generally think highly of it. As with all standout charities, we do not have as much confidence in its work as we do our top charities.
• Time-sensitivity. We understand that reducing COVID-19 transmission is likely to be more effective earlier in the pandemic.56 During our grant review, DMI told us that it expected a lack of flexible funding to hamper its ability to respond.57
• Potential leverage. GiveWell's support for this work may attract other funders. However, it may cause other funders to give less by lowering DMI's perceived financial needs.

Risks and reservations

• Other funders may have supported this work if we hadn't. DMI raised $405,000 of the total $2.25 million it sought ($250,000 each for nine countries) and was in discussions with other funders for another $625,000-$685,000 at the time we made our grant.58 We understand that significant funding is being directed toward pandemic relief in general. DMI may have raised this funding from others if GiveWell had not provided it. In other words, our contribution may have been fungible.
• Changes in typical process for testing messaging. DMI told us that it didn't expect to test its messaging in response to the pandemic through its usual in-person focus groups. It may be able to find alternative ways to test messaging.59
• Less information about this grant overall than is typical. We made this grant with less information than we typically would require:
  • Our quantitative gut-check analysis is particularly uncertain, as noted above.
  • We don't have a detailed budget for the grant.
  • We are unsure on which margin the funds will be used. DMI is raising flexible funding to support work with governments in nine countries. DMI has framed its request as $250,000 per country on average. We don't expect its costs to be uniform across countries, as it may play a different role in different locations.60
  • We don't know whether DMI's messaging will be additional. Radio is an important source of media in the countries in which DMI plans to work.61 Some or all of these countries may also use mobile phone messages, billboards, and other means of spreading public health messages.

    Radio messaging may still have an effect when people receive messages from multiple sources. In our quantitative gut check, we attempted to account for other avenues through which people may hear these messages. But we are uncertain how much this will occur and how it will impact the effectiveness of the campaigns DMI runs.

Plans for follow-up

In addition to our regular check-ins due to its status as a standout charity, we plan to stay in touch with DMI over the coming months about its COVID-19 response. We may make a larger grant to DMI if it needs additional funding for pandemic-response work and we believe that marginal funding in support of this work is more cost-effective than where we would have otherwise directed the funds.

Our process

Our process for identifying this grant opportunity is described in this blog post.

IDinsight — Support for Data Collection and Analysis for Government Partners (April 2020)

Summary

We granted $150,000 to IDinsight (IDinsight.org), a development research and advisory group with which we partner closely, to support COVID-19-related data collection and analysis for government partners.

The intervention

IDinsight plans to provide additional staffing to assist governments with their COVID-19 response and to increase the amount of data IDinsight can collect to inform that response. IDinsight may be able to leverage this work to raise additional funding for similar projects.62

The organization

We discuss IDinsight as an organization above, as well as a relationship disclosure that we hired Dr. Neil Buddy Shah, IDinsight's CEO and Founding Partner, to join GiveWell as Managing Director. Dr. Shah provided input on the April grant in his role as IDinsight's CEO.

Planned activities and budget

The $150,000 grant is unrestricted. We expect that it will be divided between:63

1. Staff costs, including the reallocation of senior staff and technical team capacity to COVID-19-related analytical projects and policy recommendations, such as how to structure and target cash transfer programs ($100,000).
2. Data collection to support governments ($50,000).

Case for the grant

• Policy is promising. Governments are making decisions that could have a large impact on the eventual burden of COVID-19. For example, they are deciding whether, when, and how to implement social distancing, how to adapt essential health services, and how to target livelihood programs.64

  Improving government policies could be a particularly effective way to reduce the impact of COVID-19. Based on our initial investigations into programs to assist government policymakers, we believe that this type of work may offer strong returns on investment.65

• We know IDinsight well and think highly of it. We have been working with IDinsight since 2014 and have a positive opinion of its analytical capabilities.
• Government relationships. We understand that IDinsight has relationships with governments in India, Ghana, Malawi, Morocco, the Philippines, and Zambia, and has received requests for assistance from those governments on COVID-19-related projects.66
• Potential leverage. IDinsight may raise additional funding by demonstrating how it can assist governments with COVID-19 response.

Risks and reservations

• Less information about this grant overall than is typical. Due to the short timeline on which we made this grant, we had less information about it than we typically would require before making a grant.

Plans for follow-up

At the time we made the grant, we remained in contact with IDinsight about potential funding needs for COVID-19 response. We noted that we might provide additional support going forward if we believed that marginal funding in support of this work was more cost-effective than where we would have otherwise directed the funds.

Our process

Our process for identifying this grant opportunity is described in this blog post.

Yale Research Initiative on Innovation and Scale — Support for Government of Bangladesh Response (April 2020)

Summary

We granted $100,000 to a multidisciplinary team at the Yale Research Initiative on Innovation and Scale (Y-RISE), led by Professor Mushfiq Mobarak, to support the government of Bangladesh's COVID-19 response.

The intervention

Professor Mobarak and his team expect to use this grant to support the Bangladesh government's response to the COVID-19 pandemic.

When we made the grant, we understood that it would most likely be used for one of the following purposes:67

1. Supporting the salary of an analyst to help respond to COVID-19-related policy requests from the Bangladesh government.
2. Scaling up mobile phone surveys of migrants in Bangladesh to collect region-specific data on COVID-19 risk.

We thought the most likely impact of the grant would be supporting an analyst, as we understood this to be a higher priority of Professor Mobarak and because he expected it would be harder to raise funds for analyst support.68

The organization

Y-RISE researches the effects of delivering policy interventions at scale. It is led by Professor Mobarak. We worked with Professor Mobarak on our assessment of former GiveWell top charity Evidence Action's No Lean Season, a Y-RISE flagship project. We have a positive opinion of Professor Mobarak's work.

Planned activities and budget

See the section on the intervention above for details.

Case for the grant

• Positive opinion of Professor Mobarak. We have a positive opinion of Professor Mobarak's work and his deep knowledge of the Bangladesh context. Our impression is that he has strong relationships with a2i, a Bangladesh cross-department government group that Professor Mobarak told us is coordinating the government response to COVID-19 between departments.69
• Time-sensitivity. We understood the scale-up of mobile phone surveys to be highly time-sensitive, as they were intended to track migrants who were returning from urban to rural areas due to the pandemic. Supporting an analyst (by renewing their contract) was also time-sensitive, but less so.70
• Longer-term opportunities. We think it is likely that there will be additional opportunities to fund Professor Mobarak's COVID-19-related work going forward. Making this grant could improve our ability to build a relationship, follow his work, and potentially make future grants.
• Potential leverage. This funding may allow Professor Mobarak to demonstrate projects that will help him raise more funding.

Risks and reservations

• Subjective case for the grant. Professor Mobarak's expertise and connections are the primary reason we made this grant. Our assessment of these is subjective and prone to bias.
• Less information about this grant overall than is typical. We didn't have a detailed understanding of the analytical work Professor Mobarak would be conducting with grant funding. We knew far less about this grant than we would a typical GiveWell grant.
• Other funders may have supported this work if we hadn't. We understand that significant funding is being directed to the COVID-19 response. Professor Mobarak may be able to raise funding from others, making our contribution fungible.

Plans for follow-up

We remain in contact with Professor Mobarak about potential funding needs for COVID-19 response. We may provide additional support going forward if we believe that marginal funding in support of this work is more cost-effective than where we would have otherwise directed the funds.

Our process

Our process for identifying this grant opportunity is described in this blog post.

Blog posts on COVID-19 grantmaking

We published blog posts on our plans for and decisions about making grants in response to the COVID-19 pandemic. They are available here.

Sources

Document	Source
Centers for Disease Control and Prevention 2020	Source (archive)
Centers for Disease Control and Prevention, "Testing Strategy for Coronavirus (COVID-19) in High-Density Critical Infrastructure Workplaces after a COVID-19 Case Is Identified"	Source (archive)
Centre for Effective Altruism, EffectiveAltruism.org, "EA Funds: Global Health and Development Fund"	Source (archive)
DMI, comments to GiveWell on draft grant page, July 31, 2020	Unpublished
DMI: COVID-19 Communications Response in Sub-Saharan Africa 2020	Source
Dr. Neil Buddy Shah, CEO and Co-Founder, IDinsight, conversation with GiveWell, April 10, 2020	Unpublished
Esther Hsu Wang, Founding Partner, IDinsight, comments to GiveWell on draft grant page, July 20, 2020	Unpublished
Exchange Rates UK, British Pound to US Dollar Spot Exchange Rates for 2020	Source (archive)
Gentilini et al. 2020	Source (archive)
GiveWell back-of-the-envelope cost-effectiveness analysis of DMI, April 2020	Source
GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020	Source
GiveWell's non-verbatim summary of a conversation with Development Media International, March 26, 2020	Source
GiveWell's non-verbatim summary of a conversation with Esther Wang and Paul Wang, May 28, 2020	Source
IDinsight, "About Us"	Source (archive)
IDinsight, "Additional Notes on the Delhi COVID-19 Opportunity"	Unpublished
IDinsight, "Concept Note: Opportunities to Support IDinsight's COVID-19 Response," redacted	Source
IDinsight, "COVID-19 Response and Recovery in New Delhi, India: Data Systems and Policy Support for the Chief Minister's Office," redacted	Source
IDinsight, "Responses for GiveWell COVID Diligence," updated June 3, 2020, redacted	Source
IDinsight, "Responses for GiveWell COVID Diligence," updated June 5, 2020, redacted	Source
IDinsight, "What We Do: Services"	Source (archive)
IDinsight, COVID-19 testing strategy framework overview (draft), May 21, 2020	Unpublished
Mobarak and Barnett-Howell 2020	Source (archive)
Professor Mushfiq Mobarak, Founder and Faculty Director, Y-RISE, conversation with GiveWell, April 13, 2020	Unpublished
Professor Mushfiq Mobarak, Founder and Faculty Director, Y-RISE, conversation with GiveWell, April 9, 2020	Unpublished
Unique Identification Authority of India, Aadhaar saturation and projected 2020 population by state or territory	Source (archive)
Walker et al. 2020	Source (archive)
Walker et al. 2020, Appendix data sources	Source (archive)
World Health Organization, "COVID-19 Solidarity Response Fund for WHO"	Source (archive)
World Health Organization, Access to COVID-19 Tools (ACT) Accelerator, 2020	Source (archive)
World Health Organization, Laboratory testing strategy recommendations for COVID-19: interim guidance, 2020	Source (archive)
Yugal Kishore Pant, Mission Director, National Health Mission of Uttarakhand, letter to Evidence Action, May 27, 2020	Unpublished

• 1

  "IDinsight is a global advisory, data analytics, and research organization that helps development leaders maximize their social impact. We tailor a wide range of data and evidence tools, including randomized evaluations and machine learning, to help decision-makers design effective programs and rigorously test what works to support communities. We work with governments, multilaterals, foundations, and innovative non-profit organizations in Asia and Africa. We work across a wide range of sectors, including agriculture, education, health, governance, sanitation, and financial inclusion." IDinsight, "About Us"

• 2

  See IDinsight, "What We Do: Services", under the headings "Work with Us" and "Services," for more details.

• 3

  "IDinsight is the Government of Delhi’s official technical partner for tackling the COVID-19 crisis for more than 20 million people. IDinsight has an embedded team within the government that is working directly and on a daily basis with the Delhi Chief Minister’s team. We are tasked with establishing data systems that streamline data from COVID-19 treatment facilities to inform emergency decision-making by the Chief Minister, and providing tailored policy support on other government priorities." IDinsight, "COVID-19 Response and Recovery in New Delhi, India: Data Systems and Policy Support for the Chief Minister's Office," redacted

• 4

  On June 12, IDinsight told us that if it did not secure additional funding, its work on the Delhi learning partnership would continue for at least another week, but the likelihood of any work beyond that was uncertain. Due to capacity constraints, the project was being staffed in part by two IDinsight alumni on a pro bono basis, while also requiring considerable time from the organization's India lead and founding partner, Ronald Abraham. IDinsight anticipated that this staffing arrangement would be unsustainable without additional funding. Esther Wang and Paul Wang, Founding Partners, IDinsight, conversation with GiveWell, June 12, 2020 (unpublished)

• 5

  "The government has also requested nimble and responsive policy support (second service offering) over a longer duration for their evolving decision-making needs. Beyond data systems, IDinsight’s broad methodological toolkit and in-house expertise of economists and public health specialists can help answer a range of questions facing the government. Here is an indicative list of questions we will [could] seek to answer for the Delhi Government:

  • How does the government ensure that other life-saving health services, such as immunization and TB treatment, aren’t displaced by the COVID-19 response?
  • What are the most effective ways of operating schools keeping in view public health needs?
  • What are the most effective ways of COVID19-proofing economic interactions in markets, neighbourhoods and other locations?
  • What are the knowledge, awareness and behavioural levels regarding COVID-19 and physical distancing in Delhi’s population, and how do we raise these levels?
  • How does the government balance health and livelihoods considerations in its decision-making on physical distancing and containment policies?"

  IDinsight, "COVID-19 Response and Recovery in New Delhi, India: Data Systems and Policy Support for the Chief Minister's Office," redacted

  Edits in square brackets were corrections suggested by IDinsight on a draft of this page after the grant was made.

• 6

  IDinsight told us that its previous work on a project for the Delhi education department enabled it to connect with the Delhi health department earlier in the COVID-19 pandemic and offer its services. At the time, according to IDinsight, Delhi's COVID-19 burden was among the highest in India. IDinsight was later contacted by the Delhi government and set up a call with a senior advisor to the Chief Minister who was leading the territory's data systems work. Through this relationship, IDinsight began to help with automating reports of COVID-19 data, which evolved into the project currently under way.

  IDinsight is now the official technical partner of the government of Delhi on COVID-19 response, and has received a letter affirming this relationship from the Dialogue & Development Commission of the Delhi government, as well as a memorandum of understanding that it intends to sign. IDinsight believes it is the only external organization currently working out of the Delhi government offices, and noted that some of the research fellows from Delhi's legislative assembly had occasionally been deputized to help IDinsight with data cleanup and other tasks. Ronald Abraham and Paul Wang, Founding Partners, IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 7

  "PRE-COVID: India: Our India government efforts are headlined by several substantial efforts.

  • We have a 20+ person team embedded in NITI Aayog, the policy advisory unit of the Government of India, that supports national policy efforts in agriculture, financial inclusion, health / nutrition, and sanitation. Using Data on Demand capabilities (described below), the ELP also provides the data foundation for India’s Aspirational Districts Programme which aims to improve socio-economic outcomes in 117 poor districts (250 million people) through a outcomes-based competitive management approach (i.e. use of data, competition and collaboration to improve outcomes).
  • IDinsight serves as the Program Management Unit to establish India’s National Data and Analytics Platform that will democratize access to and use of India’s public government data. This includes mapping the government data landscape and gaps, documenting user interests and needs, prototyping user interfaces, and facilitating interaction between disparate data sources.
  • IDinsight’s Data on Demand (DoD) platform makes accurate, representative data on almost any topic multiples faster and cheaper. DoD provides representative data for 140 million people in India via a 37,000 household panel and >1,000 local surveyors. DoD makes possible the Aspirational Districts Programme and has served many other needs including urgent information demands for India’s COVID-19 response (described more below).
  • IDinsight has executed several other ELPs with India government units, including governments of Andhra Pradesh, Himachal Pradesh and Bihar."

  IDinsight, "Responses for GiveWell COVID Diligence," updated June 5, 2020, redacted

• 8

  Ronald Abraham, a founding partner and the head of IDinsight's team in India, told us that he is spending more than 50% of his time on the project, and an economist on staff is devoting nearly all his time to it. Several other staff members are working part-time on the project, including Divya Nair, a director and public health specialist; two research associates; the data science director and one member of his team; and the monitoring systems director and one member of his team. Though it is investing considerable resources, IDinsight has so far refrained from dedicating a full team to the project due to lack of confirmed funding. Ronald Abraham and Paul Wang, Founding Partners, IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 9

  Unique Identification Authority of India, Aadhaar saturation and projected 2020 population by state or territory

• 10

  "The government has also requested nimble and responsive policy support (second service offering) over a longer duration for their evolving decision-making needs. Beyond data systems, IDinsight’s broad methodological toolkit and in-house expertise of economists and public health specialists can help answer a range of questions facing the government. Here is an indicative list of questions we will [could] seek to answer for the Delhi Government:

  • How does the government ensure that other life-saving health services, such as immunization and TB treatment, aren’t displaced by the COVID-19 response?
  • What are the most effective ways of operating schools keeping in view public health needs?
  • What are the most effective ways of COVID19-proofing economic interactions in markets, neighbourhoods and other locations?
  • What are the knowledge, awareness and behavioural levels regarding COVID-19 and physical distancing in Delhi’s population, and how do we raise these levels?
  • How does the government balance health and livelihoods considerations in its decision-making on physical distancing and containment policies?"

  IDinsight, "COVID-19 Response and Recovery in New Delhi, India: Data Systems and Policy Support for the Chief Minister's Office," redacted

  Edits in square brackets were corrections suggested by IDinsight on a draft of this page after the grant was made.

• 11

  After getting an overview of work from all CHAI country directors, IDinsight began with an initial deep dive on these two countries. IDinsight chose to focus on India and Zimbabwe because these countries had expressed the greatest interest in and need for this type of support. Currently IDinsight does not have the staff capacity to support discussions with more than these two countries; however, the project could be expanded in the future to include any of the countries where CHAI works. Alison Connor, Director of Health at IDinsight, and Alice Redfern, Manager at IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 12

  CHAI told us that IDinsight offered an unusual set of skills that were uniquely complementary to CHAI's own, and that it would have been difficult to find a similarly qualified partner to do this work if IDinsight had not been available. Originally, the CHAI diagnostics team sought out former CHAI employees as consultants, but was not able to find enough of them. Trevor Peter, Senior Director of Medical Diagnostics, CHAI, conversation with GiveWell, June 12, 2020 (unpublished)

• 13
  • "In four weeks, IDinsight has created new tools to craft optimal LMIC [low- and middle-income country] diagnostic strategies and conducted in-depth analyses which demonstrate the infeasibility of the predominant contact tracing and population sample testing tactics for LMICs. We are now at a critical stage where we will engage policymakers in specific countries (short list countries are India, South Africa, Zimbabwe, and Senegal) to help craft optimal national COVID-19 diagnostic strategies and develop generalizable resources for all LMICs." IDinsight, "Concept Note: Opportunities to Support IDinsight's COVID-19 Response," redacted
  • IDinsight told us in a June 8 call that South Africa and Senegal are no longer necessarily on the "short list" and that future countries could include any of those where CHAI works. Alison Connor, Director of Health at IDinsight, and Alice Redfern, Manager at IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 14

  Currently IDinsight does not have the staff capacity to support discussions with more than two countries; however, the project could be expanded in the future to include any of the countries where CHAI works. Alison Connor, Director of Health at IDinsight, and Alice Redfern, Manager at IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 15
  • "Early experience from COVID-19 outbreaks in a variety of settings suggests that when symptomatic workers with COVID-19 are identified, there are often asymptomatic or pre-symptomatic workers with SARS-CoV-2 present at the workplace. Testing is important to identify such individuals, as they may not know they are infected. SARS-CoV-2 transmission from asymptomatic or pre-symptomatic persons can result in additional cases and potentially outbreaks of COVID-19. Implementing screening for symptoms of COVID-19, testing, and contact tracing may be used to detect infected workers earlier and exclude them from the workplace, thus preventing disease transmission and subsequent outbreaks." Centers for Disease Control and Prevention, "Testing Strategy for Coronavirus (COVID-19) in High-Density Critical Infrastructure Workplaces after a COVID-19 Case Is Identified"
  • "As part of the Strategic Preparedness and Response Plan, WHO developed testing strategy recommendations. The foundation of this strategy is threefold:
    • All countries should increase their level of preparedness, alert, and response to identify, manage, and care for new cases of COVID-19; laboratory testing is an integral part of this strategy.
    • Countries should prepare to respond to different public health scenarios, recognizing that there is no one-size-fits-all approach to managing cases and outbreaks of COVID-19.
    • Each country should assess its risk and rapidly implement the necessary measures at the appropriate scale and prepare for a testing and clinical care surge to reduce both COVID-19 transmission and economic, public health, and social impacts."

    World Health Organization, Laboratory testing strategy recommendations for COVID-19: interim guidance, 2020

• 16

  Of the $18 million grant CHAI received from Unitaid, $14.5 million was allocated to procurement of commodities, including diagnostic tests. Trevor Peter, Senior Director of Medical Diagnostics, CHAI, conversation with GiveWell, June 12, 2020 (unpublished)
  

• 17

  Trevor Peter of CHAI reported a positive impression of the "wide range of expertise" demonstrated by IDinsight staff during a call. Trevor Peter, Senior Director of Medical Diagnostics, CHAI, conversation with GiveWell, June 12, 2020 (unpublished)

• 18

  IDinsight, COVID-19 testing strategy framework overview (draft), May 21, 2020
  

• 19

  IDinsight told us that India's strategy includes testing anyone in the population who meets certain criteria, but there are few guidelines, at either the state or national level, for adjusting the process if testing within these groups or within certain contexts proves unfeasible. IDinsight told us that Zimbabwe's current strategy is less well formed. While the Zimbabwean government has strategized on where seroprevalence testing would be used, it currently lacks the testing capacity it needs to put any antibody testing plan in action. Alison Connor, Director of Health at IDinsight, and Alice Redfern, Manager at IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 20

  IDinsight told us that so far it has primarily worked through CHAI's country teams instead of communicating directly with government officials in India and Zimbabwe. IDinsight described this gap between its team and government leaders as one of the biggest risks to the success of this project. Alison Connor, Director of Health at IDinsight, and Alice Redfern, Manager at IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 21

  Raghav Adlakha, Mansi Jain, Divya Nair, and Paul Wang, IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 22

  Yugal Kishore Pant, Mission Director, National Health Mission of Uttarakhand, letter to Evidence Action, May 27, 2020

• 23
  1. "Prioritize health areas and operations to prevent interruption. IDinsight - in collaboration with Evidence Action - will use its deep, India-specific expertise in public health, economics and data science to create bespoke models that estimate healthcare gains, transmission risks and economic costs associated with different service delivery mechanisms. To maximize efficiency, initial consultations and desk research will be used to prioritize areas for deep investigation. Through this, we will understand the costs and benefits of pursuing select delivery mechanisms. For example, we will be able to estimate net DALYs prevented per rupee under different immunization delivery mechanisms. We may find that while immunization camps are the most cost-effective way to deliver vaccines, the COVID-19 transmission risk associated with large gatherings of susceptible people may countervail their benefits. Through our models, we will be able to estimate the interaction rate, gathering size and participant profile that allows us to optimise immunization delivery.
  2. Health service redesign: Once we have identified the trade-offs associated with different service delivery models, we will develop alternative mechanisms to deliver essential services with high implementation fidelity within operational and resource constraints. Efforts will focus on the most actionable areas with high impact likelihood in consultation with multiple stakeholders. This includes activities to 1) adjust or integrate existing delivery mechanisms and 2) incubate innovative delivery mechanisms to plug gaps flagged by the epidemiological models. This effort will leverage Evidence Action implementation expertise and Evidence Action / IDinsight stakeholder networks.
  3. Support redesigned delivery model implementation: Once we have designed optimized delivery mechanisms, we will help the government establish associated implementation protocols that include elements such as training guidelines, coordination structures, supply chain practices, and monitoring systems. Delivery innovations will be piloted to serve proofs of concept for wider adoption. This phase will heavily leverage Evidence Action at-scale implementation expertise.
  4. Monitor implementation to enable implementation pivots and refinements. Using IDinsight’s rapid data collection and analytics expertise, we will lightly monitor the implementation fidelity of these services. Using Evidence Action’s technology system expertise, administrative data can also be used to assess initial outcomes and validate epidemiological model parameters. Rigorous experiments (e.g. RCTs) will be considered for delivery interventions with high impact potential, generalizability and operational requirements. All data and evidence will be consistently fed back to state policymakers for continual design and implementation refinement."

  IDinsight, "Concept Note: Opportunities to Support IDinsight's COVID-19 Response," redacted

• 24

  Paul Wang, Founding Partner, IDinsight, email to GiveWell, February 18, 2021 (unpublished)

• 25

  IDinsight hopes to complete three specific goals in the first six months of the project: (1) complete its cost-effectiveness model for this work and plan out delivery, (2) implement pilot programs for three identified interventions in two to three districts in Uttarakhand, and (3) complete monitoring and evaluation work on those three interventions. The larger aim of accomplishing these three goals will be to create a set of standard operating procedures and an implementation toolkit that can be used in additional settings and service areas. In the 12 months after that, IDinsight hopes to scale up its essential health services work, both geographically (expanding throughout Uttarakhand and ultimately into additional states in India) and by moving into additional service areas, such as postnatal care and malaria net distribution. IDinsight hopes to begin a conversation with national-level government decision-makers within the first six months, with a final goal of integrating the practices developed in Uttarakhand into national guidelines by the end of 18 months. Raghav Adlakha, Mansi Jain, Divya Nair, and Paul Wang, IDinsight, conversation with GiveWell, June 8, 2020 (unpublished)

• 26

  Yugal Kishore Pant, Mission Director, National Health Mission of Uttarakhand, letter to Evidence Action, May 27, 2020

• 27

  CHAI told us that IDinsight offered an unusual set of skills that were uniquely complementary to CHAI's own, and that it would have been difficult to find a similarly qualified partner to do this work if IDinsight had not been available. Originally, the CHAI diagnostics team sought out former CHAI employees as consultants, but was not able to find enough of them. Trevor Peter of CHAI also reported a positive impression of the "wide range of expertise" demonstrated by IDinsight staff during a call. Trevor Peter, Senior Director of Medical Diagnostics, CHAI, conversation with GiveWell, June 12, 2020 (unpublished)

• 28

  A representative from another funder told us they believed these projects were valuable opportunities because governments are unusually motivated to pursue cost-effective interventions related to COVID-19, and are under significant time pressure to do so. They also noted that the value of governments making marginally better decisions about essential health services, data collection, and diagnostic strategy would be high and unlikely to change over time. They expressed a positive view of IDinsight's alignment with GiveWell's values and its capacity for flexible response, and shared examples of IDinsight's past successful work with governments.

• 29

  IDinsight quoted one representative of a government COVID-19 task force who called IDinsight's work in that country (which consisted of designing social distancing measures and mitigating the economic impacts of the virus) very helpful, and expressed openness to getting more support from IDinsight in the future. A government representative in another country, which received a policy brief from IDinsight on the development sector's response to COVID-19, praised the brief and said that it would be among the first materials to play a key role in determining the country's path. IDinsight also noted three other countries in which its recommendations were quickly adopted or informed policy decisions. Paul Wang, Founding Partner, IDinsight, email to GiveWell, May 27, 2020 (unpublished)

• 30

  See IDinsight, "Responses for GiveWell COVID Diligence," updated June 5, 2020, redacted.

• 31

  IDinsight, "Responses for GiveWell COVID Diligence," updated June 5, 2020, redacted

• 32

  In a June 12 conversation, IDinsight explained that due to lack of flexible funding, if it received no further project-specific funding in the short term, it would continue the Delhi learning partnership and the diagnostics project for at least another week, and it would not begin work on the India health services project. Esther Wang and Paul Wang, Founding Partners, IDinsight, conversation with GiveWell, June 12, 2020 (unpublished)

• 33

  GiveWell's non-verbatim summary of a conversation with Esther Wang and Paul Wang, May 28, 2020

• 34

  IDinsight, "Responses for GiveWell COVID Diligence," updated June 5, 2020, redacted

• 35
  • For example, one charity representative told us that their organization had been most successful in fundraising for tangible direct service provision related to COVID-19, such as distributing soap and disinfectants to rural communities.
  • A representative from a leading private foundation told us that the foundation was unlikely to support policy-related COVID-19 work in countries other than those where it already had strong relationships.

• 36

  The primary fund for supporting COVID-specific responses in LMICs at the time of our investigation was the WHO COVID-19 Solidarity Response Fund, which has directed the bulk of funds raised "to WHO to procure and distribute essential commodities and coordinate response" (World Health Organization, "COVID-19 Solidarity Response Fund for WHO"). Smaller amounts from this fund have been allocated to vaccine research and development, protection of refugees and internally displaced people, shipment of commodities, and support for vulnerable communities. As our investigation was coming to a close, WHO launched the Access to COVID-19 Tools (ACT) Accelerator, a collaboration among several large funders "to accelerate the development, production and equitable access to new COVID-19 diagnostics, therapeutics and vaccines" (World Health Organization, Access to COVID-19 Tools (ACT) Accelerator, 2020).

• 37

  IDinsight, "Responses for GiveWell COVID Diligence," updated June 5, 2020, redacted

• 38

  CHAI told us that it would likely redirect some existing funding toward the diagnostics project, but has not yet determined how much. CHAI currently expects that the funding it provides will be less than half of the project's short-term needs. Trevor Peter, Senior Director of Medical Diagnostics, CHAI, conversation with GiveWell, June 12, 2020 (unpublished)

• 39

  See IDinsight, "Concept Note: Opportunities to Support IDinsight's COVID-19 Response," redacted.

• 40

  Paul Wang, Founding Partner, IDinsight, email to GiveWell, August 14, 2020 (unpublished)

• 41

  IDinsight shared data with us suggesting that the proportion of COVID-19 cases which were uncontactable in Delhi had declined substantially between early June and mid-July. This was one of the goals of IDinsight's work in Delhi. Ronald Abraham, Founding Partner, IDinsight, email to GiveWell, August 12, 2020 (unpublished)

• 42

  "Development Media International (DMI) proposes to conduct national radio campaigns to support the health communications response for COVID-19 in Burkina Faso, Côte d’Ivoire, Ethiopia, Madagascar, Malawi, Mozambique, Tanzania, Uganda and Zambia, where we have experienced in-country teams and strong relationships with both Ministries of Health and radio stations." DMI: COVID-19 Communications Response in Sub-Saharan Africa 2020, Pg 1.

• 43

  "DMI's messaging will promote three categories of behaviors to mitigate the spread of COVID-19:

  • Respiratory hygiene and handwashing – The messaging will warn against coughing on others and promote instead coughing into a tissue, disposing of the tissue, and washing hands, or coughing into a handkerchief, returning the handkerchief to a pocket, and washing hands, or coughing into the crook of an arm. The messaging will promote regular handwashing, as well as awareness that surfaces, including hands, are contaminated, thus stressing the importance of handwashing before touching one's face.
  • Social distancing – The messaging will advise cutting down severely on social contacts and keeping two arms' lengths away from as many people as possible outside one's household.
  • Protection of the vulnerable – The messaging will advise that people over 60 or those with health conditions stay two arms' lengths away from everyone, including people in their own households.

  "These recommendations align with guidance from the Centers for Disease Control and Prevention (CDC) and World Health Organization (WHO). In each country, DMI will tailor messaging to ensure it aligns with government communications, which may diverge slightly from WHO guidance.

  "DMI might incorporate promotion of government-mandated behaviors, such as adherence to a curfew, as well as individual behavior change. This would depend to some extent on how well established DMI's presence is in a given country. Where a curfew has been implemented, the government will likely immediately broadcast the news on its own radio station, and DMI could supplement that messaging if it seemed appropriate. DMI could possibly play the role of reassuring the population that a curfew is a legitimate way to reduce social interaction and not a sign of repression, but DMI may also decide not to amplify government messages if it believed they were harmful." GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020, Pg 3.

• 44"DMI is the first organisation to demonstrate through two randomized controlled trials (RCTs) that mass media interventions can change health behaviours." DMI: COVID-19 Communications Response in Sub-Saharan Africa 2020, Pg 1.

• 45

  “From 2011 to 2015, DMI conducted a randomized controlled trial (RCT) of its program in Burkina Faso to measure impact on behavior change and child mortality. At midline, DMI found some positive effects on key self-reported health behaviors that suggested DMI’s program may cause a substantial reduction in child mortality.” From GiveWell’s review of DMI.
  

• 46"Development Media International (DMI) proposes to conduct national radio campaigns to support the health communications response for COVID-19 in Burkina Faso, Côte d’Ivoire, Ethiopia, Madagascar, Malawi, Mozambique, Tanzania, Uganda and Zambia, where we have experienced in-country teams and strong relationships with both Ministries of Health and radio stations." DMI: COVID-19 Communications Response in Sub-Saharan Africa 2020, Pg 1.

• 47

  "Following WHO guidelines, our UK research experts will provide content advice and oversight of the final messages to be broadcast." DMI: COVID-19 Communications Response in Sub-Saharan Africa 2020, Pg 3.

• 48

  DMI, comments to GiveWell on draft grant page, July 31, 2020

• 49

  "DMI believes that a three-month COVID-19 health messaging campaign would cost approximately $200,000–$250,000 per country—with funding used to support production, staff costs, and airtime." GiveWell's non-verbatim summary of a conversation with Development Media International, March 26, 2020, Pg 3.
  

• 50

  "DMI believes that given sufficient funding, it can work in all nine countries it is targeting; however, its role in each country will be different. Ethiopia, for example, is a large country with powerful media, and DMI's presence in the country is smaller and newer than that of other organizations. For this reason, DMI believes it is best suited to a quality control role in Ethiopia. In Burkina Faso or Tanzania, on the other hand, DMI may take responsibility for the entire COVID-19 media campaign.

  "It is unlikely that DMI would run the entire campaign in all nine countries. The most likely outcome, if DMI receives enough funding for campaigns in all nine countries, is that it will run the campaigns in two or three countries, make a significant contribution to the campaigns in five or six countries, and make a small contribution to the campaigns in one or two countries." GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020, Pg 7.

• 51"The costs of delivering 3 month campaigns in Burkina Faso, Côte d’Ivoire, Ethiopia, Madagascar, Malawi, Mozambique, Tanzania, Uganda and Zambia are varied, but average at approximately $250,000 USD per country." DMI: COVID-19 Communications Response in Sub-Saharan Africa 2020, Pg 2.
• 52

  We do not plan to publish sources for these conversations.

• 53

  Our gut check suggested that the cost per life saved by donating to DMI's COVID-19-response work would be roughly $600. GiveWell back-of-the-envelope cost-effectiveness analysis of DMI, April 2020, cell E48. We often round estimates of the cost per life saved to the nearest thousand, to reflect the level of precision we believe our model can estimate.

  We compared this to our estimate of the cost-effectiveness of providing marginal funds to Malaria Consortium's seasonal malaria chemoprevention program, which we estimated to be slightly less cost-effective than the DMI COVID-19-response work in April 2020. This internal Malaria Consortium estimate has not been published.

  Note that our cost-effectiveness analyses are simplified models that do not take into account a number of factors. There are limitations to this kind of cost-effectiveness analysis, and we believe that cost-effectiveness estimates should not be taken literally, due to the significant uncertainty around them. We provide cost-effectiveness estimates (a) for comparative purposes and (b) because working on them helps us ensure that we are thinking through as many of the relevant issues as possible.

  Our “gut check” is at a much earlier stage than, and therefore is not directly comparable to, the cost-effectiveness analyses of our top charities. As a general rule, our estimates of a given program's cost-effectiveness tend to go down as we gain more information.

• 54

  Walker et al. 2020

• 55
  • The Imperial model estimates are based on “age-specific IFR [infection fatality ratio] from China.” Walker et al. 2020, Pg 7.
  • “It is important to note, however, that these estimates assume no substantive difference in general health/co-morbidity prevalence between Chinese and other populations.” Walker et al. 2020, Pg 8.
  • The model does take into account lower rates of intensive care capacity in low- and middle-income countries. (Intensive care capacity is described in Walker et al. 2020, Pg 6, Figure 3: Estimates of Hospital Bed and ICU Capacity, Stratified by World Bank Income Group.) It does not, however, take into account whether patients have access to basic healthcare such as oxygen.

• 56

  "If a suppression strategy is implemented early (at 0.2 deaths per 100,000 population per week) and sustained, then 38.7 million lives could be saved whilst if it is initiated when death numbers are higher (1.6 deaths per 100,000 population per week) then 30.7 million lives could be saved. Delays in implementing strategies to suppress transmission will lead to worse outcomes and fewer lives saved." Walker et al. 2020, Pg. 2

• 57"DMI frequently receives requests from governments to participate in campaigns and must be able to respond within 24 hours or less, which likely would not allow time to get permission from a funder. In situations such as these that depend on a fast turnaround, it is more cost-effective for funders to allow DMI to decide where its funding should go at that moment." GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020, Pg 7.
  
• 58
  • "DMI has raised $130,000 for a three-month COVID-19 campaign in Burkina Faso. It has also raised $275,000, from the Mulago Foundation and an individual philanthropist who came to DMI via The Life You Can Save, for a Rapid Response Fund that will allow it to respond quickly to government requests from the nine countries where it intends to conduct COVID-19 campaigns.
  • "DMI is also in talks with several other funders. Wellcome has invited DMI to complete a formal application for funding, which could result in an additional £250,000, and DMI is planning to submit a proposal to the Global Innovation Fund (GIF) for another £250,000 to £300,000." GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020, Pgs 6-7.
  • Exchange Rates UK, British Pound to US Dollar Spot Exchange Rates for 2020 indicates that on April 9, 2020, the exchange rate was one British pound to about US$1.25, so £500,000 to £550,000 would be equivalent to about $625,000 to $687,500.

• 59"DMI has been exploring ways to adapt its process of designing messages, as it is currently unable to assemble in-person focus groups. DMI may be able to use networks from its radio station partners, which often have panels of listeners that the stations may allow DMI to contact. These listeners could help with pre-testing of DMI's messaging and/or answering formative research questions, although conducting this research by phone will likely be challenging." GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020, Pg 6.
  
• 60

  "It is unlikely that DMI would run the entire campaign in all nine countries. The most likely outcome, if DMI receives enough funding for campaigns in all nine countries, is that it will run the campaigns in two or three countries, make a significant contribution to the campaigns in five or six countries, and make a small contribution to the campaigns in one or two countries." GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020, Pg 7.

• 61

  "Radio is the dominant medium in sub-Saharan Africa." GiveWell's non-verbatim summary of a conversation with Development Media International, April 9, 2020

• 62Dr. Neil Buddy Shah, CEO and Co-Founder, IDinsight, conversation with GiveWell, April 10, 2020

• 63Dr. Neil Buddy Shah, CEO and Co-Founder, IDinsight, conversation with GiveWell, April 10, 2020

• 64

  For a discussion of social distancing, see Mobarak and Barnett-Howell 2020; for a discussion of adapting health services, see Centers for Disease Control and Prevention 2020; and for a discussion of livelihood programs, see Gentilini et al. 2020.
  

• 65

  "For example, we believe programs to assist government policymakers may offer strong returns on investment, based on our initial investigations into policy organizations." GiveWell blog, "Three grants in response to the COVID-19 pandemic," April 21, 2020

• 66Dr. Neil Buddy Shah, CEO and Co-Founder, IDinsight, conversation with GiveWell, April 10, 2020 and Esther Hsu Wang, Founding Partner, IDinsight, comments to GiveWell on draft grant page, July 20, 2020

• 67Professor Mushfiq Mobarak, Founder and Faculty Director, Y-RISE, conversation with GiveWell, April 9, 2020
  
• 68Professor Mushfiq Mobarak, Founder and Faculty Director, Y-RISE, conversation with GiveWell, April 13, 2020
  
• 69Professor Mushfiq Mobarak, Founder and Faculty Director, Y-RISE, conversation with GiveWell, April 9, 2020. We have also spoken with several other funders who believe that Professor Mobarak has strong relationships with decision-makers in Bangladesh.
  
• 70Professor Mushfiq Mobarak, Founder and Faculty Director, Y-RISE, conversation with GiveWell, April 9, 2020