Note: This page summarizes the rationale behind a GiveWell grant to Johns Hopkins University. Johns Hopkins University staff reviewed this page prior to publication.
In a nutshell
In March 2025, GiveWell recommended a grant of up to $843,502 to Johns Hopkins University to fill an emergency funding gap for an ongoing randomized trial of the Cholera Hospital-Based Intervention for 7 Days (CHoBI7) in Bangladesh.
We think that CHoBI7 and PICHA7 (Preventative Intervention for Cholera for 7 Days), a related program in the DRC, are among the most promising interventions we’ve identified for diversifying our water grantmaking portfolio beyond our current focuses on chlorine dispensers and in-line chlorination. This is because CHoBI7 and PICHA7 support chlorine use during a critical window in which the families of diarrhea patients are at the highest risk of infection, and because two previous trials have shown strong results.
This grant supports an ongoing randomized trial that tests a streamlined version of the program, which will help us to understand which version of the intervention is best-suited for implementation at scale. Our grant directly replaces USAID funding that was withdrawn from the trial six months after it had begun, during the USAID funding cuts.
We’re primarily recommending this grant because:
- As above, we think this research could help us learn about what we think could be an especially promising way to diversify our water grantmaking portfolio;
- The value of what we’ll learn from this study, plus the impact of the study on informing the possible scale-up of the program in Bangladesh, places this grant at our cost-effectiveness threshold;
- Our grant allows researchers to fulfil their commitments to treatment group participants, and preserves the value of the funding already invested by USAID;
- We think that funding this will develop our relationship with researchers working on this program in both Bangladesh and the DRC, where we would expect this program to look especially cost-effective (due to its large health burden).
Our main reservations are that:
- With deeper investigation, we might conclude that these interventions are less cost-effective than we currently think;
- Although the role of this study in informing the scale-up of the program in Bangladesh is a core part of our cost-effectiveness analysis for this grant, we haven’t externally triangulated claims about the likelihood of this, due to the time-sensitivity of this funding opportunity;
- We think it’s possible that Open Philanthropy would have provided this funding in our absence;
- The trial will need to be formally terminated and then restarted, due to USAID termination requirements. This could harm the trial, although the researchers think this issue will not be insurmountable.
Published: June 2025
Background
The Cholera Hospital-Based Intervention for 7 Days (CHoBI7) is an intervention designed by a research group led by Christine Marie George at the Johns Hopkins School of Public Health.1
The intervention is designed to reduce the rate of diarrheal disease among household members of diarrhea patients admitted to health facilities, by promoting (a) handwashing with soap and (b) water treatment to the patients and their household members over a one-week period after the patient is admitted to a health facility. The research group associates this one-week period with a much higher risk of household members contracting diarrheal diseases, likely due to shared, contaminated water sources, and poor hygiene practices.2 (The name references the program’s use of picture-based education: “chobi” means picture in Bangla.)
The standard CHoBI7 package includes:3
- A visit by a health promoter to a diarrhea patient in hospital;
- The distribution of a ‘diarrhea prevention package’, including a one-month supply of chlorine tablets, a water bottle containing water and detergent, a handwashing station, and a vessel designed for safe water storage;
- Voice and text messages about water and sanitation behaviour over a period of 12 months, plus instructions to boil water once the chlorine tablets are used up.
An alternative version of CHoBI7 also includes two home visits from a health promoter in the week after the patient is enrolled in the program.4 CHoBI7 is also closely related to another program, PICHA7, designed by the same research team. PICHA7 has been trialled in the Democratic Republic of the Congo (DRC), and includes additional chlorine tablets and home visits.5
CHoBI7 was evaluated in a randomized trial conducted in urban Dhaka, Bangladesh, from December 2016 to April 2019.6 PICHA7 was tested in urban Bukavu, in South Kivu, DRC, from December 2021 to 2023.7 These RCTs demonstrated significant reductions in diarrhea prevalence (around 18% reduction in children under five) and stunting.8
In 2020, USAID’s Development Innovation Ventures awarded a $1.5m grant to support a randomized evaluation in Bangladesh of a less expensive, streamlined version of CHoBI7, in which the handwashing station and water vessel are not provided (instead, program beneficiaries are encouraged to purchase or build those materials themselves).9 The primary outcomes of the evaluation are handwashing, the presence of soapy water, and water quality (e.coli and free chlorine levels), measured at 3 month follow-up.10 This evaluation was designed to inform the potential national scale-up of the program by the Bangladesh government’s Ministry of Health and Family Welfare.11
In March 2025, this $1.5m USAID grant was terminated, during the USAID funding cuts. At the point at which the grant was terminated, the trial had been running for six months.12 Professor George’s research team had received $420,000 of the grant funds, with a further $440,000 uncollected and due to be paid. This left the trial with an emergency funding gap of $640,000.
The grant
GiveWell’s grant is designed to fill the emergency funding gap for this trial, which will allow it to be completed as originally planned. The grant consists of an initial $668,787, which will cover costs for the remainder of the study, from February 2025 until December 2026. This initial sum does not cover the cost of any work completed to date, although the research team currently has $174,715 in spent funds due to be received in arrears from USAID. If this payment does not come through, GiveWell has agreed to provide this funding, in order to allow the trial to be completed. This would take the total cost of our grant to $843,502.
The case for the grant
We are recommending this grant because:
We think that CHoBI7 and PICHA7 are promising interventions that could help to diversify GiveWell’s water grantmaking portfolio.
One of the core goals of GiveWell’s water grantmaking team for 2025 is to diversify its grantmaking beyond the two core programs that it has funded at-scale to date (chlorine dispensers and in-line chlorination; more here). Based on our general understanding of the sector, we think that programs like CHoBI7 and PICHA7 are among the most promising candidates we’ve seen for expanding our grantmaking focus. This is because:
- They target a high-risk population (diarrhea or cholera patients and their household members) during a critical window in which we would expect them to be most vulnerable,13 and potentially most receptive to an intervention that requires behavior change;
- They incorporate chlorination (usually the distribution of chlorine tablets in hospitals, and through post-discharge home visits),14 which has well-established mechanisms for reducing diarrhea transmission and mortality;15
- They can be adapted to different settings and implementation methods (including scaling through governments), since program components are easily modified (from a more intensive program with multiple home visits, as with PICHA7,16 to a more streamlined program that only includes the provision of chlorine tablets and text message reminders);
- They have demonstrated promising results to date in RCTs in both Bangladesh and the DRC (including 20-60% reductions in diarrhea prevalence and reductions in stunting17 ), although we plan to research this evidence base in more detail in the near future.
Our best guess is that a more intensive version of the program, like PICHA7, may be where we find the most promising grantmaking opportunities in the future. However, we’re also open to the idea that different versions could make sense in different settings, or that lighter-touch versions might be easier to scale within government health systems.
By funding this trial, we expect to learn about the viability of the lighter-touch model. Because we currently speculate that this is less likely to be the most cost-effective version of the program, we would still plan to consider evidence generation grantmaking opportunities around a heavier-touch version, before considering larger grants for the program assessed in this RCT.
Under two different scenarios, this grant looks above GiveWell’s cost-effectiveness bar.
Our back-of-the-envelope cost-effectiveness calculation for this grant is based on two types of impact: (a) the impact of the study on informing the Government of Bangladesh’s decision to scale-up the program in the future (which we refer to as ‘leverage’), and (b) the impact of the study on informing GiveWell’s grantmaking (which we refer to as ‘value of information’). We value these impact streams as follows:
- Effect on scale-up in Bangladesh (leverage): 7x as cost effective as cash transfers (or 24x in an optimistic scenario). We think this research could encourage the Bangladeshi government to scale up the program with its own funding, since the researchers claim that officials have expressed interest if the results of the study are positive. In this scenario, we would assume that CHoBI7 is more cost-effective than what the government would otherwise fund. For more, see our leverage calculations here.
- Effect on GiveWell’s grantmaking (value of information): 3x as cost-effective as cash transfers in the DRC only, and 10x cash for Bangladesh and the DRC combined (or 21x in an optimistic scenario). This estimates the impact from providing GiveWell with more information about the cost-effectiveness of the program assessed in this trial, which could lead us to co-fund the scale-up of this program with support from the governments of either Bangladesh or the DRC. Note that this calculation assumes that this trial does not inform our view of PICHA7, since we think that the two programs are sufficiently different that a negative update from this trial would not affect our interest in generating additional evidence on PICHA7. For more, see our value of information calculations here.
Based on these figures, this grant would be around our cost-effectiveness bar of 10x cash transfers either if (a) the Bangladesh government decided to scale the program, and GiveWell concentrates on identifying grantmaking opportunities in the DRC, or (b) if GiveWell pursued grantmaking opportunities in both Bangladesh and the DRC.
We think there are a number of other benefits to funding this trial.
- Our funding will allow the researchers to maintain ethical obligations to trial participants. The abrupt termination of the trial would, absent new funding, mean that the researchers were unable to provide individuals in study’s treatment arm with the intervention that they promised. We think this raises ethical concerns, and we are also concerned about the impact on trust in health research of cancelling trials mid-process.
- Our funding preserves the value of the work and funding already invested. USAID has already invested $420,000 into this study, and data collection is already underway.18 This grant ensures that these resources do not go to waste.
- By funding this trial, we think we’ll develop our relationship with the research team, who we think are implementing promising programs in both Bangladesh and the DRC that GiveWell may wish to consider funding in the future. In particular, we understand that the research team is working in likely higher-burden areas of the DRC, such as South Kivu, where we think PICHA7 could look especially cost-effective. We think that funding this trial will make it marginally easier for us to dig into these interventions and scope potential future grantmaking opportunities.
Risks and reservations
Our main reservations about this grant are:
- With further investigation, we may conclude that CHoBI7 and PICHA7 are less cost-effective than we currently think. Our research into these programs prior to recommending this grant has been limited, due to the time-sensitive nature of this funding opportunity. It’s possible that we are missing something, or that with more time our understanding of these interventions will materially shift. However, we think this is an acceptable risk given the relatively small size of this grant.
- We’re not sure how likely the Bangladesh government is to scale the program, if the study shows promising results. Our understanding that the government is interested in program scale-up is based on the claims of the research team, who have said that the Ministry of Health and Family Welfare has collaborated on the design of the program assessed in this trial, and who report having worked with the ministry for over a decade.19 However, we (a) haven’t externally verified this due to the time-sensitivity of the funding opportunity (though we factor this uncertainty into our leverage estimate), and (b) are unsure whether a program implemented by the government would be as effective as one delivered by the research team.
- We would be more interested in new evidence on a PICHA7 or CHoBI7-style program in the DRC than in Bangladesh. This is because we associate the DRC with a higher health burden and a higher share of children under five (who are at the greatest mortality risk), and therefore with higher program cost-effectiveness. If we pursue implementation funding beyond this grant, it is likely that we will focus on the DRC, or other similarly high-burden countries. As above, our leverage estimate takes this limitation into account.
- It’s possible that Open Philanthropy would have provided this funding in our absence. Open Philanthropy has already committed to covering the $440,000 in funds invoiced to USAID but not yet received, so we think it is possible they would have covered our contribution in addition.20 However, we aren’t significantly concerned about this because we think that the counterfactual value of Open Philanthropy’s funds is high, and because we see intrinsic benefit in working closely with the researchers ourselves, and in diversifying our water grantmaking portfolio.
- There could be practical issues associated with restarting the trial. Due to USAID’s requirements for terminated programs, this trial will need to be officially terminated and restarted under new funding. As a result, the trial will need to be re-registered, and either previously enrolled participants will need to re-consent to participate in the trial, or new participants will need to be enrolled. The researchers have told us that this will not significantly impact the trial, on the basis that (a) trial enrollment is still in early stages, and (b) it is currently the low season for diarrheal disease in Bangladesh, which means that relatively few participants (around 40-50 households) are currently in active follow-up.21
Plans for follow up
We plan to schedule check-in calls around key milestones in the trial, approximately every six months. In these calls, we will try to learn about:
- Trial implementation: whether the research team has encountered any operational challenges, or issues with re-registering the trial or re-consenting those enrolled in the study;
- Timeline: whether the study remains on-track to be completed by the end of 2026;
- Government engagement: whether the Ministry of Health and Family Welfare has, or looks likely to, make concrete commitments around program scale-up;
- Whether there are opportunities to implement or evaluate a similar program in the DRC.
Once we have received the final RCT results, we will update our back-of-the-envelope calculation of CHoBI7’s cost-effectiveness, based on what we learn about costs and the effects of the program on diarrhea. This could inform additional grantmaking for the implementation of a program like CHoBI7 in other settings in the future.
Our process
- In 2024, GiveWell ran a request for proposals (RFP) for research to inform our grantmaking (alongside Open Philanthropy). Professor George submitted a separate proposal to this RFP, which introduced us to CHoBI7 and PICHA7.
- When we learned about the USAID funding cuts, we conducted a fast investigation into this grant opportunity, given the time-sensitive nature of the funding request. We received the previous RCT results for both CHoBI7 and PICHA7, created a back-of-the-envelope cost-effectiveness model for both programs in both Bangladesh and the DRC, had two conversations with Professor George and members of her research team, reviewed the trial protocol and budget, and coordinated with Open Philanthropy r.e. our and their contributions to the cost of the evaluation.
Sources
- 1
See George et al. 2019.
- 2
“During the time a patient with diarrhea presents at a health facility for treatment, the household members of the patients are at a much higher risk of developing diarrheal diseases (>100 times for cholera) than the general population. This risk is highest during the 7-day period after the patient is admitted to the health facility. This is likely because of a shared contaminated water source and poor hygiene practices in the home … In an effort to develop a standard of care to reduce diarrhea among household members of diarrhea patients, our research group developed the Cholera Hospital-Based Intervention for 7 Days (CHoBI7) … This targeted WASH intervention focuses on promoting handwashing with soap and water treatment to diarrhea patients and their household members during the 1-week period after the patient is admitted to the health facility.” George et al. 2021.
- 3
"The CHoBI7 mHealth program is initially delivered during a health facility visit by a health promoter bedside to a diarrhea patient and their accompanying household members…A diarrhea prevention package containing the following items is also provided: a 1-month supply of chlorine tablets…water and detergent powder, a handwashing station, and a water vessel with a lid and tap to ensure safe water storage. Households are instructed to boil their drinking water once their supply of chlorine tablets is completed, and encouraged to make more soapy water after their provided bottle is finished. After health facility delivery of the program, diarrhea patient households receive weekly voice and text messages from the CHoBI7 mHealth program over a 12-month period." George et al. 2021, e2561-e2562.
- 4
"The mHealth with home visits arm households received the same activities as the mHealth with no home visits arm plus two – 30 min home visits by a health promoter during the first week of intervention delivery." George et al. 2019
- 5
- "This may be because of the higher intensity of the PICHA7 program which had quarterly in-person intervention visits over a 12 month period compared to only two in-person visits during the 7-day high risk period after diarrhea patient admission in the CHoBI7 program." George et al. 2024 (preprint)
- Professor George confirmed our understanding during a review of this grant page.
- 6
"This study was a three-arm cluster RCT conducted in urban Dhaka, Bangladesh from December 4, 2016–April 26, 2019." George et al. 2022
- 7
"The RCT of the PICHA7 mHealth program enrolled 2334 participants in Bukavu, South Kivu province, DRC from December 2021 to December 2023." Sanvura et al. 2025 (preprint)
- 8
This is the main effect of the program on diarrhea in children under 5 in the mHealth + no home visits arm of the CHoBI7 trial. We use the results from this arm since (i) they were similar to the more intensive arm and (ii) this is the arm modeled in the cost-effectiveness analysis from which we derive costs.
- "Compared to the standard message arm, children <5 years of age had significantly lower 12-month diarrhea prevalence in the mHealth with 2 home visits arm (prevalence ratio [PR]: 0.73 [95% CI: 0.61-0.87]) and the mHealth with no home visits arm (PR: 0.82 [95% CI: 0.69-0.97]) (Table 2). This impact was stronger for children <2 years of age in both the mHealth with 2 home visits arm (PR: 0.69 [95% CI: 0.58-0.83]) and the mHealth with no home visits arm (PR: 0.78 [95% CI: 0.65-0.93]). For all age groups combined (children and adults), the diarrhea prevalence was 29% lower in the mHealth with 2 home visits arm (PR: 0.71 [95% CI: 0.60-0.84]), and 18% lower in the mHealth with no home visits arm (PR: 0.82 [95% CI: 0.69-0.97])."
- "Children <2 years were significantly less likely to be stunted in both the mHealth with 2 home visits arm (33% vs 45%; OR, 0.55 [95% CI: .31–.97]) and the mHealth with no home visits arm (32% vs 45%; OR, 0.54 [95% CI: .31–.96]) compared with the standard message arm after adjustment for baseline growth measures."
- 9
- "Fiscal Year Awarded: 2020…Award Ceiling: $ 1,500,000."
- "With support from Development Innovation Ventures (DIV), Johns Hopkins is conducting a randomized controlled trial with to test a scalable version of the CHoBI7 program that examines…“How large are impacts when beneficiaries receive chlorine tablets and soapy water for free, but must purchase and build the other materials themselves?"
USAID, Testing a Scalable Model of the CHoBI7 Mobile Health Program in Bangladesh.
- 10
"Study Design. Four hundred households with diarrhea patients will be recruited and prospectively followed for 12 months (1600 participants total, 4 per household)....The primary outcomes are: (1) handwashing with soap (by structured observation and spot checks); (2) chlorine (proxy of water treatment practices) and E.coli in stored drinking water; and (3) the presence of soap or soapy water in the cooking and latrine areas of households." Johns Hopkins School of Public Health, Development Innovation Ventures Application Form, 2020, p. 20 (unpublished)
- 11
"During our proposed Stage 2 award we developed the below draft scale-up strategy for delivery of the CHoBI7 program in Bangladesh in partnership with the Bangladesh Ministry of Health and Family Welfare." Johns Hopkins School of Public Health, Development Innovation Ventures Application Form, 2020, p. 16 (unpublished)
- 12
Professor George confirmed our understanding during a review of this grant page.
- 13
"Household contacts of cholera patients are at a 100 times higher risk of becoming infected with cholera compared with the general population during the 1-week period after the index cholera patient obtained care at a health facility." Burrowes et al. 2017
- 14
"The CHoBI7 intervention includes: … a diarrhea prevention package containing chlorine tablets." Johns Hopkins School of Public Health, Development Innovation Ventures Application Form, 2020, p. 6 (unpublished)
- 15
For more of our previous writing on water chlorination, see our Water Quality Interventions report here.
- 16
"...The higher intensity of the PICHA7 program which had quarterly in-person intervention visits over a 12 month period..." George et al. 2024 (preprint)
- 17
PICHA7 diarrhea prevalence: “Among all participants (all ages), 12-month diarrhea prevalence was significantly lower (diarrhea prevalence over the 12-month study period assessed through monthly visits) in the PICHA7 program arm (Prevalence Ratio (PR): 0.39 (95% Confidence Interval (CI): 0.32, 0.48) compared to the standard arm (Table 2). The impact on 12-month diarrhea prevalence was similar for children 0-4 years (PR: 0.38, 95% CI: 0.31, 0.46) and children 0-1 years (PR: 0.43, 95% CI: 0.35, 0.53)).” George et al. 2024 (preprint)
CHoBI7 diarrhea prevalence: “Compared to the standard message arm, children <5 years of age had significantly lower 12-month diarrhea prevalence in the mHealth with 2 home visits arm (prevalence ratio [PR]: 0.73 [95% CI: .61–.87]) and the mHealth with no home visits arm (PR: 0.82 [95% CI: .69–.97]) (Table 2). This impact was stronger for children <2 years of age in both the mHealth with 2 home visits arm (PR: 0.69 [95% CI: .58–.83]) and the mHealth with no home visits arm (PR: 0.78 [95% CI: .65–.93]). For all age groups combined (children and adults), the diarrhea preva-lence was 29% lower in the mHealth with 2 home visits arm (PR: 0.71 [95% CI: .60–.84]), and 18% lower in the mHealth with no home visits arm (PR: 0.82 [95% CI: .69–.97]).” George et al. 2021, e2563.
PICHA and CHOBI stunting: The effect sizes are:- PICHA7: 55% lower odds of stunting in the intervention arm (52% vs 63%, OR: 0.45, 95% CI: 0.21, 0.98). George et al. 2024 (preprint)
- CHoBI7: 45% lower odds in the home visits arm (33% vs 45%, OR, 0.55 [95% CI: .31–.97]), 46% lower odds in the no home visits arm (32% vs 45%; OR, 0.54 [95% CI: .31–.96]). George et al. 2021
- 18
John Hopkins University, Overall Budget, February 2025 (unpublished)
- 19
Professor George confirmed our understanding during a review of this grant page.
- 20
John Hopkins University, Overall Budget, February 2025 (unpublished).
- 21
This is based on several conversations with Professor George, and her view of this grant page.