University of California, Berkeley — In-Line Chlorination Research (May 2025)

Note: This page summarizes the rationale behind a GiveWell grant to the University of California, Berkeley. UC Berkeley staff reviewed this page prior to publication.

Published: August 2025

The organization

The grant will support the work of Dr. Amy Pickering, the Blum Center Distinguished Chair in Global Poverty and Practice at the University of California, Berkeley.1 Dr. Pickering has extensive experience with in-line chlorination research and implementation. She led a randomized controlled trial of ILC in Bangladesh,2 and co-developed one of the chlorination devices being tested in this grant, the TuriTap.3 Dr. Pickering’s research team has already conducted preliminary ILC installations in Nigeria and built relationships with local water authorities.4

The grant includes a sub-award to Self Help Africa, a nonprofit organization that will serve as the implementing partner in Nigeria. Self Help Africa works on rural development programs across sub-Saharan Africa and has existing water infrastructure programs in Nigeria.5

The intervention

We think that water quality is one of the most promising grantmaking areas for GiveWell, primarily due to the evidence of its effect on reducing under-five child mortality. Our full report on water quality interventions can be found here.

In-line chlorination (ILC) is one of the two chlorination interventions to which GiveWell has directed significant funding to date, alongside chlorine dispensers.6 Unlike other chlorination methods, such as dispensers or chlorine vouchers, which require users to add chlorine to their water manually, ILC works automatically, without requiring behavior change from water users. The devices are typically installed on a pipe feeding a water storage tank. As water flows past the chlorinator, it disinfects the water.7

There are a wide variety of commercially available ILC devices.8 This grant will test two types: tablet erosion chlorinators, which GiveWell has previously funded at scale in Malawi and India, and the TuriTap liquid injection chlorinator, a new device developed by Dr. Pickering’s team.9

The grant

This grant provides funding for two main activities: (i) a small-scale pilot cluster randomized trial of ILC in two states in Nigeria (Kano and Cross River) with Self Help Africa as the implementing partner, and (ii) research on chlorine acceptability in Nigeria and India.

Cluster randomized controlled trial in Nigeria

The research team will assess 100 communities across Kano and Cross River states, identifying 40 with ILC-compatible water infrastructure.10 Of these, 20 would be randomly assigned to receive ILC installations in compatible water sources, with the device type (either a TuriTap or tablet-based doser) potentially also to be randomly selected.11 The installations are expected to cover 16,000 households, or 80,000 people, based on an average of two devices per community.12

Data collection will include:13

• Baseline surveys of 15 households per community on water usage patterns, including secondary sources of drinking water, and perceptions of chlorination;
• Three follow-up survey rounds at 4, 8, and 12 months post-installation, of 600 household surveys each;
• Weekly monitoring of chlorine concentrations throughout the 12-month study period
• Water quality testing of 5 households per community for chlorine residual and microbiological contamination;
• Analog water meters, which will be installed to objectively measure consumption and validate self-reported usage.14

Chlorine acceptability research

This research includes a number of activities designed to assess chlorine taste thresholds and ways to increase chlorine acceptability, in order to inform the most appropriate chlorine dose for implementation in India and Nigeria. These activities are as follows:

• A taste test experiment to see if users can identify rice fermented with chlorinated water in Odisha, India. Through a previous research grant, we learned that community resistance to chlorination was higher than expected and residents thought chlorinated water was inhibiting fermentation of a popular local rice dish, pakhala. The researchers would enroll 150 individuals from both treatment and control villages in Odisha and ask participants to do a blind taste test of pakhala samples.15
• Desk research on other cultural practices or cuisines that could be affected by chlorination, beyond pakhala.16 This research should help to contextualize the pakhala findings and determine whether this is likely to be an issue in other areas.
• A series of experiments designed to determine the taste detection threshold for chlorinated water, and identify whether prolonged exposure to chlorinated water influences this threshold. These studies would enroll 300 participants in both Nigeria and India and conduct experiments in which participants are given several drinking water samples, and asked to identify the one that has chlorinated water, and whether they find it acceptable to drink. This research will be powered to detect a taste acceptability threshold difference of 0.18 mg/L.17

Budget and timeline

This $2.1 million grant will fund research activities over 2.5 years.18

The budget breaks down as follows:

• University of California, Berkeley – $1,798,03219
  • Nigeria cluster RCT – approximately $1,200,000 (community assessments, ILC device installations, surveys, and monitoring)
  • Nigeria chlorine acceptability research – approximately $270,000
  • India chlorine acceptability research – approximately $340,000
• Self Help Africa – $340,000
  • Nigeria cluster RCT – $340,000 (implementation support and water meter installations)

The timeline for the grant is as follows:20

• Month 6 (November 2025) – Launch field pilot trials in Nigeria
• Months 12-14 (May-July 2026) – Interim results from Kano on ILC performance and acceptability
• Month 16 (September 2026) – Interim results from Cross River
• Months 17-18 (Oct-Nov 2026) – Taste experiment results from both countries
• Months 18-20 (Nov 2026-Jan 2027) – Draft mortality trial proposal submission
• Month 24 (May 2027) – Final revised mortality trial proposal
• Months 24-30 (May-Nov 2027) – Final analysis and preparation for potential trial launch

The case for the grant

We are recommending this grant because:

• We think that through this research, we will meaningfully advance toward a randomized trial of chlorination that is powered to detect child mortality effects. This is one of the top priorities of our water grantmaking team. This research directly addresses barriers that prevented us from funding such a trial previously:
  • Establishing whether there is sufficient treatment contrast: Statistical power in any trial depends on achieving meaningful differences in the uptake of the intervention between treatment and control groups (or ‘treatment contrast’). The pilot cluster RCT will allow for measurement of treatment contrast across seasons, while accounting for source-switching behavior.
  • Testing passive, rather than active chlorination methods: Previous preparatory work for the mortality trial has included work on chlorine vouchers, which have lower take-up rates (and therefore treatment contrast) than ILC.21 Since ILC is automatic (whereas vouchers must be redeemed and chlorine consciously applied), we would expect high treatment contrast to be more likely with an ILC-focussed trial, particularly since we assume that baseline chlorination rates in Nigeria are very low.22
  • Identifying a viable implementation partner: At present, a lack of proven implementers remains a key constraint both for the mortality trial and for funding in-line chlorination at scale. This research would involve working with Self Help Africa, and will allow us to evaluate their potential as an implementing partner for the full trial.
  • Assessing the feasibility of the trial in promising settings: We think that Kano and Cross River states could be promising locations for the mortality trial, due to Kano’s high under-five mortality rate,23 and Self Help Africa’s strong existing presence in both Kano and Cross River.24 This research will help us assess whether there are enough ILC-compatible water points in the two states to make a well-powered trial feasible.
  • Concrete path to a fundable proposal: Dr. Pickering’s team plans to submit an initial proposal for a full mortality trial around 18 to 20 months after the start of the grant, with a final, revised proposal by month 24 (see timeline above). This increases our confidence that this work will culminate in the mortality trial that we would like to fund.
• Beyond the mortality trial, this research is likely to inform our future in-line chlorination grantmaking decisions. We think this research could be informative for future grant opportunities that we plan to consider, including pilot proposals that have come through our chlorination request for information. For example, we’ll learn about:
  • ILC devices: by using tablet dosers and the TuriTap side by side, we’ll learn about their relative effectiveness, and we’ll develop more sophisticated implementation protocols for their usage on different types of infrastructure. Data on chlorine concentrations, refill frequency, installation difficulty, and operational costs should allow future implementing organizations to choose the most appropriate technology for their contexts.
  • Chlorine dosing: The taste threshold research directly addresses an implementation challenge we’ve observed in previous work, including in India with the pakhala, as above. This research will establish: (a) precise chlorine dose targets for specific contexts, (b) evidence on whether gradual dosage increases can improve acceptance, and (c) rough heuristics on which cultural contexts (within Nigeria and India) require additional sensitization, and which are immediately receptive to chlorination. We think this is important because implementers currently have to make dosing decisions with minimal data; this research could improve the efficacy of in-line chlorination programs.
  • Infrastructure compatibility: This work will include an assessment of whether there are enough ILC-compatible water points in Kano and Cross River to make either a mortality trial feasible or extensive scale-up possible,25 by triangulating Self Help Africa’s data with data from local government authorities. A lack of compatible water points is a potential barrier to successful chlorination programs, so we expect this work to be informative for our grantmaking plans in the future.
  • Cultural acceptance: The taste research will help to determine whether there are cultural practices that pose significant barriers to the implementation of chlorination at scale, and could help future implementing organizations to target areas with higher chlorine acceptance, or plan for additional sensitization in areas with lower acceptance.
• The research fits within our broader research agenda on ILC’s barriers to scale. We think that in-line chlorination is one of the most promising water quality interventions, but our understanding is that it is still a relatively new approach to chlorination, with a number of potential barriers to rapid scale-up. This research could inform our understanding of these barriers in both India and Nigeria, and potentially suggest ways to overcome them.
• We think that Dr. Pickering and her team are uniquely well-positioned to lead this research. Dr. Pickering is a leading ILC expert, with background on our previous work in support of a potential mortality trial. Her team also has existing relationships with key stakeholders, including the water authority in Kano, and Self Help Africa. We think these factors increase the likelihood that this research results in a promising proposal for a larger mortality trial.

Risks and reservations

Our main reservations about this grant are:

• There’s a risk this represents continued, incremental funding for a mortality trial that may not happen. To date, we have invested $2.7 million across two grants26 for mortality trial preparatory research, and we initially expected to fund the study by Spring 2024.27 However, a proposal for the trial did not materialize as a result of this initial work, primarily due to lower-than-expected treatment contrast.28 We’re concerned that this grant could also fail to result in a trial proposal, although we think this risk is mitigated in this case because (a) there is a concrete deliverable for a proposal by month 20, (b) we have a strong qualitative impression of Dr. Pickering’s team, and (c) the research directly addresses issues encountered in previous work, as above.
• We’re uncertain about the practical impact of the taste acceptance research. While taste aversion is frequently cited as a barrier to chlorination scale-up by experts,29 and surfaced during the prior preparatory research for the mortality trial, we're not certain whether further research on this question (which represents 34% of the budget) will meaningfully improve implementation. Our decision to include it is partly based on the explicit support of Evidence Action, the main chlorination implementer we currently fund at scale, who have signalled an interest in applying the findings to their own work.
• The findings may have limited generalizability. While the findings of this work will directly inform our work in the specific places where this research will take place (Kano, Cross River, and Odisha), we recognize that infrastructure and cultural factors vary substantially both within and between countries. Since our primary interest is in using this research to inform a trial and future grantmaking in these specific regions, we will interpret the results of this work for other contexts as hypotheses to be tested further, rather than universal lessons to inform all of our ILC grantmaking.
• There is a trade-off between mortality burden and feasibility in the selected sites. Our impression is that Kano has a higher under-five mortality burden than Cross River,30 but that Cross River likely has more ILC-compatible infrastructure,31 and may be easier for Self Help Africa to operate in, due to security concerns in Kano. This contrast should help to determine whether mortality rates or implementation factors are more constraining for a future trial.
• There are ethical considerations around withholding the intervention: The trial design raises ethical concerns by establishing control communities that won't receive chlorination for 12-18 months, particularly in Kano, which has a high mortality rate. While researchers will obtain local IRB approval and government support, withholding a potentially life-saving intervention remains ethically complex. We think that in this case, the concern is somewhat mitigated by the fact that opportunities to fund large-scale ILC implementation aren’t currently available in these regions, as far as we’re aware.
• There is potential for a conflict of interest: Amy Pickering's team developed the TuriTap liquid doser being tested, creating potential bias in the comparison with tablet dosers. While concerning, we believe this can be managed through objective measurement protocols for both technologies, with clear documentation of performance metrics.

Plans for follow up

We will track progress through regular milestone-based check-ins, including around the launch of field trials, the interim results in Kano and Cross River, and the development of the proposal for a mortality study. Towards the end of the grant period, we expect to make a decision on whether or not to fund the larger trial.

Internal forecasts

For this grant, we are recording the following forecasts:

Confidence	Prediction	By time	Resolution
60%	We will fund a mortality trial proposal resulting from this grant.	November 2027	-
75%	The ILC pilot will demonstrate at least a 50% reduction in E. coli contamination in treatment communities compared to control communities.	September 2026	-

Our process

• We initially funded the Development Innovation Lab (DIL) at the University of Chicago to begin work preparing a potential mortality trial of in-line chlorination.
• DIL indicated that a trial wasn't feasible in the near term, but recommended that we we work directly with Amy Pickering on ILC-specific research
• We received and reviewed Dr. Pickering's initial proposal focusing on one Nigerian state, but after consulting with external experts and with Evidence action, we requested an expanded proposal covering two states and additional taste research.
• We also had multiple calls with Dr. Pickering and Self Help Africa.

Sources

Document	Source
Development Innovation Lab at the University of Chicago, Vouchers for Safe Water Kenya - Phase 1 Report, June 2024	Unpublished
Email from Amy Pickering, May 8 2025	Unpublished
GiveWell Blog, “Research strategy: Water.”	Source
GiveWell Blog, GiveWell Is Looking to Fund Pilots of Water Chlorination Programs	Source
GiveWell, Chlorine Vouchers	Source
GiveWell, Development Innovation Lab at the University of Chicago — Bridge Grant for RCT of Water Quality Interventions (January 2024)	Source
GiveWell, Development Innovation Lab at the University of Chicago — RCT of Water Quality Interventions: Phase 1 (January 2023)	Source
GiveWell, Evidence Action — Scale-Up of In-Line Chlorination in India (September 2023)	Source
GiveWell, Evidence Action's In-Line Chlorination Program — General Support (July 2022)	Source
GiveWell, Water Quality Interventions	Source
IHME, 2021 Global Burden of Disease project	Source
Lindmark et al. 2022	Source (archive)
Mangrove Water, Our Mission	Source (archive)
Pickering et al. 2019	Source (archive)
Self Help Africa, Our Work in Nigeria	Source (archive)
UC Berkeley Civil & Environmental Engineering, "Amy Pickering"	Source (archive)
UC Berkeley, Optimizing the benefits of inline chlorination in Nigeria and India, 2025	Source
UC Berkeley, Running list of questions from GiveWell, 2025	Unpublished
UC Berkeley, Summary Budget: Optimizing the benefits of inline chlorination in Nigeria and India, 2025	Unpublished

• 1

  “Amy Pickering is the Blum Center Distinguished Chair in Global Poverty and Practice, jointly appointed in the Department of Civil and Environmental Engineering and the Blum Center for Developing Economies. Pickering’s research focuses on using tools from multiple disciplines (engineering, economics, microbiology, epidemiology) to identify low-cost and scalable interventions to interrupt disease transmission in low-income countries. She has 15 years of experience collaborating with partners in Benin, Kenya, Tanzania, Mali, Mexico, Sri Lanka, Bangladesh, and India on research to improve human health and well-being.” UC Berkeley Civil & Environmental Engineering, "Amy Pickering"

• 2

  Pickering et al. 2019

• 3

  For more on the TuriTap, see Mangrove Water, established by Amy Pickering, Megan Lindmark, and Jeremy Lowe. Note: GiveWell is separately funding Mangrove Water to provide technical assistance for several ILC pilots funded through our 2025 request for proposals. Our grant page for this work is upcoming.

• 4
  • The UC Berkeley team has conducted initial scoping work in Kano state. "For reference, in terms of ILC-compatible water points, in our exploratory work in Kano we selected communities with boreholes, and then sent our team out to confirm… As part of our pilot work, we were also able to get data on communities (total number) and list of communities boreholes from RUWASSA/LGA staff from 5 LGAs [local government areas]. UC Berkeley, Running list of questions from GiveWell (unpublished).
  • Additionally, the UC Berkeley team's positive experience working with local government stakeholders in Ogun State gives us confidence in their ability to manage these relationships in Kano and Cross River. "We have already established a working relationship with RUWASSA [Rural Water Supply and Sanitation Agency] in Ogun for the installation of three ILC devices." UC Berkeley, Running list of questions from GiveWell (unpublished).

• 5

  See Self Help Africa's Work in Nigeria for a list of programs.

• 6

  For more on GiveWell’s approach to chlorination to date, see our blog post: GiveWell, “Research strategy: Water.”

• 7

  See this section of a previous GiveWell grant page for an in-line chlorination program in Malawi, which compares ILC to Evidence Action’s “Dispensers for Safe Water”, for more details. Note that this description is of a tablet erosion chlorinator, the only type of in-line chlorinator GiveWell has funded at scale to date.

• 8

  For more, see Lindmark et al. 2022. Table 2, “Evaluations Conducted on Passive Chlorinators Identified in Peer-Reviewed Literature,” lists many different in-line (“passive”) chlorination devices.

• 9

  "The TuriTap is our innovative, open-source, liquid chlorine in-line chlorination device. It accurately and passively doses chlorine into drinking water by harnessing the Venturi effect—a principle of fluid dynamics that enables chlorine dosing without the need for electricity. This innovative design not only reduces operational costs but also extends the reach of safe water solutions to off-grid and resource-limited areas." Mangrove Water, Our Mission.

• 10

  “We plan to conduct scoping activities across 100 communities to identify 40 communities with ILC-compatible drinking water infrastructure to be enrolled in an RCT in each state. We consider infrastructure to be compatible with ILC if drinking water is supplied at least once daily or when needed, everyone in the community has access to the drinking water source(s), baseline turbidity levels are below 0.5 NTU, and no known priority chemical contamination exists.” UC Berkeley, Optimizing the benefits of inline chlorination in Nigeria and India, 2025, p. 7

  Note: In the original proposal, Dr. Pickering and her team planned on conducting a trial in Ogun State. During our investigation, GiveWell and the UC Berkeley team decided to conduct research in Cross River State instead due to the relatively low mortality burden in Ogun, Self-Help Africa's preference for Cross River due to their existing work and relationships there, and security concerns in other potential locations. Source: UC Berkeley, Running list of questions from GiveWell (unpublished).

• 11
  • "In each state, following enrollment of 40 communities, we will randomly select 20 communities to receive an ILC installation in compatible drinking water sources. We plan to install two types of ILC devices to evaluate their technical performance and cost: 10 communities will receive the AquaTabs tablet-based dosers and 10 communities will receive the TuriTap liquid chlorine dosers." UC Berkeley, Optimizing the benefits of inline chlorination in Nigeria and India, p. 8
  • "At this stage, it is hard to say whether the range of water infrastructures between communities allow us to randomize doser-type allocation." UC Berkeley, Running list of questions from GiveWell (unpublished).

• 12

  “We expect to install an average of 2 ILC devices per community, with each device serving ~200 households, for a total of 16,000 households across the 40 treatment communities (80,000 people assuming 5 per HH).” Email from Amy Pickering, May 8 2025 (unpublished).

• 13

  "We plan to conduct a baseline survey among 15 households per community using ILC compatible water points enrolled in the study to understand household drinking water usage and existing perceptions of chlorination…Following ILC implementation, we will conduct 3 additional follow-up survey rounds (n=15 households per community) at 4-month, 8-month, and 12-month timepoints post-installation (600 household surveys per time point)…During each household survey round, a subset of 5 households from each community will have their stored drinking water tested for residual chlorine concentrations and microbiological contamination by measuring concentrations of E.coli (n=200 samples per round). We will also conduct weekly chlorine concentration monitoring thereafter for the 12-month study period." UC Berkeley, Optimizing the benefits of inline chlorination in Nigeria and India, p. 8

• 14

  "The analogue meters will be used to monitor how much water has been treated with ILC devices (and consumed by communities) and will provide valuable data on uptime for the water points enrolled in the study." UC Berkeley, Running list of questions from GiveWell (unpublished).

• 15

  “Participants will be asked to taste a series of 3 fermented rice samples (~100 grams each) in which 2 samples will be prepared with non chlorinated drinking water from nearby villages and 1 sample will be prepared with chlorinated drinking with a chlorine concentration varying between 0.2-2.0 mg/L. Participants and surveyors will be blinded to which samples were prepared with chlorinated versus non chlorinated drinking water. Participants will then answer questions about the samples they prefer the most and if they can identify which sample has been prepared with chlorinated water. To characterize taste acceptability, participants will also be asked to rate the taste of the fermented rice on a 5-point scale.” UC Berkeley, Optimizing the benefits of inline chlorination in Nigeria and India, 2025, p. 8

• 16

  "We plan to conduct desk research on cultural practices or cuisine that could be affected by chlorine in India, in addition to the issue of consuming pakhala in Odisha. We plan to conduct a literature search of published and grey literature, as well as interview experts and those with knowledge of Indian cuisine and cooking practices." UC Berkeley, Optimizing the benefits of inline chlorination in Nigeria and India, 2025, p. 9

• 17

  "In each state, we plan to enroll 300 individuals across both treatment and control communities (600 in each country) and conduct chlorine taste detection threshold experiments…For each set, participants will be asked if they can identify which drinking water sample has been prepared with chlorine and if they find the water acceptable…we will be able to detect a minimum difference in the median taste acceptability threshold of 0.18 mg/L between treatment (chlorine exposed) and control groups." UC Berkeley, Optimizing the benefits of inline chlorination in Nigeria and India, 2025, p. 9

• 18

  UC Berkeley, Summary Budget: Optimizing the benefits of inline chlorination in Nigeria and India, 2025 (unpublished)

• 19
  • This amount includes approximately $150,000 that will be sub-awarded to the University of Michigan for one co-investigator's time. Source: UC Berkeley, Summary Budget: Optimizing the benefits of inline chlorination in Nigeria and India, 2025 (unpublished)
  • "Here is a rough approximation of the splits, please note there are some synergies with conducting the taste test experiment along with the Nigeria cRCT (e.g. field managers overseeing both activities), and doing the taste test experiment across countries.
    • Nigeria cRCT: 66%
    • Nigeria Taste Detection: 15%
    • India Taste Detection: 19%"

  Source: UC Berkeley, Running list of questions from GiveWell (unpublished).
  .66 x $1,798,032 = ~$1,200,000
  .15 x $1,798,032 = ~$270,000
  .19 x $1,798,032 = ~$340,000

• 20

  UC Berkeley, Running list of questions from GiveWell (unpublished).

• 21

  See GiveWell, Chlorine Vouchers.

• 22

  “In Kano, assuming a baseline prevalence of E. coli of 93.3% in drinking water samples (based on the original pilot)...Across the study sites, chlorination is not commonly used, and we expect the baseline proportion of samples positive for free chlorine to be low.” UC Berkeley, Running list of questions from GiveWell (unpublished).

• 23

  IHME's 2021 Global Burden of Disease project estimated an annual under-5 mortality rate of 3.2% in Kano.

• 24

  See Self Help Africa's work in Nigeria, including in Kano and Cross River, here.

• 25

  "[GiveWell]: Are there enough ILC-compatible water points in Kano to make a mortality trial feasible?...

  [UC Berkeley]: We believe we can collect the data needed to answer this question within the first 6 months of the grant period…we could seek access to RUWASSA [Rural Water Supply and Sanitation Agency]/LGA [Local government area] data, as we did for our pilot, and potentially triangulate that data with SHA [Self Help Africa] data (for the available LGAs)." UC Berkeley, Running list of questions from GiveWell (unpublished).

• 26

  See Development Innovation Lab at the University of Chicago — RCT of Water Quality Interventions: Phase 1 (January 2023) ($1.8 million) and Development Innovation Lab at the University of Chicago — Bridge Grant for RCT of Water Quality Interventions (January 2024) ($900,000). $1.8 million + $900,000 = $2.7 million.

• 27

  "This is Phase 1 of a two-stage grant. In fall 2023, we expect to decide whether to fund a Phase 2 grant for multi-site RCT or a scaled-up Kenya RCT, based on the findings from Phase 1." GiveWell, Development Innovation Lab at the University of Chicago — RCT of Water Quality Interventions: Phase 1 (January 2023)

• 28

  Development Innovation Lab at the University of Chicago, Vouchers for Safe Water Kenya - Phase 1 Report, June 2024 (unpublished).

• 29

  See for example Pickering et al. 2019, which states "Chlorination is widely used for drinking-water disinfection, but taste acceptability is often cited as an important barrier to adoption."

• 30

  IHME's 2021 Global Burden of Disease project estimated an annual under-5 mortality rate of 3.2% in Kano and 1.2% in Cross River.

• 31

  Our understanding of Cross River State's higher share of ILC-compatible infrastructure is based on our discussions with Self-Help Africa and because Kano is a relatively poorer state than Cross River, and we expect access to improved water infrastructure to be negatively correlated with indicators like poverty rates and mortality rates.