Well Water Arsenic Mitigation in Bangladesh: Benefits Outweighing Risks of Sharing Private Environmental Data

Abstract

This paper argues that environmental data should be collected and shared more widely to benefit public health while maintaining privacy, using Bangladesh’s well water arsenic crisis as a case study. Arsenic of natural origin in well water causes about 5% of adult deaths in Bangladesh, but a recent study shows that mortality returns to background levels within a decade after exposure ends. This finding motivates comparing the Bangladesh government’s testing of millions of wells for arsenic in 2000–2005 and again in 2021–2023. Despite progress, the data reveal that 20 million Bangladeshis still consume unsafe water today, largely due to a lack of information and suboptimal interventions. We propose reviving arsenic mitigation through public sharing of anonymized test data via a web application combined with a solution-focused national media campaign to encourage targeted installation of safe wells. The Bangladesh arsenic crisis illustrates how a wider distribution of private but anonymized environmental data could serve public health across other domains by raising awareness and enabling more effective individual and governmental responses.

Introduction

Environmental toxicants are often heterogeneously distributed, making identification of hot spots crucial for reducing exposure. The proliferation of inexpensive environmental testing tools and smartphones to record and share data is transforming both risk assessment and mitigation. One key question is how large quantities of environmental data generated privately can be shared for public health benefit without compromising privacy. We address this question using Bangladesh’s recent testing of 6 million wells to present generalizable arguments and context-specific recommendations.

In Bangladesh, millions of wells have been installed privately because groundwater filtered by floodplain sediments contains fewer microbial pathogens than surface water. However, naturally occurring arsenic in the water pumped from many of these wells has been responsible for about 5% of adult deaths in Bangladesh for several decades, primarily from cardiovascular disease. A longitudinal study tracking 12 000 adults for 20 years recently showed that mortality in populations exposed to arsenic halved and returned to that of unexposed populations a decade after they stopped drinking water high in arsenic. The striking reversibility of such a major health impact, similar to the benefits of no longer smoking for a decade, underscores the importance of mitigation.

Arsenic Mitigation to Date

A comparison of national surveys of arsenic exposure conducted 2 decades apart highlights both notable progress and continuing challenges in arsenic mitigation in Bangladesh. The first survey, conducted in 2000–2005 under the Bangladesh Arsenic Mitigation and Water Supply Program (BAMWSP), tested 4.3 million wells using field kits (Figure ). The second survey, conducted in 2021–2023 under the Arsenic Risk Reduction Program (ARRP), tested 6.3 million wells with a different kit but also used smartphones to record results, GPS coordinates, and well depths in the field (Figure ). Both sets of kit readings require adjustments for consistency with laboratory analyses (Figures S1–S3 of the Supporting Information). Combined, the data provide the most comprehensive assessment to date of changes in groundwater arsenic exposure over time across Bangladesh (Tables S1–S4), even if some villages were skipped because of a cap placed on the number of wells tested per union under ARRP (Figures S4 and S5).

1.

Well water arsenic maps of Bangladesh based on two campaigns of field kit testing reveal consistent geographic patterns while showing an overall decrease in populations exposed to arsenic concentrations exceeding 50 μg/L from 40 to 26% (Supporting Information).

The most pronounced improvements occurred within a ∼75 km-wide southern belt of Bangladesh, where the number of unions (administrative units of which there are 4600 total) with >80% of wells exceeding 50 μg/L arsenic decreased by half, from 357 to 165 (Table ). Despite these gains, the overall proportion of drinking water wells containing >50 μg/L arsenic declined only modestly from 41 to 26% nationwide over the same time span (Supporting Information). This corresponds to a decline in the exposed population from approximately 30 million in 2000–2005 to 20 million in 2021–2023. While the World Health Organization guideline is 10 μg/L, we reference the national standard of 50 μg/L that is the current national policy target. Independent surveys conducted by the Bangladesh Bureau of Statistics and UNICEF reported a reduction in the exposed population to 20 million already in 2009, with evidently little further improvement since. Given that mortality declines dramatically after lowering arsenic exposure, the current stagnation underscores the need for a different approach.

1. Comparison of Two Well Surveys for Arsenic 2 Decades Apart Supervised by the Government’s Department of Public Health Engineering (DPHE).

	BAMWSP 2000–2005	ARRP 2021–2023
number of wells tested	4.7 million	6.3 million
number of unions >80% >50 μg/L	357	165
proportion >50 μg/L	41%	26%
population exposed >50 μg/L	30–33 million	22–23 million
number of DPHE deep (>500 ft) wells	73 025	>231 669

Barriers to Mitigation

Essential for reducing arsenic exposure is knowing the status of one’s own well. The problem is that wells last only about a decade on average before needing replacement (Figure S6), while the government’s two testing campaigns took place 2 decades apart. In combination with a doubling in the number of mostly private wells due to their popularity, this turnover explains why only 20% of wells tested in 2021–2023 could have been tested previously during the 2005 campaign. Without interim testing and no way to distinguish safe from unsafe water visually, millions have unknowingly been exposed to arsenic for years.

Another barrier is the poorly targeted provision of government wells >500 ft deep that, by and large, are low in arsenic (Figure S7). The number of deep wells installed by the government’s Department of Public Health Engineering (DPHE) increased from 73 025 in the early 2000s to more than 231 669 in 2021–2023. The ARRP data show, however, that 95 880 deep wells were installed in 32 836 villages with ≤40% high-arsenic wells (Table ). These villages received an average of 2.5 deep wells, only slightly below the 3.3 average in 2927 villages where >80% of wells are high in arsenic. In addition to such misallocation, government wells have often been allocated based on political connections, and many have been effectively privatized through inaccessible placement or by discouraging neighbor usage.

2. Criteria for Evaluating Types of Interventions Based on ARRP Data.

proportion of wells >50 μg/L	<40%	41–80%	>80%
number of villages	32 836	8524	2927
proportion of villages with >95% wells 150–300 ft <50 μg/L	56%	5%	3%
number of government deep (>500 ft) wells	95 880	53 073	9536
deep wells per village	2.9	6.2	3.3

^aAmong the subset of villages with at least 1 well in the 150–300 ft depth range.

A final reason for limited progress has been diminished media coverage and government communication about arsenic, reflected in declining search trends in both Bangla and English (Figure S8). The lack of awareness diminishes the chances that households take measures to reduce arsenic exposure but also means that drillers have had few incentives for targeting safe aquifers.

Spatial Variability: Challenge and Opportunity

Arsenic levels in groundwater vary greatly both laterally and with depth, which makes prediction challenging and implies that every well in an impacted area needs to be tested. However, a close up of a portion of the area where the exposure and health of 12 000 individuals was tracked for two decades illustrates how such spatial variability can also help reduce exposure (Figure ). An early survey of the study area showed that, although half the wells were unsafe, 90% of affected households lived within 100 m of a safe well. After 2 years, half of these households had reduced their exposure by switching their consumption to a neighbor’s low-arsenic well, despite potential social barriers and added inconvenience. Overall, the new ARRP data show that 32 836 of 44 287 tested villages contain no more than 40% contaminated wells (≥50 μg/L arsenic; Table ), underscoring the potential for well sharing without requiring any new well installations.

2.

Close up based on ARRP data of a representative 4 km² area in Araihazar Upazila (subdistrict) containing approximately 12 neighboring villages with contrasting arsenic distributions across (a) private shallow wells (<150 ft depth), (b) intermediate private wells (150–300 ft depth), and (c) primarily government-installed deep wells (>300 ft depth). The maps were generated with an interactive web application that provides local access to government test data, with privacy protection ensured by rearranging well locations into rectangular arrays centered on village coordinates while preserving the broader spatial patterns (tinyurl.com/Nolkup).

Depth-stratified maps of the ARRP test data for the same study area show, moreover, that several villages in the study region and likely others further east offer opportunities to install new low-arsenic wells within the 150–300 ft depth range (Figure a and b). Installing new wells to this depth range has reduced exposure for many in Bangladesh but could do so for many more if the necessary information was made available. , In fact, in more than half of the villages where up to 40% of wells are contaminated, 95% of wells in this depth range remain safe (Table ). In areas where neither well sharing nor private mid-depth wells are feasible, government resources should be focused on installing deeper wells.

Overcoming Barriers with Wider Data Access

The documented reversibility of increased mortality combined with new test data should motivate a reinvigorated public health campaign through traditional and social media, paired with information about safe aquifers available locally. The premise is that well water arsenic concentrations remain stable over time, which is the case with the notable exclusion of some transitional areas within 30 km of Dhaka. Other mitigation approaches, particularly household water treatment, have proven unsustainable in rural Bangladesh. , The key challenge, therefore, is effective communication of test data and convincing households and local governments of the value of this information when making decisions.

Since the 2000–2005 campaign, smartphone ownership has become widespread among rural Bangladeshi households. Beyond social media campaigns, a navigable web app displaying government test data could raise awareness of and highlight locally relevant mitigation options. The Bangladesh cabinet member overseeing DPHE has recently requested the rapid rollout of such an app, named Nolkup (tube well in Bangla; tinyurl.com/Nolkup). Crucially, individual well test displays in the app are anonymized while maintaining information granularity by presenting locations at the village level in geometric arrays that do not reflect actual locations (Figure a–c). Anonymization does mean households with an unsafe well need to be urged to ask their neighbors about sharing a safe well instead of relying on the displayed locations. Households could also decide based on the data that the chances of reaching a safe aquifer <300 ft deep warrant an attempt to install a new well at their own expense.

The Nolkup app also highlights villages where neither switching nor installing private wells up to 300 ft deep is an option (Figure b). This should motivate local residents to call on the government for the installation of a deeper, low-arsenic well. Using a simple algorithm, a different app based on the ARRP data could show local government officials how to optimize their annual allotment of deep wells for the greatest impact. Past experience has shown that representative local committees should be involved in well site selection to minimize elite capture.

A final recommendation relates to the ongoing installation of 1 million new wells annually in Bangladesh just to replace wells that fail. One way to ensure that the status of these wells is determined and shared is a network of certified, paid testers equipped with kits and smartphones. While this network does not need to be governmental, the government should subsidize the $1–2 per well testing cost, as field trials have already shown low household willingness to pay for testing.

Broader Implications

The case of well water arsenic in Bangladesh demonstrates how sharing anonymized private data about toxicants serves the public interest. Data access raises awareness, enables individual action where there is such an option, and compels the governmental response where it is not. Following the Fukushima nuclear accident, data deficiency led to government distrust and citizen-led development of inexpensive radiation monitoring. In contrast, New York City has offered free lead testing of tap water and posted anonymized results to contextualize concerns for years. Such examples show that the public rightfully expects the collection of and access to spatially resolved data concerning environmental health.

Biographies

Kazi Matin Ahmed is a professor of hydrogeology at the Department of Geology and also the dean of the Faculty of Earth and Environmental Sciences, University of Dhaka, with research interests on arsenic, salinity, and contaminations of groundwater, and has published over 300 journal papers and book chapters with more than 20 500 citations. His current works include making a master plan for restoration of aquifers of the Dhaka watershed, assessment of impacts of industrialization on Gazipur aquifers, arsenic hazard mitigation in different geological environments, studying the impacts of over abstractions of groundwater due to the influx of displaced Rohingya people in Cox’s Bazar and Bhasan Char areas, and water system management in the SW coastal region of Bangladesh.

Lex van Geen is a Lamont research professor at Columbia University and member of its Earth Institute faculty. His 25 years of field research on arsenic in Bangladesh has shown that contamination patterns are highly spatially heterogeneous. While this complicates prediction, mapping these patterns can reduce risk. For arsenic as well as other toxicants, such as lead in soil, he advocates widespread use of field kits by non-specialists to engage affected communities. He collaborates with public health scientists and economists to evaluate kit effectiveness through randomized controlled trials.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.5c06441.

• Calibration of field kits used for 2000–2005 Bangladesh Arsenic Mitigation and Water Supply Program (BAMWSP) and 2021–2023 Arsenic Risk Reduction Program (ARRP), potential impact of well mislabeling on calibrations, coverage of BAMWSP and ARRP surveys derived from Araihazar data, national arsenic exposure estimates from BAMWSP and ARRP, replacement rate of wells inferred from repeated surveys of Araihazar, distribution of deep well allocations, and search trends for arsenic in Bangladesh (PDF)

The manuscript was written based on contributions from all authors. All have given approval to the final version of the paper.

Core funding was provided for 2 decades by the U.S. National Institute for Environmental Health Sciences Superfund Research Program (Grant P42 ES10349), with additional support from numerous National Institutes of Health and National Science Foundation grants as well as our own institutions.

The authors declare no competing financial interest.