Editorial Perspectives: what is “safe” drinking water, anyway?

If you are like us, you are routinely asked variations on the question: “Is my drinking water safe to drink?” And unfortunately, these days it is getting harder to answer that question unequivocally.

Take the framework used by the United States to define “safe”. The Safe Drinking Water Act (SDWA), first enacted in 1974, established a water treatment-centric approach for delivering “safe” drinking water to the tap. SDWA established national primary drinking water regulations (NPDWRs), legally enforceable standards in the form of either maximum contaminant limits (MCLs) or required treatment techniques (TTs) for all public water systems. Today, NPDWRs exist for 89 chemical and microbial contaminants in drinking water.¹ Accordingly, if a water system meets these requirements when the water enters the distribution system, then drinking water in the United States is deemed “safe”.

For comparison, the World Health Organization (WHO) has adopted a definition for “safe drinking water” which refers more directly to a concept of “risk”, rather than solely or primarily to the achievement of particular water quality standards or “guideline values”. The WHO Guidelines for Drinking-water Quality² state that safe drinking water is that which “does not represent any significant risk to health over a lifetime of consumption, including different sensitivities that may occur between life stages”. The basic and essential requirements to ensure safety of drinking water are a “framework” for safe drinking water, comprising health-based targets established by a competent health authority, adequate and properly managed systems (adequate infrastructure, proper monitoring, and effective planning and management) and a system of independent surveillance. This framework entails systematic assessment of risks throughout a drinking water supply, from a catchment and its source water through to the consumer. This includes identification of the ways in which these risks can be managed, including methods to ensure control measures are working properly.

The risk management framework approach to providing safe drinking water has also been adopted in various national jurisdictions, including the Australian Drinking Water Guidelines³ and the Guidelines for Drinking-water Quality Management for New Zealand.⁴ These documents are framed around the philosophy that “safe drinking water” is not merely defined by water quality that is currently being achieved, but also by the management of vulnerabilities that may exist in systems being relied upon to collect, store, treat and distribute drinking water. These vulnerabilities, if not effectively managed, may result in failure to provide drinking water that is fully protective of public health at some time in the future.

Water quality guideline values and standards themselves are not solely based on public health considerations. In the United States, an MCL is supposed to be set as close as possible to a corresponding maximum contaminant level goal (MCLG), where risk assessment has established a level where there is no known or expected risk to health. For example, the MCLG for any suspected carcinogen is zero, acknowledging there is no other “zero-risk” exposure level. But MCLs must also weigh factors like economic and technological feasibility of a standard, particularly when it is implemented across all sizes (from very small to large) of public water systems. Accordingly, corresponding MCLs for carcinogens are not also equal to zero. Although this is a necessary compromise (an MCL of zero is simply not practically feasible), it has had the unintended consequence of interjecting politics into the US regulatory process through cost–benefit analysis and questions over what constitutes the “best available science” upon which regulations are based. This has slowed the pace of standard setting in the US to the point where SDWA is simply not nimble enough to respond to emerging public health crises (e.g., PFAS) or when new science calls for stricter standards (e.g., the growing evidence between nitrate exposure and cancer).⁵

 

So what constitutes “safe” when a drinking water standard of zero is not practically achievable?

Agencies around the globe including the WHO and the US EPA have effectively defined an acceptable or ‘tolerable level of risk’ for managing exposure to carcinogenic chemicals. This approach follows currently applied dose–response relationship modelling for exposure to carcinogens. Such dose–response relationships are conventionally modelled by downward linear extrapolations of experimentally-derived data, producing an implication that only zero-exposure can produce zero-risk. Limitations of these linear extrapolation models, for the representation of realistic dose–response relationships, have been increasingly recognized in recent years. However, their ongoing application requires that unless ‘zero-exposure’ can be achieved and verified, some other non-zero level of risk must be accepted. Thus ‘tolerable levels of risk’ have been set, such as an excess lifetime cancer risk of 1 in 1 million (10⁻⁶) or 1 in 100 000 (10⁻⁵). In this sense, the concept of ‘safe drinking water’ has been quantitatively defined.

Safe or tolerable levels of risk associated with exposure to pathogens have been more elusive to define. Conventionally, adequate disinfection to manage risks has been the basis of principal regulatory requirements, such as requirements to achieve minimum chlorine doses. While the terminology is not always applied, this is an adaptation of a Hazard Analysis and Critical Control Points (HACCP) risk management concept, widely applied in food and pharmaceuticals production. Pathogen disinfection requirements are increasingly related to the achievement of specified low levels of pathogen risk of infection or illness. In the US, these have been defined in terms of pathogen concentrations that relate to a specified risk of infection. The approach adopted by the WHO recognizes that illnesses caused from infection by one pathogen may be more or less serious than illnesses caused by another pathogen. Therefore an equivalent level of risk from such pathogens relates to different levels of infection.

The WHO methodology is based on the quantifiable unit of ‘Disability-Adjusted Life Years’ (DALYs), which incorporate both the number of life-years that may be lost from across a population due to exposure to pathogens, as well as the incapacitation that may be experienced with illnesses such as gastro-intestinal disease. Much like the approach taken for carcinogenic chemicals, the application of DALYs requires the establishment of a ‘tolerable level of risk’. For drinking water quality, these have most commonly been defined as pathogen risks that correspond to the loss of 10⁻⁶ DALYs per person per year in a given population. This can be thought of as one person losing one year of life (death occurring one year earlier than it otherwise would have), from a population of one million, per year. More realistically, it relates to a much larger number of people experiencing much less severe illnesses over that period of time. The application of DALYs has brought us much closer to a more universally quantitative definition of ‘safe drinking water’.

 

Going forward, one thing is certain: regulatory frameworks will continue to strain from the sheer number of chemical and microbial contaminants that can threaten modern drinking water systems.

This is seen in the US where high-profile drinking water pollutants including PFAS, microcystins, and perchlorate do not currently have federally enforceable NPDWRs. The development and inclusion of new guideline values in WHO, Australian and New Zealand drinking water guidelines are also slow and struggle to keep pace with ongoing developments in water quality science. To be effective, future developments will need to adopt procedures whereby chemical and microbial contaminants may be confidently managed without being exclusively based on the need to specifically identify and measure all contaminants of potential concern. Concepts such as source water protection, broad-spectrum multiple barrier treatment, and distribution system management will be central requirements. Much of the monitoring focus then will be on ensuring the performance of these system management procedures, rather than exclusively on water quality itself.

 

How can the research community help to better define and ensure delivery of safe drinking water? Here are opportunities where the audience of ESWRT can help lead the way:

• Safe drinking water requires reliable control and monitoring technologies. Although fundamental research is driven by discovery of new technologies and tools, we should not ignore the complementary goal of making existing or recently discovered approaches more widely available and accessible, particularly to smaller water systems that may have greater resource constraints. This requires efforts to validate technologies at appropriate scales and focus on lowering their cost of adoption. Such work is especially valuable from the US regulatory perspective, where the cost associated with adopting new technologies has long stymied the development of new standards. In optimizing treatment and monitoring technologies for cost-effectiveness, there remain opportunities to apply a more “systems-level” approach to water treatment, where net reductions in associated energy and chemical demand across the treatment plant may help to minimize capital and operating costs of implementing new technologies.

• A regulatory approach that defines acceptable levels for specific contaminants also requires improved understanding of what is present in source water and finished drinking water supplies. There is a long-standing role for the research community to develop analytical methods that identify new and emerging threats not yet subject to regulation. Modern chemical and bioanalytical tools can provide greater detail of the complex mixtures in source and finished water.

• Research should work to identify and better mitigate risks from vulnerabilities of existing regulatory approaches. This involves continued scrutiny on how aging infrastructure systems influence drinking water quality after the compliance point but before it reaches the consumer, and how such deterioration in quality can be avoided with alternative materials used in premise plumbing and distribution. As another example, there are opportunities for toxicologists to evaluate alternative paradigms for risk assessment that better address the cumulative effects of mixtures (for example, the framework put forth by Kortenkamp and Faust)⁶ and threats from so-called “known unknowns” (e.g., transformation products).

• We must continue to evaluate alternative approaches to define and regulate safe drinking water. Comprehensive risk management frameworks offer a means of protecting drinking water along all points of the supply chain, from catchment to consumer. But these frameworks also require further development, including improved understanding of critical control points within the supply chain and best management practices that can be implemented to maintain or improve quality at these control points.

• As another alternative, there may be merits toward shifting countries toward a “technology forcing” statute, akin to the Clean Water Act in the US. While there are clear infrastructure costs that need to be considered, would the safety of drinking water be improved if all community water systems were required to achieve a demonstrable level of safety based upon performance standards set by market-available technologies?

• Finally, we must not overlook those that rely on water supplies that fall outside any regulation or oversight. In both the US and Australia, for example, many people rely on sources of drinking water (e.g., private wells) that fall outside of the control of enforceable water quality regulations. For these consumers, the burden of ensuring a safe water supply falls squarely on the consumer. In tackling questions related to safe drinking water, there are opportunities for our community to use their collective expertise to help those consumers and communities for which little is known about their water quality and associated public health burden.

Notes and references

1. US EPA, NPDWR Alphabetical List EPA 816-F-09-004, Washington, D.C., 2009.
2. Guidelines for drinking-water quality: fourth edition incorporating the first addendum, World Health Organization, Geneva, 2017, Licence: CC BY-NC-SA 3.0 IGO.
3. NHMRC and NRMMC, Australian Drinking Water Guidelines Paper 6 National Water Quality Management Strategy, National Health and Medical Research Council, National Resource Management Ministerial Council, Commonwealth of Australia, Canberra, 2011.
4. Ministry of Health, Guidelines for Drinking-water Quality Management for New Zealand, Ministry of Health, Wellington, 3rd edn, 2017.
5. M. H. Ward, R. R. Jones, J. D. Brender, T. M. De Kok, P. J. Weyer, B. T. Nolan, C. M. Villanueva and S. G. Van Breda, Drinking Water Nitrate and Human Health: An Updated Review, Int. J. Environ. Res. Public Health, 2018, 15(7), 1557.
6. A. Kortenkamp and M. Faust, Regulate to reduce chemical mixture risk, Science, 2018, 361(6399), 224–226.

---