Proceedings of a Workshop—in Brief

Convened October 15, 2024

State of the Science and the Future of Cumulative Impact Assessment
Proceedings of a Workshop—in Brief

INTRODUCTION

Communities often face overlapping stressors like pollution, climate change, and social inequities that combine to create more significant health risks and environmental challenges than any single factor alone. These layered challenges disproportionately impact underserved populations, exacerbating health disparities and environmental injustices. Cumulative impact assessments can help scientists and communities understand the impacts of multiple environmental stressors by accounting for the totality of exposures and their cumulative effects over the life course, providing a scientific basis to help guide more equitable and effective decision-making to improve public health, well-being, and environmental resilience.

While the general impetus for cumulative impact assessment is well recognized, challenges and opportunities remain in applying this approach at the community, state, and national levels. To understand best practices and inform future activities, the National Academies of Sciences, Engineering, and Medicine (the National Academies) assembled an ad hoc committee to convene state of the science workshops and develop a consensus report to advise the U.S. Environmental Protection Agency (EPA) on further developing the scientific foundation underlying the practice of cumulative impact assessment. The committee held its first public workshop, State of the Science and the Future of Cumulative Impact Assessment: Workshop 1, on October 15, 2024, to gather information for developing their eventual consensus report. The workshop brought together participants from academic and private research organizations, non-governmental organizations, and government agencies in an online forum to discuss fundamental concepts and methods pertinent to cumulative impact assessment.

Committee Chair Weihsueh Chiu of Texas A&M University provided opening remarks for the workshop, which was designed to elicit broad perspectives in examining key concepts for understanding cumulative impacts; current knowledge on the combined impacts of the built, natural, and social environments on health outcomes; and approaches for collecting and integrating relevant data for assessing these impacts. Before the event, pre-recorded presentations were available online as background for the workshop discussions. The workshop began with a live keynote presentation, followed by three panel discussions addressing cross-cutting themes. Each panel was moderated by a member of the study committee and included invited presenters as well as a member of the community and tribal liaison group¹ that has been

__________________

assembled to provide the committee with community and tribal perspectives. Materials from the workshop, including the workshop agenda, recordings, and participant biographies are available online.² This Proceedings of a Workshop—in Brief was prepared by a rapporteur to provide a high-level summary of the content of the workshop based on recordings, slides, and transcripts. The views contained in this Proceedings of a Workshop—in Brief are those of individual workshop participants and do not necessarily represent the views of all workshop participants, the study committee, or the National Academies.

CONTEXT ON EMERGING OPPORTUNITIES IN CUMULATIVE IMPACTS

Janet Currie of Princeton University delivered a keynote presentation that explained how cumulative early life exposure to multiple pollution sources drives group-level disparities. She focused on emerging opportunities to use administrative records—data routinely collected by government agencies or other organizations, such as medical claims and education, vital, and juvenile justice records—to extract insights on the cumulative exposures communities face and the impact of policies addressing these exposures. She described how administrative records and artificial intelligence-integrated datasets, alongside natural experiments, can advance understanding and uncover relationships between exposures and outcomes.

Leveraging Administrative Data

Combining administrative datasets from various sources allows researchers to track the long-term impact of environmental exposures on health outcomes. For example, using National Vital Statistics System³ records, Currie’s research team analyzed proximity to Superfund locations, finding that Black women were more likely to live near such sites than white women. Using information from birth records, the research team showed that cleaning up Superfund sites was associated with reductions in rates of low birth weight and premature birth.⁴ Currie also used a natural experiment to evaluate the health impacts of changes in traffic patterns related to the implementation of the E-ZPass electronic tolling system in Pennsylvania and New Jersey between 1997 and 2001.⁵ Her team found that living closer to toll plazas was associated with higher rates of premature birth and low birth weight, attributed to exposure to pollution from idling vehicles. After the E-ZPass system reduced idling at tolls, these rates dropped by about 10 percent.

Administrative data sets can also be combined to study the later impacts of early-life exposures and shed light on the potentially lifelong impacts of policy decisions that affect those early-life exposures. One research group demonstrated this by using the Longitudinal Employer-Household Dynamics file from U.S. Census data, information from unemployment insurance records, and counties of birth to study young adults whose early-life exposures to air pollution either were or were not affected by the passage of the Clean Air Act during early childhood or in utero.⁶ The study revealed that people who were exposed to less air pollution in early life as a result of the Clean Air Act had a significantly better earnings distribution—with fewer individuals not working and more people obtaining average earnings by the time they were about 30 years old. In another study, Currie’s team used Rhode Island’s comprehensive blood lead and education records to examine outcomes after implementing a lead exposure reduction program, revealing significant improvements in health and academic achievement.⁷

The Role of Monitoring

Currie emphasized how monitoring pollutants is critical for enabling research insights. She noted that a lack of measurement impedes progress because problems are unlikely to be addressed if they are not captured and quantified. For example, she said that large areas of the United States lack EPA fine particle air pollution (PM2.5) monitors, including in those estimated to have high levels of PM2.5 pollution. She added that similar gaps in water quality monitoring have allowed communities to be

__________________

exposed to contaminants for decades in places like Flint, Michigan. Even now, she said, many chemicals are likely present in community water supplies but not routinely monitored. To inform impactful interventions, she said, it is critical to identify the right pollutants and then deploy monitoring systems to ensure adequate coverage everywhere. In addition to increasing the sophisticated networks of high-precision monitors that the EPA uses, she suggested that communities use inexpensive sensors and draw on their knowledge of local conditions to expand data collection.

Using machine learning models in combination with administrative data can also help address knowledge gaps and create an even more comprehensive understanding of exposures and policy impacts. For example, Currie and colleagues merged individual U.S. Census data with newly available estimates of national pollution exposures obtained by applying a machine learning model to all available pollution data.⁸ They found that from 2000 to 2015, Black individuals consistently faced higher PM2.5 exposure than white individuals, but the gap narrowed over time. The research underscores the role of policy in driving exposure trends, with a steep decline in exposures from 2005 to 2009—driven by enhanced Clean Air Act enforcement—explaining over 60% of the convergence in exposure levels between the racial groups.

Addressing Disparities

A common theme across the studies Currie highlighted and in the subsequent discussion is that environmental exposures disproportionately affect certain groups, exacerbating cumulative impacts. However, she also noted that policies that effectively target the exposures with the most significant health impacts will bring these communities the most significant health benefits. “Because people of color and poor people are more likely to be exposed to all of these things, any measure that cleans up these kinds of pollutants is likely to disproportionately benefit the people who are being most exposed,” said Currie. “Being more intentional about building things that are going to serve low-income people, and especially children, would go a long way to helping people to start life healthier.”

To further illuminate disparities and guide strategies to address them, Currie reiterated the value of drawing on administrative data and emerging opportunities to combine and analyze these data in new ways. “I think that creative use of administrative data offers a potential way to look at the experience of people who are currently not seen by our monitoring systems,” said Currie. “It also offers a potential way to follow people over time as they accumulate different exposures and to assess the effect of environmental policy on health and other outcomes.” To use administrative data responsibly, she stressed the importance of addressing privacy concerns and involving communities to ensure equitable participation and shared benefits.

When asked how cumulative impact assessments could contribute to anticipating the impacts of proposed policies, Currie suggested that the technology is ready, but the regulatory framework is not. Historically, she said, pollution standards have been set and then repeatedly lowered as scientific evidence reveals previously unrecognized health impacts at lower exposure levels. In light of the uncertainty involved in setting pollution standards, she suggested a more anticipatory approach as harms of not imposing or not complying with environmental regulations might be underestimated while the costs of imposing those regulations might be overestimated.

KEY CONCEPTS RELEVANT TO CUMULATIVE IMPACTS

The workshop’s first panel discussion focused on key concepts relevant to cumulative impacts. Speakers shared takeaways from their pre-recorded presentations and engaged in an open discussion about cross-cutting issues and themes.

Unique Facets of Cumulative Impact Assessment

Na’Taki Osborne Jelks of Spelman College began the session by providing context on what distinguishes cumulative impact assessment from other approaches used to understand exposures, impacts, and risks. She noted that the EPA defines cumulative impacts as a totality of exposures to combinations of chemical and non-chemical stressors and their effects on health, well-being, and quality of life outcomes. Cumulative impact assessments evaluate both quantitative and qualitative data to inform a decision. In this context, cumulative exposures

__________________

are examined over the life course at different ages and stages, reflecting a wide array of stressors.

Emphasizing that addressing hidden hazards—including non-chemical stressors—will require moving beyond standard publicly available data sources, Osborne Jelks underscored the value of tools like participatory science and community-generated data and the need to effectively integrate these approaches with traditional quantitative methods. She stressed that the unique value of cumulative impact assessments is their ability to embrace both quantitative and qualitative data. Further, she noted that integrating community knowledge and lived experiences into these assessments provides a more robust perspective and can help community-generated data to fill some of the gaps in publicly available data. To access and elevate community perspectives, she said, it is essential to identify community assets, authentically engage with them, honor their voice, and utilize multiple ways of knowing to get to the root of the problems communities face.

Osborne Jelks pointed to the EPA’s Environmental Justice Collaborative Problem Solving Model⁹ as a cooperative approach in which communities partner with municipalities or industry to examine a wide range of health impacts and collaboratively address those challenges. She highlighted an example in which EPA Region 4 in Atlanta, Georgia worked with the Proctor Creek Watershed, a local community inundated with environmental challenges and stressors. Through this community-centered collaborative process, the team identified numerous hazards to include in analyses of community health, investment priorities, and strategies to address cumulative exposures.

Shirlee Tan of the Seattle & King County Public Health Department, a member of the committee’s community and tribal liaison group, built on this discussion by emphasizing that cumulative impact assessments could be used to highlight and prioritize needs for communities most heavily affected by environmental exposures. “Because traditional processes have really been focused on risk and single exposures, it is really important for communities to be able to use cumulative impact assessment. [This will help them] understand the burdens on their communities, be able to start reflecting those burdens in decision-making processes, and be part of the discussion,” said Tan. “It allows them to ask and demand for things that will reduce the impacts they are seeing.”

Tan said there are many examples of state or local governments trying to incorporate cumulative impact assessments, but they tend to hit barriers before reaching full implementation. She said that ensuring these communities can bring their lived experience and information that would not usually be collected or considered to these assessments can help inform research and policy at the local, state, and national levels.

Regulatory Policy Context

William Boyd of the University of California, Los Angeles highlighted lessons learned through the regulatory history of risk assessment. He said that risk assessment, as currently practiced, is not working and has been unable to deliver the information necessary for effective regulation in a timely manner. To ensure that the burden of regulatory delay does not fall on the public, specific individuals, or communities, he identified potential approaches like generic methods to assess entire chemical classes and harm types or applying conservative safety and adjustment factors to account for uncertainties, recognizing the potential for compounding harms.

Boyd pointed to per- and polyfluoroalkyl substances (PFAS) pollution as an example of the failure of the Toxic Substances Control Act’s (TSCA) approach to pollution control. Boyd said that the approximately 65,000 chemicals already “in commerce” in 1976, when TSCA was enacted, were not assessed and presumed to be safe with no requirement for testing. Even in the few cases where the EPA has tried to regulate existing chemicals, it has taken decades to complete risk assessments for some of the most prominent chemicals, including trichloroethylene and formaldehyde. Under the current system, discussions or decisions about regulation can only happen once these risk assessments are complete.

Boyd posited that cumulative impact assessment offers a comprehensive approach that could help to overcome some of the shortcomings of traditional risk assessment. Using PFAS as an example, he described how a cumulative impacts approach would recognize PFAS as a broad,

__________________

highly persistent class of chemicals—commonly called “forever chemicals”—that accumulate in human bodies and pose long-term health risks. Cumulative impact assessments also provide frameworks to examine populations with high PFAS exposure across the entire life cycle of these chemicals, including workers in facilities where they are produced or used, communities near production sites, and those exposed to PFAS-contaminated water or PFAS-laden products. Finally, a cumulative impacts approach can account for non-chemical stressors affecting vulnerable groups, integrating community voices into the assessment process to address compounding impacts.

Looking ahead, Boyd cautioned that viewing cumulative impact assessment as additive or complementary—or done in tandem with the detailed quantitative methods required for traditional risk assessment—could cause further delays to an already lengthy regulatory process. Instead, he suggested that “when we think about the conceptual architecture of cumulative impact assessment, we should be thinking about alternatives to the basic risk assessment framework that have, as the ORD [EPA Office of Research and Development] report¹⁰ puts it, a bias for action.” For example, Boyd suggested screening across multiple potential types of stressors and establishing simple, hazard-based triggers for initial regulatory action.

Science and Policy to Address Inequities

Panelists then discussed developments at the intersection of science and policy that have helped to elucidate key concepts in cumulative impact assessment and advance strategies to address disparities. Tracey Woodruff of the University of California, San Francisco summarized findings from a consensus publication she organized that recommends approaches to address structural barriers and improve how science is used to reduce health-harming chemical exposures.¹¹ In the publication, Woodruff and colleagues emphasize incentivizing data generation using validated methods to support decision making and ensure accuracy. The publication calls for establishing provisional toxicity values, more precise definitions of susceptible subpopulations, expanded TSCA criteria, and incorporating community knowledge into hazard and risk assessments to address aggregate exposures and cumulative impacts.

Woodruff also drew attention to potential sources of bias in generating scientific evidence and other processes involved in traditional risk assessment, including as addressed in guidance from the National Academies¹². “It needs to be recognized that there are financial conflicts of interest in the assessment process,” she said. “The [National Academies] has made recommendations on how to remove some of the bias due to financial conflicts of interest in generating science.”

While cumulative impact assessments are not currently required to support policy making, Woodruff suggested that they could be a natural complement to tools like the Justice40 Initiative¹³ in helping to provide resources and protections to communities with high exposure burdens. She also suggested that cumulative impact assessments could be integrated into existing regulatory mechanisms that include specific rules around addressing susceptible populations like the Clean Air Act and TSCA. However, like Boyd, she noted that TSCA currently falls short of its mandate and fails to adequately define and address the experiences of vulnerable subpopulations. Recognizing that slow regulatory processes and limited resources can constrain progress, she added that using adjustment factors in risk assessments—as Boyd also suggested—could

__________________

be an achievable way to integrate cumulative impacts into existing regulatory frameworks.

Darryl Hood of Ohio State University posited that successfully deploying cumulative impact assessment will require new approaches to better understand the links between environments and health disparities in vulnerable populations. He introduced the concept of the public health exposome as a social-ecological framework to study the cumulative impacts of environmental, social, and behavioral exposures on health. The public health exposome framework is also used to study chemical and non-chemical exposures, supported by extensive databases, for querying impacts. However, Hood noted that essential limitations remain, such as restricted authority at the state level, which creates policy implementation challenges. He described how his research group used large data sets to explore the relationship between stressors and health disparities in Columbus, Ohio, and are now scaling their findings to examine policy and track risk trajectories to chronic disease at the national level. He highlighted how a climate pollution reduction program in Columbus fostered innovation—particularly in hazard identification and community engagement—and noted that engaging community members in meetings has improved responses and aided in navigating complex policy barriers.

CURRENT UNDERSTANDING OF THE COMBINED IMPACTS OF BUILT, NATURAL, AND SOCIAL ENVIRONMENTS ON HEALTH AND COMMUNITY WELL-BEING

The workshop’s second panel discussion focused on what available research has revealed about how different facets of the environment combine to influence the health and well-being of individuals and communities.

Understanding Combined Impacts on Community Health

Joan Casey of the University of Washington described how social exposures, environmental exposures, and climate change can impact community health. While some communities may have baseline resilience to various exposures, she noted that climate change can introduce new stressors that are difficult to study and critical to consider when addressing health disparities. For example, she noted that some communities suffer disproportionate exposures to heat waves, wildfire smoke, and high temperatures—and are also more susceptible to these impacts because of factors like poor housing quality, less access to adequate health care, or difficulty engaging with the legislative process. “For these reasons [it is] really critical to think about both health and health equity when we are trying to study the impact of climate change and the way it interweaves with social environmental exposures,” Casey said.

Recent years have seen growing attention to the notion that environmental factors cannot be fully understood without considering social factors and climate change, Casey said. Even as researchers seek to understand how these factors interact, she said, it is crucial to focus on finding the fastest and most effective ways to improve health and health equity. She said it is important to attend to both upstream and downstream factors that influence health to inform action. Upstream factors include external drivers like national housing policies, the built environment, and the communities where people live, while downstream factors focus on individual experiences, such as stress.

Casey also emphasized the importance of carefully allocating resources and prioritizing the construction of new infrastructure to avoid exacerbating environmental disparities, which can lead to health inequities. She pointed to California as an example, where high-income communities have by far the greatest uptake of solar energy. To mitigate the risk of inadvertently widening equity gaps, she suggested incorporating and tracking equity in every new policy as a proactive measure.

Casey pointed out that emerging tools and data sources can advance understanding of the exposures people experience from birth through death and how they interact with one another. The impacts of climate change also intersect with a plethora of intergenerational exposures and justice issues, and, importantly, tools for understanding cumulative impacts are increasingly incorporating these factors. However, because exposures can be highly correlated, it can be challenging to identify true drivers or whether some are acting together to cause additive and amplifying effects. Considering these challenges, Casey suggested that statistical analysis methodologies are needed to help parse causes and effects and inform action-oriented approaches.

Social Vulnerability, Resilience, and Capacity

Christopher Emrich of the University of Central Florida discussed the connections between cumulative impacts and social vulnerability, resilience, and capacity. While vulnerability and resilience are often considered opposites, he countered by stating that “social vulnerability is not the opposite side of the coin as resilience. You can be both very socially vulnerable and very resilient to disasters.” He pointed out that both social vulnerability and resilience can either exacerbate or attenuate the ability to respond to environmental shocks and stresses, which are important to consider when thinking about cumulative exposures. Both social vulnerability and resilience can also be mapped and analyzed across geographic areas at fine scales of resolution.

Emrich cautioned that it is essential to understand the nuances of various social vulnerability models and data sets before they are used in risk analysis or decision making. When thinking about the factors that make a community or individual vulnerable or resilient, he posited, it is helpful to separate ascribed vulnerabilities and resiliencies—factors that cannot be controlled—and achieved vulnerabilities and resiliencies—factors that can be controlled. “I think if we break vulnerabilities and resiliencies down into these two simple categories—ascribed versus achieved—we can really start to think about how can we influence these things in meaningful ways to help these individuals and communities to either become more resilient or to decrease their vulnerability,” he said.

Achieved vulnerabilities and resiliencies, which are influenced by factors like education or housing, can be potent targets for interventions to decrease vulnerability or increase resilience. Although ascribed vulnerabilities cannot themselves be changed, Emrich noted, it is possible to design programs and policies to help those people. A policy intervention aimed at helping older people prepare for disasters would be one example.

Elizabeth Vasquez of the State University of New York, a member of the committee’s community and tribal liaison group, highlighted how the social environment is essential for promoting community health and well-being, further noting that these elements all fall under the umbrella of a life course approach. Focusing on later life stages, she underscored the opportunities and challenges of understanding disparities and outcomes among older people using cumulative impact approaches. She pointed out that older adults in communities affected by redlining face compounded challenges as their cumulative impacts are intensified by this population’s limited capacity to relocate. Older adults can also face more significant resource limitations than younger generations, making it harder for them to engage in decision making processes and support issues affecting their communities. As a result of these factors, Vasquez said, older adults in heavily impacted communities are not always as visible as other groups and their needs are not always heard—therefore, more research and policy focus may center on younger and middle-aged adults.

Such gaps underscore the importance of proactively engaging with communities to deeply understand their needs and priorities. While there has been a significant refocus from doing research on communities to doing research for communities, Emrich stressed the continuing need to prioritize individual lived experiences and embrace the idea that research and interventions should be undertaken with communities in a user-centered, use-inspired process. “No longer should it be acceptable for someone up in an ivory tower or anywhere else to identify a problem and a set of solutions,” Emrich said. “Those solutions are only going to be acceptable, accepted, and implemented through deep integration with communities.”

Promoting Health and Well-Being

Several panelists focused on how a cumulative impact approach can advance positive outcomes for people and communities. Denise Dillard of Washington State University, who is Iñupiaq and originally from Alaska, spoke about opportunities for promoting health and community well-being. She echoed other presenters’ opinions, emphasizing that effective cumulative impact assessments need to include participatory methods. For Indigenous peoples, this can mean acknowledging the sociopolitical status, history, and culture of each tribe. She said that models should conceptualize well-being as a holistic interplay of physical, mental, social, and spiritual domains and recognize how the health of humans, animals, and the environment are interrelated.

To highlight how using cumulative impact assessments could help address issues at a community level, Dillard pointed to work in Alaska that is examining how connected Indigenous youth are to their culture and developing interventions to more deeply involve these youth, like opportunities to hunt for seals with an elder. For the communities involved in this program, seal hunting represents a way of transmitting knowledge, building connections with elders, and fostering pride in the knowledge carried across generations.

Dillard noted that a lack of running water in many Alaska Native villages has been recognized for decades. Yet, many villages still do not have access to running water and live in sub-standard housing despite known negative health impacts. Dillard said it is important to recognize issues but even more important to act. “We need to be able to first measure what these community-level factors are, but it is really important that we also address them,” said Dillard. “Just doing assessment after assessment after assessment is not really going to get us where we want to go.”

Dillard noted that recognizing historical factors that have affected specific communities can help researchers and policy makers more deeply understand their challenges. For example, historical events such as how the U.S. government used boarding schools to assimilate Indigenous children affected communities in different ways. While some individuals had positive experiences, she noted that others faced abuse, making it essential not to generalize. Dillard said acknowledging past injustices is critical for fostering community participation and effectively addressing complex issues. She added that there are significant opportunities to focus on community-level assessments but emphasized that robust measurement and a shift toward holistic, partnership-driven approaches are critical.

Sacoby Wilson of the University of Maryland discussed the concept of salutogenesis in the context of cumulative impacts. Aaron Antonovsky coined this term to capture ways to promote health, wellness, and well-being across many dimensions, including the built, natural, economic, political, and spiritual environments. Salutogenesis also lends itself to the concepts of resiliency and resilience. “When you think about salutogenesis, it is really about public health promotion and it is really about integrating, in a serious way, the social determinants of health into how we do our work around understanding cumulative impacts,” Wilson said, adding that this applies to “not just cumulative impacts from a negative perspective, but cumulative impacts from a positive perspective.”

Environmental justice science elucidates how disproportionate exposures, differential burdens, and cumulative impacts can lead to poor health outcomes and health inequities for historically underserved populations. In advancing this work, Wilson suggested that researchers could increase their impact by shifting from a reductionist approach to a positive, comprehensive understanding of health and resilience—recognizing that health is not just the absence of disease but also encompasses mental well-being, cultural connections, and quality of life. He also noted the importance of ensuring that communities drive the work when addressing cumulative impacts.

Wilson also suggested that resilience, often framed as the ability to “bounce back” from hazards, could be reimagined as “thriving resilience” to recognize and address inequities in adaptive capacity and resources. Some communities repeatedly bounce back simply because they have no other choice—poor housing quality or other systemic barriers have put them in a position where they will either keep facing the same challenges or be forced to leave their community. To build healthy, lasting communities that are resilient in a more positive sense, Wilson highlighted the need for systemic change to support thriving and well-being and address issues like climate gentrification, redlining, and disparities in access to healthy housing and nature. “Having a more positive frame moves us away from surviving to thriving,” he said.

METHODS FOR COLLECTING AND INTEGRATING QUANTITATIVE AND QUALITATIVE DATA ACROSS DIFFERENT DOMAINS, LEVELS, AND SCALES

In the third panel discussion, participants focused on methods and approaches to collect and integrate data in cumulative impact assessments.

Geospatial Tools

Marcos Luna of Salem State University focused on available geospatial tools to advance cumulative impact

assessment. He served on a National Academies ad hoc committee that authored a recent report on geospatial tools for environmental justice.¹⁴ In its planning work, the committee identified over three dozen geospatial tools used nationwide to address environmental equity or justice. To focus on cumulative impact assessment, the committee narrowed the list to tools that measured the magnitude, types, and interactions of multiple stressors, finding at least 17 tools that met this definition.

California’s CalEnviroScreen¹⁵ was the most followed model for cumulative impact measures, although Luna said that other states have developed similar models. CalEnviroScreen uses a risk model approach to calculate interactions between environmental and social stressors, which are then used to create a final assessment, including a score that can be used to allocate resources, prioritize communities for monitoring, and even in some cases, for enforcement. Luna said that the committee found that most tools are largely consistent with the current understanding of the science behind cumulative impact assessment but are typically driven by policy priorities.

Luna pointed to the Justice40 Initiative as an example of place-based priority investing. This federal program aimed to invest at least 40 percent of federal climate and environment resources in historically underserved communities and issued scorecards to track where funding went and whether their goals were met. As part of its cap-and-trade program, California has been allocating resources to communities using a similar approach, Luna noted.

While recognizing that well-constructed, validated tools are essential for understanding problems and justifying policies and investments, Luna said that effective community engagement is a hallmark of the most successful programs. “Affected communities have to be part of the decision making about what constitutes a problem and what constitutes a solution. At the end of the day, we are looking at problems that are very hard to pin down, and we do not even have a uniform definition for a lot of the things we are trying to measure,” Luna said. “When you are dealing with people, how you do things matters as much as what comes out of that process.”

Emerging Data Sources

Reed Walker of the University of California, Berkeley spoke about emerging data sources and analytical techniques for assessing environmental exposures and outcomes. “With the advent of new technologies like low-cost sensors or remote sensing data from satellite imagery, combined with new statistical tools like machine learning, we can observe, in some cases for the very first time, comprehensive measurements of environmental exposures with both high spatial and temporal resolution for the entire U.S.,” he said.

Walker highlighted how partnerships with federal agencies like the U.S. Census Bureau, Internal Revenue Service, and Centers for Medicare & Medicaid Services have enabled access to longitudinal data for the entire U.S. population. These data enable researchers to track individuals over decades, linking environmental exposures to long-term outcomes like life expectancy and mortality while supporting robust analyses and natural experiments to uncover causal relationships. While these capabilities hold enormous potential, he emphasized the need for responsible data stewardship.

Walker also highlighted challenges in conducting credible cumulative impact assessment research, including numerous confounding factors, especially when analyzing concentration-response relationships. He said it is essential to acknowledge these difficulties and prioritize methods to address confounding, unobservable variables, and omitted variable bias. He also expressed concern that these unresolved challenges persist even in top-tier studies used for regulatory impact analyses.

Discussing opportunities to further improve research designs and techniques, Walker noted that addressing confounding variables requires creativity and statistical expertise. While tools like machine learning and deep learning excel at prediction, he said they are inadequate for causal inference and counterfactual analysis, which require control groups or robust research designs. He warned against overreliance on such tools for causal questions, suggesting instead thoughtful methodologies to address fundamental issues in data analysis.

__________________

Combining Cumulative Impacts Across Domains

Panelists then discussed opportunities and challenges in combining data and analysis tools from different domains. Bill Rish of ToxStrategies discussed a project that evaluated tools, methods, and data for cumulative impact assessment. The project team found that no models thoroughly combine impacts across all domains, as cumulative impact assessments are multidisciplinary and rely on quantitative and qualitative approaches. He also noted that most models assess the relative potential for cumulative impact rather than quantifying cumulative impacts directly, often due to the high degree of scientific uncertainty. However, he said that effects-based strength of association approaches can help inform action in the near term while waiting for more scientific evidence and clarity. Rish also emphasized prioritizing local data collection and integrating insights from community experts.

Rish said that integrating cumulative impact assessments into regulatory decision making processes will allow for planning based on risk factors from all domains and reap improvements in community health. “I do not really feel like cumulative impact assessment’s primary role is to demonstrate that there are inequities—we know there are inequities,” he said. “It is to actually identify ways to improve community health from multiple perspectives.”

He noted that while chemicals have been historically regulated on an individual basis, it is now possible to move away from determining acceptable risk with a hazard index or excess cancer risk threshold. Cumulative impact assessments make it possible to look more broadly at how improvements to infrastructure and access to services interact with environmental exposures, as well as how they can be addressed—from multiple perspectives—via regulation to improve community health. To advance the use of cumulative impact assessments in practice, Rish emphasized eliminating siloed environmental regulations and coordinating efforts across agencies. While much of the evidence on non-chemical stressors comes from observational studies with limited causal insights, he suggested that interdisciplinary collaboration can help to spur progress, especially through collaborations involving community members who can help identify disproportionate impacts so that the most influential factors can be targeted for intervention.

Building on these points, Ben Trump of the U.S. Army Corps of Engineers discussed strategies to combine interdisciplinary data and use these data to evaluate risks across a variety of domains, from critical infrastructure to medicines and consumer products. “We have new and exciting tools like machine learning and deep learning to begin to read between the lines and provide more assessment and battle test our tools even further,” said Trump. “It expands the reach of risk and decision science to be able to ask much more thought-provoking questions on a much shorter timescale with a substantially reduced cost.”

When engaging in research, Trump emphasized the importance of understanding the limitations and uncertainties of data and tools, particularly for tasks like cumulative impact assessments or when applying decision tools like multi-criteria decision analysis. Although machine learning and other tools can enhance analysis by integrating quantitative and qualitative data, they are imperfect and can leave gaps. Trump stressed that it is crucial to identify these gaps, question assumptions, and approach findings with humility to avoid misrepresenting reality.

Trump underscored the need to characterize uncertainty in cumulative impact assessments, suggesting that dealing with uncertainty is often more critical than providing specific estimates, as uncertainty directly impacts the validity and effectiveness of translating findings to broader applications. Especially when combining qualitative and quantitative methods, he emphasized the need to address biases, measurement errors, and data validity. Highlighting the importance of scalable, multi-system approaches to assess resilience, he noted that effective evaluation can benefit from detailed environmental mapping, infrastructure analysis, and understanding system interdependencies.

Finding the Right Tool for the Job

Berneece Herbert of Jackson State University, a member of the committee’s community and tribal liaison group, highlighted practical considerations for using data and tools that she has encountered in her work to understand the impacts of urban flooding. As an urban planner, she said that it is key to understand ultimate goals. For example, is the end result a policy directive

or instrument meant to protect human life? Although Jackson, Mississippi, faces numerous challenges, Herbert highlighted the lack of data as one of the most significant barriers the community deals with. This challenge encompasses access to quality and longitudinal data, methods for obtaining and cleaning data, and ensuring its usability for replicating methodologies over time.

Herbert is interested in using screening tools that can highlight vulnerabilities and then be built into more detailed tools for impact assessment. She noted that most existing tools are conceptual and tend to be policy instruments. They also vary widely, with some using 17 indicators, some using 37 indicators, and others, like one from the Federal Emergency Management Agency, utilizing up to 87. In coastal Mississippi, the challenges of addressing coastal flooding are significantly different from those of urban flooding. Herbert said that determining the most suitable approach for a specific context is an ongoing challenge, underscoring the importance of carefully selecting tools and indicators to effectively measure the aspects of equity, vulnerability, and resilience that are most pertinent to a given situation and goal.

DISCLAIMER This Proceedings of a Workshop—in Brief was prepared by Anne Frances Johnson, Nancy Lamontagne, and Kathryn Z. Guyton as a factual summary of what occurred at the workshop. The statements made are those of the rapporteur(s) or individual workshop participants and do not necessarily represent the views of all workshop participants; the planning committee; or the National Academies of Sciences, Engineering, and Medicine.

COMMITTEE MEMBERS Weihsueh A. Chiu (Chair), Texas A&M University; Julia Green Brody, Silent Spring Institute and Brown University School of Public Health; Zhen Cong, Chapman University; Deborah Cory-Slechta, University of Rochester Medical School; Andrew L. Dannenberg, University of Washington; Mia Gallo, State University of New York’s University at Albany; Rima Habre, University of Southern California; Jerreed Dean Ivanich, University of Colorado and Johns Hopkins University; Jonathan Levy, Boston University School of Public Health; Cris B. Liban, Los Angeles County Metropolitan Transportation Authority; Kristen Malecki, University of Illinois Chicago; Rachel Morello-Frosch, University of California, Berkeley; David Slusky, University of Kansas; Yoshira “Yoshi” Ornelas Van Horne, Columbia University Mailman School of Public Health; Courtney Woods, University of North Carolina at Chapel Hill; and Lauren Zeise, retired, California Environmental Protection Agency’s Office of Environmental Health Hazard Assessment.

REVIEWERS To ensure that it meets institutional standards for quality and objectivity, this Proceedings of a Workshop—in Brief was reviewed by Darryl Hood, Ohio State University. We also thank staff member David A. Butler for reading and providing helpful comments on this manuscript.

SPONSORS This workshop was supported by the U.S. Environmental Protection Agency.

STAFF Kathryn Z. Guyton, Study Director, Board on Environmental Studies and Toxicology (BEST); Leslie Beauchamp, Senior Program Assistant, BEST; Elizabeth Boyle, Senior Program Officer, Health and Medicine Division; Anthony DePinto, Program Officer, BEST; and Austin, Scheetz, Program Officer, Division of Behavioral and Social Sciences and Education.

For more information about this workshop, visit https://www.nationalacademies.org/event/43693_10-2024_state-of-the-science-and-the-future-of-cumulative-impact-assessment-workshop-1

SUGGESTED CITATION National Academies of Sciences, Engineering, and Medicine. 2025. State of the Science and the Future of Cumulative Impact Assessment: Proceedings of a Workshop—in Brief. Washington, DC: The National Academies Press. https://doi.org/10.17226/29058.

Division on Earth and Life Studies Copyright 2025 by the National Academy of Sciences. All rights reserved.