Former EPA Officials, Scientists Urge Overhaul Of Risk Assessment Methods

January 12, 2023

A group of more than 40 scientists from academia, citizen groups and state and federal agencies are floating a slate of proposals to overhaul both how EPA implements the reformed TSCA and its approach to chemical risks more broadly, saying the agency’s current methods are rife with “deficiencies” that understate the dangers of chemical exposure.

The coalition published its consensus statement and a series of papers detailing their policy recommendations on Jan. 12 in the journal Environmental Health, calling for changes to the Toxic Substances Control Act (TSCA) program, EPA’s methods of assessing risks to vulnerable communities, data-gathering mechanisms and a variety of scientific tools used for chemical assessment.

“We conclude that these changes are critical to correct existing deficiencies in exposure, hazard, and risk assessments that lead to insufficient information to identify risks, underestimates of risk, inequitable distributions of risk, and unconsented transfer of risks from manufacturers to the public,” the consensus statement says.

Among the co-authors are Linda Birnbaum, the retired director of the National Institute of Environmental Health Sciences; Lauren Zeise, head of California’s Office of Health Hazard Assessment; Jennifer Sass, a senior scientist with the Natural Resources Defense Council; Sheela Sathyanarayana, a pediatric medicine professor at the University of Washington’s medical school and a former chair of EPA’s children’s health advisory committee; Tracey Woodruff, director of the Program on Reproductive Health and the Environment (PRHE) at the University of California San Francisco and a former EPA scientist; Eve Gartner, an attorney with Earthjustice, and other academics, environmentalists, state agency scientists and EPA advisors.

The authors, with expertise across scientific disciplines ranging from toxicology and risk assessment to sociology and law, say they developed the recommendations following a “multi-year process involving meetings and workgroups.” That effort was organized by the Science Action Network for Health and the Environment, a project of PRHE.

The consensus statement outlines five main scientific principles, drawn from recommendations previously issued by the National Academy of Sciences (NAS), the International Agency for Research on Cancer and World Health Organization (WHO) committees and chemical policies in Europe and California. All are aimed at improving “how EPA conducts exposure, hazard and risk assessment and risk management analyses” -- key elements of TSCA chemical reviews.

Specifically, they say, agencies including EPA should put responsibility for gathering chemical safety data “on the chemical producers that benefit from their production and use;” affirm that a lack of such data -- a frequent hurdle for TSCA reviews -- “does not equate to lack of hazard, exposure, or risk;” better protect susceptible populations; stop assuming that non-carcinogenic chemicals will have some “‘no-risk’ level of chemical exposure in the diverse general population;” and ensure that chemical assessments “evaluate and account for financial conflicts of interest in the body of evidence.”

The authors also fault several more specific aspects of EPA’s chemical policies, including what they say is a deeply flawed response to language in the reformed TSCA requiring it to account for risks to potentially exposed or susceptible subpopulations (PESS) among the broader populace, such as children and workers.

“EPA has not applied a consistent or complete approach to identifying PESS thus far and does not currently incorporate factors or utilize methodologies to approximate the combined risk resulting from exposure to multiple chemicals (and non-chemical stressors such as food insecurity, pre-existing disease, poverty, or racism) into its risk evaluations,” the joint statement says.

TSCA Implementation

The authors further call for several targeted improvements to the TSCA program, urging EPA to use caution on adopting new approach methods (NAMs) of chemical testing as alternatives to traditional live-animal studies, and reworking its approach to addressing susceptible populations.

There, they urge EPA to revise the definition of PESS it has used in TSCA reviews of existing chemicals, with an example “explicitly identifying racism and other extrinsic factors, and that whole communities, rather than groups of individuals, can have increased exposure or susceptibility” -- drawing on language the Obama EPA proposed when it first drafted the a “framework” rule for the risk evaluations.

For chemical tests, they note that NAMs coupled with “key characteristics” of toxicology can help fill in data gaps on less-studied chemicals but warn EPA to keep in mind the limitations of the discipline.

“[I]t is critical to recognize the important limitations of in vitro, in silico, and in chemico toxicology data and that these data cannot currently be used to make definitive determinations to rule out chemical health hazards, and should only be used to identify potentially hazardous chemicals and then address those hazards under the appropriate statute,” the consensus statement says.

They also warn against “expedited timelines to replace mammalian tests with high-throughput assays” -- a likely reference to the Trump EPA’s now-abandoned goal to generally cease using animal tests by 2035. Biden administration officials dropped that policy in 2021. The authors say NAMs remain “limited in their ability to provide critical information,” particularly for PESS.

Instead of a broad changeover, the scientists urge EPA to “ensure that a hazard classification is not weakened based on speculative or limited data,” and outline areas where they say NAMs could be used “to provide ‘actionable evidence.’”

For instance, they say, the advanced tests could help with “facilitating dose-response assessments to support regulatory standards; investigating the impact of complex chemical mixtures; identifying susceptible populations and quantifying differences in risk; [and] investigating how non-chemical stressors interact with chemical exposures.”

And they urge EPA to address the widespread data gaps on chemicals and their risks -- a particularly acute problem for both new industrial chemicals and chemicals that entered the market before Congress first enacted TSCA in 1976, resulting in thousands of “existing chemicals,” many of which have databases too limited to support risk analyses.

“To correct this gap, EPA should require data sufficient to characterize health hazards, predict potential exposures, characterize cumulative risk, as well as chemical standards,” in part through its TSCA authority to require industry to report already-known data, and to conduct needed tests on chemicals under review.

The authors propose an approach where EPA would “develop a set of properties to assess completeness of the database -- such as physical and chemical properties, exposure parameters . . . and health endpoints,” allowing “EPA and other federal agencies [to] use this list to outline data gaps, identify those most critical to decision making, and request necessary data for each chemical currently in or proposed to enter the market, which would facilitate timely data generation consistent with other programs in Canada and Europe.”

They also suggest that EPA could “incentivize [companies’] data generation [by] . . . publishing provisional toxicity values for chemicals and applying multiple default adjustment factors as needed to account for lack of data.”

And they emphasize that when key data is not available, EPA and its peers must not assume that lack of information implies a lack of risk.

Risk Assessments

The papers also criticize the traditional approach EPA and other agencies have long used to conduct chemical risk assessments, with separate methods for assessing cancer and non-cancer dangers -- and a default assumption for non-cancer studies that any given chemical has a minimum safe level for exposure, with no harms expected below that threshold.

They argue that there are “fundamental scientific flaws” with the threshold-based model, which they say “often leads to public health policies that fail to sufficiently protect individuals, groups, or communities, especially PESS.” They note NAS findings that some environmental contaminants that can cause non-cancer effects, such as lead or mercury, appear to have no such threshold.

Further, “the diverse human population is comprised of individuals who have varying health susceptibilities from intrinsic . . . and extrinsic factors” as well as exposures to multiple chemicals through multiple pathways and sources, they note.

“The variability in characteristics affecting an individual’s response to chemical exposure results in a wide range of individual thresholds and an expectation of dose-response relationships in the population that extend to low, commonly experienced doses, with probability of risk at doses below the traditional” non-cancer risk estimates.

They urge EPA and other agencies to reverse that traditional assumption and instead assume “low levels of exposure are associated with some level of risk, unless there are sufficient data to indicate a threshold level below which there is no risk for the entire exposed population.”

Agencies should replace the traditional non-cancer risk calculations with “methods proposed by authoritative bodies such as the NAS and WHO to quantify risks of non-cancer effects and derive risk-specific doses,” they write. -- Maria Hegstad (

Not a subscriber? Sign up for 30 days free access to exclusive environmental policy reporting.