Avoiding paralysis by analysis: EPA proposes a sensible approach to identifying chemicals of concern

Richard Denison, Ph.D., is a Senior Scientist.  Thanks to my colleagues Jennifer McPartland and Allison Tracy for their analysis of the EPA proposal discussed in this post.

Last week, the Environmental Protection Agency (EPA) held stakeholder meetings to get public input into the criteria it will use to identify additional chemicals of concern beyond the 11 chemicals or chemical classes it has already identified.  EPA used these meetings (as well as an online forum open until September 14) as an opportunity for the public to respond to a “discussion guide” it issued in August that sets forth draft criteria and identifies data sources it intends to use to look for chemicals that meet the criteria.

The day before the EPA meetings, the American Chemistry Council (ACC) issued its own “prioritization tool” which lays out its own criteria and ranking system for identifying chemicals of concern.  This post will make a few observations about EPA’s proposal.  My next post will provide a critique of ACC’s proposed tool.

EDF and the Safer Chemicals Healthy Families coalition strongly support EPA in this endeavor – both for what it is, and for what it is not.   

The word “action” was for many years virtually missing from the EPA chemicals program’s vocabulary.  (I guess you could say the program was kinda “missing in action.”)  Of the more than 60,000 chemicals on the market at the time TSCA was adopted in 1976, fewer than two percent have received any substantive, data-informed review.  So it is a welcome development that EPA is actually looking at chemicals in commerce – despite the lack of a mandate to do so under the Toxic Substances Control Act (TSCA) – and, for those posing concerns, initiating at least those limited actions allowed by TSCA.

For the chemicals of concern EPA identifies, it expects to develop “chemical action plans” similar to those it has developed for the first 11 noted above.  These plans identify “a range of actions … from voluntary phase-outs and alternatives assessments in cooperation with industry and other stakeholders, to the development of test rules to require the development of additional data under section 4 of TSCA, to controls or use restrictions under sections 5 or 6 of TSCA.”

What EPA’s proposal is not

That’s the purpose of the criteria EPA is now proposing to formalize.  Equally important is what EPA’s purpose is not.  As EPA states on its website:

“EPA’s goal is to identify priority chemicals for near-term evaluation, not to screen and prioritize the entire TSCA Inventory of approximately 84,000 chemicals.”  (emphasis added) 

EPA has been clear that the latter task – a comprehensive review and ranking of all chemicals in commerce – is beyond its current authority and resources, and that any such effort – to the extent it is desired – must await TSCA reform.

Clarity as to the more limited purpose of EPA’s current initiative is important to note for two reasons.  First, it means that EPA is not claiming that chemicals it identifies as priorities are necessarily those that have somehow been shown to pose the greatest risk in comparison to all other chemicals.  Rather, they are chemicals for which there is sufficient evidence or reason for concern that they warrant further scrutiny.  And, of course, in order to actually regulate the production or use of such chemicals, EPA would have to meet the very high burdensimposed on it under TSCA. 

Second, some in industry have been arguing that EPA cannot even name a chemical of concern unless it first shows it is at the top of the list, identified through some kind of comprehensive ranking system that is applied to all chemicals in commerce.  That approach is indeed awfully close to what ACC has proposed as its “comprehensive” prioritization tool, about which I’ll have more to say in my next post. 

This impression is amplified by ACC’s invoking of the Canadian approach to prioritization to support its tool.  What ACC fails to mention is that Canada’s approach – which entailed a review of all 23,000 chemicals on Canada’s equivalent to the TSCA Inventory – was mandated by statutein amendments to the Canadian Environmental Protection Act (CEPA) adopted in 1999.   

Moreover, Canadian agencies were given seven years and a major infusion of new resources to complete just the first phase of its process.  With 84,000 chemicals on the TSCA Inventory … well, I’ll let you do the math to guesstimate how long and how many resources it would take for EPA to carry out the same approach.  Without the authority and the resources, well, that’s just a recipe for paralysis by analysis.

So, what is EPA’s proposal? 

While the EPA description of its proposal is merely a “discussion guide” and more detail will be needed to fully comprehend it, the agency proposes to utilize a basic set of criteria about which there is little controversy at least at the 30,000-foot level.  We are pleased to see an emphasis on chemicals that can adversely affect children’s health, on PBTs (persistent, bioaccumulative and toxic chemicals), and on chemicals detected in biomonitoring.

EPA also identifies a list of sources of information it would use initially to identify chemicals meeting each of the criteria, and then additional sources it would use to refine the list, conduct reviews of the selected chemicals and initiate risk assessment and risk management actions as warranted.  Again, the sources EPA identifies are pretty straightforward.   

The idea, according to EPA, is to generate and make public an additional list of chemicals of concern this fall, and then to proceed on to identify more chemicals over time using the same criteria.  Release of that list would provide all parties with an opportunity to provide more information to the agency.

What else is needed? 

We generally support EPA’s approach and believe it strikes the right balance between clarity and transparency and avoiding paralysis by analysis.  Nonetheless, we offer the following 10 additional suggestions for improvement:

1.      Cast a wide net in Step 1:  EPA need not and should not limit the sources it relies on in Step 1 to a small number, as it suggests it will do, especially if those sources are intended to identify the longer list from which a subset will be selected for further review and action.

  • Many of the sources EPA plans to use in Step 2 could identify chemicals that might otherwise be missed in Step 1.  For example, databases of chemical releases to or presence in air, water, fish, sediment, etc. should supplement the human biomonitoring data sources identified for use in Step 1.
  • While an exhaustive search of all possible data sources is not warranted especially to develop a “starter list,” EPA should be able to efficiently conduct searches of multiple data sources by relying on its own and others’ integrated databases and portals, such as:
    • ToxRefDB:  EPA’s own Toxicity Reference Database captures thousands of in vivo animal toxicity studies on hundreds of chemicals.
    • ExpoCastDB:  EPA’s Exposure Database includes studies where chemicals were measured in environmental and biological media, including air, house dust and food, and human biological fluids and tissues.
    • The OECD’s eChem Portal:  The Portal consolidates chemical data from many different international and national programs across OECD member countries.
    • European Union classification databases and lists, including ESIS (the European chemical Substances Information System), which includes lists of chemicals classified using criteria developed under the Globally Harmonized System (GHS) for Classification and Labeling as carcinogens, mutagens, reproductive toxicants, aquatic toxicants, PBTs, etc.

 2.      EPA should not preclude using published, peer-reviewed literature as a primary source of information to identify priority chemicals:  While reliability and data quality always need to be considered in any weight-of-the-evidence approach, there is no a priori reason to exclude such information, any more than to exclude industry-generated data that populate many of the databases just noted.

 3.      Add criteria for environmental hazard and exposure:  EPA’s proposed criteria are heavily weighted toward human health and need to be balanced by adding criteria that address both hazards to wildlife and ecosystems and environmental release and exposures.

 4.      Expand health effects of prenatal and postnatal concern for children’s health:  Reproductive and developmental toxicity are appropriate focuses, but need to be supplemented with an additional explicit focus on neurodevelopmental effects, a major concern for many chemicals for which early life exposure may occur.

5.      Expand the scope of exposure considerations for children:  While EPA articulates a broad concern for children’s health, its main focus on products intended for use by children is far too limited.  Children may be directly exposed to products used in the home whether or not they use them.  And exposures that occur in utero via transfer of chemicals from pregnant women or through breast feeding may be just as or more important than children’s product exposures.

6.      Consider a broader range of vulnerable subpopulations:  While a focus on children’s health is warranted, EPA also needs to consider chemical exposures of workers and other subpopulations (e.g., environmental justice communities) who may be more susceptible or disproportionately exposed relative to the general population.  For workers, this focus should extend beyond chemical or product manufacturing workplaces to include exposures to chemicals in industrial or commercial products or materials they use (e.g., building materials, automotive products) or manage after their use (e.g., product and material recycling, disposal).

7.      Consider aggregate exposure to chemicals:  In making prioritization decisions, EPA should factor in the range of sources and uses of a chemical that contribute to overall exposure, not just those uses that fall under its TSCA jurisdiction.  While legal issues would need to be addressed if and when EPA decided that regulatory action would be needed, it makes no sense for EPA to ignore at this stage uses or sources of a chemical that may contribute substantially to overall exposure.  Just as EPA’s proposed reliance on biomonitoring data represents a measure of exposure integrated across all sources, so too should its consideration of other exposure information sources.

8.      Don’t exclude chemicals with high hazard or high exposure for which data gaps leave uncertainty as to risk:  Where strong evidence of high hazard or pervasive or high exposures exists, EPA should be able to prioritize such a chemical.  This is critical if the limited data gaps are to be addressed – otherwise, EPA will simply continue to look at the same data-rich chemicals over and over again.  Chemicals with high hazard or high exposure for which there is concern about the other parameter need to be prioritized at a minimum for data development to determine the level of risk they pose.  (I’ll have more to say about ACC’s insistence that only chemicals with affirmative evidence of both high hazard and high exposure should be identified as priorities.)

9.      Go beyond the TSCA Inventory Update Reporting (IUR) data on use and exposure wherever possible:  As we’ve blogged about repeatedly in the past, and as EPA has forthrightly acknowledged in finalizing major enhancements to its chemical information reporting system, chemical use information available to EPA through the IUR  are woefully incomplete and limited.  To the extent possible, EPA should look for other sources of such information to identify priority chemicals, and certainly should not exclude high hazard chemicals on the basis of such information.  That was a common mistake EPA made in its earlier, ill-fated ChAMP initiative.

10.   Provide more clarity as to how EPA intends to proceed from Step 1 to Step 2:  This is an area where EPA discussion guide is particularly lacking in detail and needs to be clarified.

As noted, my next post will provide a critique of the ACC prioritization tool it released last week, so stay tuned!


This entry was posted in Health policy, Regulation and tagged , , , , , , , , , , , , , , , . Bookmark the permalink. Both comments and trackbacks are currently closed.