EPA's Assessments of Chemical Dangers -- Too Slow

A government investigation of the U.S. Environmental Protection Agency's (EPA) process for assessing dangerous chemicals concludes the agency is so slow and lacking in credibility that the system is in "serious risk of becoming obsolete."

The Government Accountability Office (GAO) completed a new extensive review of EPA's Integrated Risk Information System (IRIS), a publicly searchable database for studies and information on the human health effects of chemical substances. The GAO investigation concluded that recent EPA changes to the IRIS assessment process had made a bad situation worse.

This database is a significant tool to protect public health and the environment. Health risk assessments made using IRIS data directly influence the development of public health policies. The EPA's IRIS program is supposed to assess more than 540 chemicals now in the IRIS database, but from 2006 to 2007, it finalized evaluations of only four chemicals. At that rate, it will take almost three centuries to complete the assessments, assuming no new chemicals will require evaluation between now and then. The agency is also supposed to reevaluate old decisions to incorporate new scientific data.

EPA has a significant backlog of chemical assessments and a growing number of outdated assessments. The GAO reports that assessments of certain, especially dangerous chemicals, such as dioxin and trichloroethylene (TCE), the most frequently reported contaminant in groundwater, have been in progress for over 17 years and over 10 years, respectively. Unlike many other EPA programs that have statutory requirements, the IRIS program has no required deadlines.

In April, EPA released its new assessment process. The new process was not made available for public comment. This lack of transparency and public feedback occurred despite Office of Management and Budget (OMB) assurances that EPA would circulate a draft to the public before moving forward with the final process. Changes included one apparently demanded by OMB, which allows other agencies, including OMB, to comment on IRIS assessments. The comments from OMB and other federal agencies about the scientific assessments will not be made public nor be noted in any peer review process. Additionally, EPA changed the definition of the scientific assessment process to include policy considerations, where previously, science and policy were distinct.

GAO's Sept. 18 testimony before the House Committee on Energy and Commerce Subcommittee on Oversight and Investigations indicated that EPA's new procedures have failed to improve the program. Several factors contributed to the failure, including the fact that OMB and several other federal agencies now have an even larger role in evaluating the EPA's work, slowing down the process.

Under EPA's current IRIS procedures, OMB, as well as several other federal agencies, may intercede in the scientific assessment process multiple times. The comments and changes to the IRIS assessments made by these interceding agencies are not revealed to the public. In the GAO's analysis, this lack of transparency violates the principles of sound scientific analysis. Moreover, these federal agencies are often affected by IRIS assessments, which poses an apparent conflict of interest.

Another IRIS problem identified by GAO is that EPA management decisions to postpone completion of assessments to wait for more scientific analyses compound existing delays. Several of EPA's assessments are essentially stuck in a loop where the evaluation process stretches out over several years, during which time the scientific research used in the initial assessment becomes outdated, and the agency starts a new evaluation to incorporate more recent information. This cycle, combined with the new, longer evaluation process, is largely responsible for the significant delays in finalizing risk assessments, according to GAO. Without risk assessments, policymakers at the federal and state levels — and even in other countries — are not able to make informed risk management decisions.

For instance, in 2005 in the aftermath of Hurricane Katrina, the Federal Emergency Management Agency (FEMA) provided trailers to those without housing, which caused health problems because of high levels of formaldehyde. FEMA officials cited the lack of a standard for formaldehyde exposure in mobile homes as one of the problems that delayed action. Apparently, EPA had initiated an assessment of formaldehyde in 1997 to update the data in IRIS, but the process had not been completed by 2005 when FEMA took action. Instead, EPA chose to rely on an industry-funded assessment of formaldehyde, which projected the risk from the chemical to be 2,400 times lower than that determined in studies by the National Cancer Institute and the National Institute of Occupational Safety and Health, according to the GAO analysis.

At the hearing, EPA officials defended the program. Dr. George Gray, EPA's assistant administrator for research and development, cited increased staff levels and funding for the IRIS program. He also stated that there would be no way to conduct "scientific shenanigans" with the IRIS process because of existing independent peer reviews. Marcus Peacock, EPA's deputy administrator, also proclaimed that EPA "[does not] tolerate political interference with science."

back to Blog