RISK newsletter:
Risk Assessment As a Major Tool for EPA Policy Decisions
Source: The Society for Risk
Analysis' RISK newsletter,
Second Quarter 1995
(Lynn R. Goldman, M.D., Assistant Administrator for the
Office of Prevention, Pesticides and Toxic Substances at the U.S.
Environmental Protection Agency , was the speaker at the plenary
session of the Society for Risk Analysis' 1994 Annual Meeting in
Baltimore, Maryland. Given below is RISK newsletter's
summary of her address.)
Using "Risks Along the Regulatory Superhighway" as
the topic of her address at the 1994 SRA Annual Meeting, Dr. Lynn
R. Goldman, pediatrician, environmental epidemiologist, and,
since 1993, assistant EPA administrator, summarized how risk
assessment has evolved as a major tool in shaping policy
decisions at the EPA.
Dr. Goldman pointed out that risk has always had a role in
EPA's regulatory decision making, but initially no formal methods
existed for assessing risk. Instead, EPA's earliest set of
"cancer principles" was developed from the proceedings
of adjudicatory hearings on the carcinogenicity of the persistent
chlorinated pesticides, during which scientists from opposite
sides presented their findings.
The result, she said, was that "the cited findings
included that any chemical producing cancer in animals should be
considered a human carcinogen, and there is no method for
establishing a no-effect level for a carcinogen." With no
way to model or extrapolate low-level cancer risks for chemicals,
the EPA administrator at that time had only two options: ban the
chemical (protecting everyone) or retain it (possibly putting
some people at risk).
By the mid-1970s, however, the EPA had issued its interim
cancer risk assessment guidelines, becoming the first agency to
attempt to quantify cancer risks. "Precedents were set by
using consistent risk assessment practices, both in regards to
our handling of scientific uncertainties as well as in the courts
where the relative merits of our assessment practices were
debated and usually settled in favor of the agency."
But the process remained adversarial, Dr. Goldman said, and
the EPA sought advice from the National Research Council, which
responded in 1983 by publishing the well-known Red Book (the
report Risk Assessment in the Federal Government: Managing the
Process). It recommended keeping risk assessment and risk
management separate, to be linked together later by risk
characterizations as summations for policy makers.
In 1986, the EPA revised and added to its guidelines, making
them consistent with the NRC recommendations. Even so, concern
and criticism of EPA's practices persisted---and still do.
"Many in the regulated community view EPA's risk assessments
as too conservative," Dr. Goldman said.
In 1989, a report by the American Industrial Health Council
recommended an even-handed exposition of data and their
evaluations. The EPA subsequently developed guidance for the
conduct of risk characterization, which was later updated by the
agency's Science Policy Council. In addition, EPA joined in
sponsoring a study on risk characterization by the National
Research Council.
In 1990, the U.S. Congress asked the NRC to recommend how EPA
could best develop new risk assessments. The result was the 1993
report Science and Judgment in Risk Assessment. In addition,
Congress, through the Clean Air Act Amendments, required the U.S.
president to establish a Risk Assessment and Management
Commission, which is examining the implications of the NRC report
and will give recommendations on the assessment of residual risks
under the Clean Air Act.
Meanwhile, the agency is revising its 1986 cancer risk
assessment guidelines to reflect significant changes in the
science and understanding of carcinogenic processes. With the
explosion of information in basic biology and toxicology, major
emphases of the revised guidelines will be (1) the use of all
available cancer mode-of-action information and (2) modification
of hazard classification guidelines to give narrative
descriptions of hazards, together with three descriptors of
conclusions (known/likely, unlikely, and cannot say).
Dr. Goldman added that the agency will continue to fill in
the gaps with default assumptions, including the assumption that
a linear dose-response assessment is appropriate when the
mechanism is not known for certain.
At the same time, EPA is performing or sponsoring ongoing
projects to further improve the science base for decision making.
For example, in keeping with recommendations in the 1993 NRC
report on Pesticides in the Diets of Infants and Children, the
agency recently published in the Federal Register a proposal to
examine combined exposures from various routes (food, water,
etc.) and sources (initially the triazines, a major group of
herbicides that all produce mammary tumors in female rats).
In future research directions, the EPA will focus on ecology,
for which research, testing, and risk assessment methodology have
lagged behind. The agency also is planning to study non-cancer
effects, such as the effect of endocrine disruptors both in the
environment and on human health.
To streamline its operations, the agency's Office of Research
and Development is reorganizing along the lines of the NRC's
decision-making paradigm. It will include a National Health and
Environmental Effects Research Laboratory, a National Exposure
Research Laboratory, a National Risk Management Research
Laboratory, an Office of Research and Science Integration, a
National Center for Environmental Research and Quality Assurance,
a National Center for Environmental Assessment, and an Office of
Resources Management and Administration.
In addition, EPA Administrator Carol Browner has directed all
offices to develop standard operating plans to ensure appropriate
peer review for all agency science efforts.
In conclusion, Dr. Goldman said, "As we make
improvements in the scientific processes, it is important to keep
in mind that the legislative basis for environmental policy
decisions is multifaceted. Generally, risk is only one component
in decision making."
Also, she noted that risk assessment is used in many
different ways in the various agency programs to support
regulatory decisions. "Sometimes, a bounding estimate on
risk is used to set priorities for more in-depth review. In
others, it forms the basis for determining that more testing
information is needed on the chemical. Some assessment
information is used to classify chemicals and appropriately label
substances as to toxic or safety hazards. In the new industrial
chemicals program at EPA, some 2,000 chemical evaluations are
made per year based on minimal or no test data. Other assessments
are major, like those for the setting of air and water standards,
where there is a considerable body of data spanning an array of
health and ecological effects."
Dr. Goldman said the Clinton administration is committed to
the appropriate use of risk assessment and other analytic tools
in developing policies and regulations. And Congress is
developing several bills which prescribe certain types of risk
assessments as a part of regulatory decision making.
"Let us work together to create the best assessment
tools we can," Dr. Goldman admonished. "The challenges
that face us today demand no less of us than making government
work for the people of this country, and that entails using the
best science."
Editor's Note: Not included in the above summary are Dr.
Goldman's additional remarks on EPA's participation in efforts to
harmonize international risk assessment activities.
Go to: