;)M RISK ASSESSMENT FRAMEWORKS

Presented at trie
2nd U.S. - Japan workshop on Risk Assessment/Risk Management

Octooer i?8, 19b7

Peter W. Preuss, Ph.D.

Director, Office of Health and Environmental Assessment

United States Environmental Protection Agency
Washington, B.C. 20460

The views expressed in this paper are those of the author
and are not necessarily the views or policies of the United
States Environmental Protection Agency.

-------
INTRODUCTION

Earlier in this worksnop I spoke about Me U.S. Environmental Protection
Agency's (EPA's) guidelines for risk assessment. In tnat talk, I presented
some concepts of risk, some definitions for risk assessment and risk management,
and some specific information about the guidelines themselves.

Now, I would like to turn to a discussion of the risk assessment process
itself. Here, I will present some information on the development of risk
assessment, on the history of its use in decision-making, on some impediments
to its use, and on the ways it can be used.

HISTORY OF RISK ASSESSMENT

Risk assessment is not an invention of the 1970s, but has been with us
since Adam assessed the risk and chose to bite into the apple in the Gardes, of
Eden. As the modern science of toxicology developed, scientists developed
techniques for setting safe levels of threshold toxicants. These were all
variations of a theme: finding the No-Observed-Adverse-Effect-Level (NOAEL),
and dividing that by a Safety Factor, Uncertainty Factor, Margin of Safety, or
Margin of Exposure which yields what has traditionally been called an Accept-
able Daily Intake or now, at the United States Environmental Protection Agency-
(EPA), a Reference Dose. This process is still widely used at the EPA for
establishing criteria for air pollutants, water pollutants, pesticide toler-
ances, and the like.

During th« past several decades, an emerging and very pressing concern
has been tim risk from cancer; the techniques developed for evaluating the
risks fro* cancer are thought of by most people when, today, they refer to
risk assessment.

1

-------
During tms time, :ie role of risk assessment evolved from an incidental
role in any one regulatory decision toward a principal part of the decision-
making process. It is important to understand tnat there were a number of
forces, many associated with the legal and governmental process, that acted
together (perhaps even synergistically) to accelerate the use of risk assess-
ment in the decision-making process. We would like to point out briefly four
of these forces: court decisions, congressional interest, regulatory agency
actions, and advances in the biological understanding of disease mechanisms.

The first of these is court decisions. For instance, courts in the United
States have suggested that quantitative risk assessment is required for an
appropriate regulatory decision to be made in decisions on benzene [US Supreme
Court, 1980] and urea-formaldehyde foam insulation [US Fifth Cir, 1983],

The United States Congress has also been encouraging quantitation and
the use of comparative risk assessment. Their interest led to the key study
on risk assessment in the federal government conducted by the National Academy
of Sciences [NAS, 1983], and interest by individual congressmen has continued.

The regulatory agencies have also attempted to define risk assessment
and describe how it is done. (See Preuss, et al, 1987, for details.) Two
reports deserve special mention. The President's Office of Science and
Technology Policy (OSTP) published a major report entitled Chemical Carcinogens:
A Review of the Science and its Associated Principles [OSTP, 1985], Perhaps
the most la^ortant policy document, which has and will continue to ensure a
focus on topic of risk assessment, 1s the National Acadenty of Sciences
(NAS) stirfjT tntUltd Risk Assessment in the Federal Government: Managing the
Process [NAS, 1983] which I discussed earlier in this conference, and will
also touch on in a moment.

2

-------
Finally, there have been both apparent and real advances in the art of
risk assessment and in basic biological understand^ng of disease mechanisms.
As risk assessment techniques developed, scientific analyses often were
not separated organizationally from the decision process. Generic procedures
and assumptions, though supportable scientifically, were often unwritten or
inconsistently applied. The apparent or real inconsistencies that developed
led to a public perception that a scientific analysis was often fitted to the
decision rather than the reverse. These problems led Congress to commission
the study by the National Academy of Sciences referred to above. That report
recommended that government agencies separate risk assessment from risk manage-
ment, both procedurally and organizationally. They also recommended that
agencies issue inference guidelines similar to those we discussed earlier.

Once risk assessment became a clearly articulated component of the decision
process, it was often possible to quantify risk, to evaluate the degree to
which risk could be reduced in various regulatory to evaluate the degree to
which risk could be reduced in various regulatory strategies, and to include
these risk reduction analyses into the decision-making process.

This process of ensuring consistency and technical quality in risk assess-
ments culminated at EPA with, the publication of the risk assessment guidelines.
As discussed previously, both the government-wide and EPA interest in this area
began about ten years ago. It intensified in 1983 when former Administrator
Ruckelshaus committed the Agency to following the NAS recommendations [US EPA,
1984a], «fflt to the development of the risk assessment guidelines, five of.
whlch werf*published in September 1986 [Summary . Preuss and Ehrllch, 1987;
detailed guidelines - US EPA, 1986a - e]. I discussed these guidelines earlier

3

-------
-n tnis >vorksnop. Pus developing scientific, congressional, administrative,
and scientific framework has seen accompanied by a parallel expansion of the
application and use of risk assessment in the regulatory decision process, for
example, vinyl chloride at EPA [US EPA, 1976]; benzene at the Consumer Product
Safety Commission [US CPSC, 1978; withdrawn, 1981], the Occupational Safety
and Health Administration [OSHA, 1987], and EPA [US EPA, 1984b, c]; and many
others.

IMPEDIMENTS TO USE OF RISK ASSESSMENT

That quick overview describes how risk assessment has developed as an
analytical tool and how it has begun to be used as a decision-making tool by
evaluating risk reduction; that is, in EPA's situation, the risks and the de-
gree to which those risks can be reduced by selection of appropriate regulatory
strategies. There are, however, several problems that have slowed its imple-
mentation. These include dealing with the inherent uncertainties in the risk
assessment process, statutory requirements which ignore or downplay the use of
risk assessment, and the lack of public understanding about—and confidence
in--the use of risk assessment information. Let us first talk about some of
the inherent uncertainties in the scientific process itself [Goldstein, 1984;
Preuss, 1987]. Perhaps the most-discussed group of uncertainties are those
having to do with the statistical quality of the dose-response assessment; that
is, extrapolating the experimental results to inferences about human risk.

Some of th«s« concerns have to do with uncertainty about the biological basis
for the iMthtnatlcal models. Within the field of epidemiology, there are'
problems tftth the statistical power of the studies, and proper allowance for
confounding factors, With animal studies, there are problems with choice of
the appropriate mathematical model for extrapolating from high doses to low
doses, problems with using the right scaling factor for extrapolating from

4

-------
ani:na 1 data to humans, and adjusting for the route of exposure if experimental
and environmental exposure are by different routes of entry into the body.

An emerging area of concern has been in the field of pharmacokinetics.

This rapidly growing area of research will, I hope, help us to reduce the
uncertainties in relating external exposure to absorbed dose, in properly
characterizing transport of the toxicant to the target organ, and in understand!ng
the role of chemical transformations before the toxicant's metabolites reach
the target organ.

Another area of uncertainty, in which EPA has been active for a number of
years, is exposure modeling. Here there are uncertainties that develop because
on-the-spot measurement data are often unavailable, because there is a need to
estimate dispersion of the pollutants in the environment from available data
through mathematical models which have their own uncertainties, and because of
chemical and biological transformations as various toxicants disperse through
the environment.

A fourth area of uncertainty is the lack of wel1-developed techniques for
assessing risks for other kinds of end points besides carcinogenicity, muta-
genicity, and--perhaps--developmental and reproductive toxicity. That is,
there are significant deficiencies in our traditional techniques of evaluating
threshold toxicants, since those assessment techniques are typically not spe-
cific for a particular target organ, such as the liver or kidney, or for a
particular typ« of effect, such as neuro-behavioral responses. Beyond that, even
when the techniques are available, appropriate toxlcologlcal Information may
not have bMft collected for specific chemicals of concern.

As EPA has become more and more concerned with hazardous wastes, we have
also become more Involved with risk assessment for mixtures [US EPA, 1984d, e;
Stara and Erdreich, 1985]. This presents several more uncertainties 1n addition

5

-------
to those discussed "n tne last few paragraphs. For example, we do not always
know what toxicological significance there is to tne physical form of a toxicant
in a mixture relative to the physical form in the experiment from which the
toxicity information was originally developed. There are uncertainties in
selecting the appropriate methods for adding effects of several different
toxicants in a mixture. There are uncertainties about the role of interactions
in the mixtures (synergisms and antagonisms). Finally, we know that there are
data gaps for known components of the mixtures, and there may be contributions
from unidentified components of the mixtures.

The second impediment to the use of risk assessment as a regulatory tool
occurs when Congress writes laws that ignore or downplay the use of risk assess-
ment. In some cases, statutes provide only for banning upon a particular
showing of adverse health effects. The classic case is the Oelaney clause for
food additives that are thought to be carcinogens. Another approach, directed
in some EPA statutes, is the technology-based standard, where EPA is required
to mandate the use of pollution control technology (commonly referred to as
Best Available Technology) independent of any direct consideration of risk
reduction or cost-benefit ratios. We will discuss the influence of such laws
later in this paper. (For details, see Preuss et al, 1987.)

The third impediment to which we referred is the lack of public under-
standing about, or confidence in, risk assessment Information. That Issue is
being discussed at this workshop by several people.

At EPA» we have been doing several things to try to resolve some of these
problems. First, we have been trying to reduce some of the uncertainties in the
scientific risk assessment process through targeted research [Preuss, 1987],

6

-------
Some examples include developing new techniques in dose-resporse modeling for
carcinogens and for developmental toxicity. We have also been interested in
advancing the science of pharmacokinetics, and recently held a workshop on the
use of pharmacokinetic models in risk assessment [Gillette and Jollow, 1987],
We recently convened two other workshops, on tumor promoters [US EPA, 1987]
and on the relationship of maternal and developmental toxicity [Kimmel et al,
1987]. We expect these workshops to be the first of several on controversial
risk assessment issues. Finally, we have been placing more emphasis on non-
cancer health effects, specifically looking at techniques for evaluating addi-
tional health effects beyond developmental and reproductive toxicity to iimnuno-
logical effects, sensitization, and neurotoxicity; techniques for quantifying
inhalation toxicity; and statistical techniques for assessing risks from these
threshold toxicants. We are also doing more research on complex mixtures and
total human exposure.

We have been attempting to develop ways to communicate both the concepts
of risk assessment and information about risk assessment to the public. We
have recently developed, and made available to the public, what we call the
Integrated Risk Information System or IRIS [Preuss and Ehrlich, 1987]. This is
a computer-based file of EPA risk assessment and risk management information
for chemical substances. It is designed especially for federal, state, and
local environmental health agencies as a source of the latest Information about
EPA health assessments and regulatory status for specific chemicals. It is
intended ffcp users without extensive training in toxicology, but with some
knowledgtQfC health sciences.

In the past several years, EPA has put greater emphasis on working direct-
ly with the public concerned about environmental hazards or potentially affect-
ed by proposed EPA regulations. For example, the Agency worked directly with

7

-------
people living in the Takoma, Washington, area who were faced with the pjssi-
oility of closure of a shelter because of EPA's arsenic regulation, to explain
the trade-offs between estimated health risk and economic impact, and to give
the community a voice in the ultimate decision [Ajax and Meyer, 1985]. We have
also been developing pilot activities in specific geographic areas that inte-
grate risk assessment, communication of these risk assessments in laymen's
terms, and discussions with affected government and private groups to address
environmental problems on a regional basis. Earlier in this conference
Dan Beardsley has talked about some of those activities. EPA works closely
with people living near hazardous waste sites and with their communities in
regulating and cleaning up the sites. EPA is also in the process of Implementing
the Community Emergency Preparedness and Right-to-Know Act (Title III of the
Superfund Amendments and Reauthorization Act of 1986, PL 99-499), whose basis
intent is to inform the public about chemical risks in their communities and
establish community planning for chemical emergencies. As a last example, EPA
is attempting to gain a better understanding of how the general public perceives
EPA's explanations of the degree and nature of risks it is regulating so that
EPA can improve its communications with general public. You have heard from
Vincent Covello about some of those activities.

HQM RISK ASSESSMENT IS USED

Risk assessment 1s used in three major ways - for priority setting, for
regulatory dtclslonHMklng and for evaluating the benefit side of a cost-benefit
analysis. The first use Is quite obvious. Many risk assessments lead to-
specific numbers - Reference Ooses, LDso's, carcinogen risk estimates, and
the like. It Is quite commonplace for managers to use lists of numbers like
these to decide which chemicals should be regulated first, or which problems

8

-------
should -eceive tne 1 aryest share of a. program budget. The very existence of
toe numbers encourage this use.

The use of risk assessment in regulatory decision-making ,is best illustrated
with examples from different kinds of regulations. We pointed out earlier that
health and safety regulations can be based on three kinds of theories, depending
on the specifics of each environmental law. The three theories are risk-based,
technology-based or some kind of balancing between risk and other factors like
costs or technological feasibility. Let us review regulatory development under
each concept.

One example of the risk based approach is the development of regulations
for hazardous air pollutants under Section 112 of the Clean Air Act. The first
step is actually a priority ranking exercise similar to those I just referred
to. In the air program, this step is referred to as a screening assessment.
We rank chemicals we are considering by a combination of numerical ratings -
volume of production (as an estimate for exposure), health effects indices, and
the like. From this screening assessment, if a chemical remains of concern, a
brief risk assessment, called a Tier I, is done consisting of a review of
acute toxicity, estimate of oncogenic potential, and estimate of sources and
magnitude of exposure. If this brief analysis continues to indicate a cause
for concern, both a comprehensive Health Assessment Document and detailed
exposure assessment are prepared. Part of this preparation is a peer-review
process consisting of Invited expert review, public comment on external review
drafts, evaluation of the documents by our Science Advisory Board, and revision
of the documents in light of those comments. Once these steps are completed,
the process shifts from an emphasis on risk assessment to an emphasis on
risk management. The a<~ office makes a preliminary decision to regulate and
publishes that intent to hst the chemical as a hazardous air pollutant. In

9

-------
the United States a regulatory decision like that is first proposed for public
comment, the comments considered, and the regulation, in this case the listing
decision, subsequently issued in final form. Once work on a listing decision
has been completed, a specific regulation deciding limits for that chemical is
first proposed for public comment and then issued after the public comments
are considered.

Examples of a technology-based approach can be found in the effluent
guidelines program under EPA's Clean Water Act (or Water Quality Act, as it is
now called). Here, risk assessment has been used prior to consideration of
effluent guidelines to establish independent water quality criteria, but those
criteria are used only as guidelines for control of the potential sources of
of pollution. Once a decision has been made to regulate a particular industry,
the focus of the regulations is controlling the amount of each pollutant that
can be emitted by mandating the use of the best technology available, taking
into account the feasibility and cost of the treatment and the established
water quality criteria for the receiving stream. The controls might be in
terms of best practicable control technology or best available technology (for
existing production facilities discharging wastes directly into receiving
waters), pretreatment standards for production facilities discharging into
municipal sewage systems, or new source performance standards for new production
facilities. The effluent guidelines are, 1n turn, enforced by a permit system
which Includes the specific technology, permitted emission levels, and required
monltortfif for compliance.

The Safe Drinking Water Act, 1n comparison, contains a two-tiered approach,
starting off with a risk-based approach and switching to a technology-based
approach. When a potential problem is identified, a risk assessment 1s done,
and - 1f regulation is considered necessary, a Maximum Contaminant Level Goal

10

-------
(MCLG) - strictly neal*:i oased - is established. For instance, the MCLG is
set, as a matter of policy, at a risk of zero for known or probable human
carcinogens, below the RfD for possible numan carcinogens, and.at the RfD for
other threshold pollutants. In concept, once the MCLG is established, the
Agency then shifts to a technology-based approach, selects the best available
technology (BAT), and sets Maximum Contaminant Levels (MCLs) based on that
BAT. In practice, under the revisions to the Safe Drinking Water Act, the
proposal and subsequent promulgation of the MCLG, and MCL takes place together.
Proposed MCLGs and MCLs for several chemicals should be published soon and
selects the best available technology that comes as close to the MCLG as possible.
For threshold toxicants, the BAT usually achieves protection well below the
MCLG. For carcinogens, since obtaining zero is not possible, the technology
is evaluated in terms of parameters like analytical detection limits, degree
of risk reduction, and residual risk.

EPA uses a risk-balancing approach for regulation of pesticides under the
Federal Insecticide, Fungicide and Rodenticide Act (FIFRA). One can see why a
risk-balancing approach is legislated for pesticides, since it is clear that
these are toxic materials, and one must carefully balance the benefits to
society from their use against the risks to society from their use. One must
also carefully control how they are used and to what extent they will be permitted
FIFRA may also be a good example to discuss at this conference because it is
one of the few EPA examples in which government approval 1n advance is required
before the chemical can be used. We understand this 1s similar to regulatory
approaches used in many other countries of the world, Including Japan.

In the case of a pesticide, once the Agency has completed the risk
assessment, i.e. both determined the degree and nature of the hazard and
determined whether the pesticide is released into the environment, 1t then

11

-------
balances that information against the need for the pesticide and a consideration
of appropriate restrictions on its use. The restrictions can include:
° labelling instructions

0 restriction of use to licensed application who have the required skills

and safety equipment
0 restriction of approval to specific physical forms (e.g. solids,

aerosols, aqueous liquids, organic liquids)

0 denial of approval for its use

We have used examples from EPA, out other Agencies approach their regulations
in similar ways, again, depending on whether their legislation requires a risk-
based approach, technology-based approach, or balancing approach. Thus, for
instance, the Food and Drug Administration uses a risk-based approach for
direct food additives, but balances risks and benefits for unintended additives
or for drugs; and the Occupational Safety and Health Administration balances
risks and costs in setting its work-place standards.

The third use of risk assessment is for estimating the benefits of a
particular regulation in order to conduct a cost-benefit analysis, also frequently
called an economic impact analysis or a regulatory impact analysis. Once a
risk assessment is completed, we can estimate the degree to which the risk is
reduced, and can then estimate the value of the reduction in injury costs,
increased longevity, reduced environmental damage, and the like. (Of course,
once econoMtc factors have been added, the risk assessment has been transformed
into, or iMfeomt part of the risk management decision.) These kinds of analyses
are now rMptfred of Federal agencies, and they are routinely performed whether
the authorizing statute provides for such assessments or not.

12

-------
Tht FUTURE OF KISK ASSESSMENT

In discussing where -ve go from here, I want to point out again that there
is no general agreement or consensus regarding the use of risk assessment.

Some agencies strongly question whether risk assessments should be done at
all, or, if so, prefer that they be restricted only to qualitative analyses
of the risk; other agencies use risk assessment to a greater or lesser extent.
The extent to which risk assessment is used depends, in part, on how agencies
have reacted to the combination of two managerial issues: (1) whether the
risk assessment responsibility is centralized in headquarters or decentralized
to regions and specific sites, and (2) the extent to which risk assessment and
the evaluation of risk reduction are part of the agency's decision-making
process.

We see EPA at one end of the spectrum. EPA has been firmly committed to
the use of risk assessment, the evaluation of risk reduction, and the making
of careful risk management decisions, regardless of whether those decisions are
being made in headquarters or the field. There has been a major shift in EPA
toward local analysis of environmental problems and local control strategy
development at hazardous waste sites, municipal incinerators, and the like.
These local analyses include such factors as matching available technology to
existing facilities; adjusting exposure assessments to local variations in
geography, terrain, and population; and accounting for local variations in the
political, social, economic, and legal environment. Therefore, 1n addition to
upgrading and Improving risk assessment techniques themselves, as I have dis-
cussed throughout this presentation, we have also had to adjust our risk
assessment processes to ensure effective use of risk assessment 1n those
decentralized activities.

13

-------
Tnerefore, as we have discussed and will discuss over the next few days,
you may want to keep in mind that we expect to see more and better risk assess-
ments as an integral part of the regulatory process, greater decentralization
in the conduct of risk assessments, and a much greater expansion of knowledge
and expertise in the science, or perhaps more properly art, of risk assessment.

14

-------
REFERENCES

Ajax RL, Meyer J (1985): Policy consideration in the selection of national
emission standards for hazardous air pollutants for the Takoma Smelter.
Presented at the annual meeting of the Society for Risk Analysis,

October 8, 1985.

Gillette J, Jollow 0 (eds) (1987): "Proceedings of the Workshop on Pharmaco-
kinetics in Risk Assessment." Washington: National Academy Press, in
press.

Goldstein BO (1984): Strengthening the assessment of risk. EPA J 12(10):5-7.

ILSI Risk Science Institute (1987): "Review of Current Research Activities to
Improve Risk Assessment for Carcinogens." Draft report prepared by Envi-
ron Corporation for ILSI Risk Science Institute.

Kimmel GL, Kimmel CA, Francis E (eds) (1987): "Proceedings of the Consensus
Workshop on the Relationship of Maternal and Developmental Toxicity,
May 12-14, 1986." Teratogen Carcinogen Mutagen 7(3):201-338.

National Academy of Sciences, National Research Council (NAS) (1983): "Risk
Assessment in the Federal Government: Managing the Process." Washington:
National Academy Press.

Occupational Safety and Health Administration (OSHA) (1987): Occupational
exposure to benzene; final rule. Federal Register 52:34460-34578.

Office of Science and Technology Policy (OSTP) (1985): Chemical carcinogens: a
review of the science and its associated principles. Federal Register
50:10372-10442. Reprinted with additional appendices and references in
Environ Health Perspect 67:201-282 (1986).

The President (1981): Executive Order 12291 of February 17, 1981: federal
regulation. Federal Register 46:13193-13198.

The President (1985): Executive Order 12498 of January 4, 1985: regulatory
planning process. Federal Register 50:1036-1038.

Preuss PW (1987): Reducing the uncertainty in assessing environmental risk.
EPA J 13(5):6-7.

Preuss PW, Ehrllch AM, White PD (1987): Welcoming address. Proceedings of the
conference on grappling with risk assessment: or the frontiers of screme
and law, Granllbakker, Lake Tahoe, California, October 6-8, 1986, to be
published 1n Am J Ind Med.

Preuss PW, Ehrllch AM (1987): The Environmental Protection Agency's risk
assessment guidelines. J Air Poll Control Assoc 37:784-791.

15

-------
Stara JF, Erdreich LS (eds) (1J85): Advances in health risk assessment for
systemic toxicants and chemical mixtures. Toxicol Ind Healtn 1(4):
Dec, 1985.

US Consumer Product Safety Commission (US CPSC) (1978): Consumer products

containing benzene as an intentional ingredient or as a contaminant under
the Consumer Product Safety Act; proposal to ban. Federal Register
43:21839-21852.

US Consumer Product Safety Commission (1981): 8enzene-containing consumer

products; withdrawal of proposed rule. Federal Register 46: 27910-27911.

US DC Circuit Court of Appeals (US DC Cir) (1987): NRDC v EPA, US Court of
Appeals, District of Columbia Circuit, No. 85-1150, July 28, 1987.

US Environmental Protection Agency (1976): National emission standards

for hazardous air pollutants: standard for vinyl chloride. Federal Regis-
ter 41:46560-46573.

US Environmental Protection Agency (1984a): "Risk Assessment and Management:
Framework for Decision Making." EPA-600/9-85-002. Available from Office
of Policy Planning and Evaluation, US Environmental Protection Agency,
Washington, DC 20460.

US Environmental Protection Agency (1984b): National emission standards for
hazardous air pollutants; regulation of benzene; response to public
comments. Federal Register 49:23478-23495.

US Environmental Protection Agency (1984c): National emission standards for
hazardous air pollutants; benzene equipment leaks (fugitive emission
sources); final rule. Federal Register 49:23498-23520. Codified at 40
CFR 61.110-61.112 and amendments to 40 CFR 61.240-247.

US Environmental Protection Agency (1984d): Approaches to risk assessment for
multiple chemical exposures. EPA-600/9-84-008, NTIS No. PB84-182369,
March, 1984.

US Environmental Protection Agency (1984e): Selected approaches to risk assess-
ment for multiple chemical exposures: progress report on guidelines
development at ECAO-Cincinnati. EPA-600/9-84-014a, NTIS No. PB86-226992,
June, 1984.

US Environmental Protection Agency (1986a): Guidelines for carcinogen risk
assessment. Federal Register 51:33992-34003.

US Environmental Protection Agency (1986b): Guidelines for mutagenicity risk
assessment. Federal Register 51:34006-34012.

US Environmental Protection Agency (1986c): Guidelines for the health risk
assessment of chemical mixtures. Federal Register 51:34014-34025.

US Environmental Protection Agency (1986d): Guidelines for the health assess-
ment of suspect developmental toxicants. Federal Register 51:34028-34040.

16

-------
US Environmental Protection Agency (1986e): Guidelines for estimating expo-
sures. Federal Register 51:34042-34054.

US Environmental Protection Agency (1987): "Report of the Environmental Pro-
tection Agency Workshop on the Development of Risk Assessment Methods
for Tumor Promoters." EPA-600/9-87-013, June, 1987.

US Fifth Circuit Court of Appeals (US Fifth Cir) (1983): Gulf South Insulation,
et al v Consumer Product Safety Commission. Federal Reporter 2d 701:
1137-1150.

US Supreme Court (1980): Industrial Union Department, AFL-CIO v American
Petroleum Institute. United States Reports 448:607-724.

17

-------