United States
Environmental Protection
Agency
Office of Policy
Planning and Evaluation
Washington DC 20460
EPA-230-12-b;001
November 1986
Using Risk Assessments
in Policy Decisions
EPA
230
12
87-
001
-------
00
USING RISK. ASSESSMENTS IN POLICY
Ann 1-isher*
Hay 1986
The Environment Directorate of the Organization for Economic
Cooperatio. «nd Development (OECU) has requested that each menhir
count r;, reft; :nl ic authorities. This pa^er describes how the Lr.it^u
States federal government has been using risk assessment in its
public pvliry tU-- ^isi
Because of th^ 'jroad scope of the OLCU questionnaire and the
I* ii'f,l ' i: • available for response, boundaries had to be estab-
lished for the paper's coverage. The questionnaire defines risk
t-«5R"-.ent in. terras of attein^ts to quantify the effects of expo-
ures to individuals. At ^resent, most risk assessment activities
,M---f or-ied by government det»^rt*nents within the Unit.ea Status are
qualitative , rather tli^n quantitative, liany only identity poten-
n i • n bm-an hazard rather thnn estimate how many people will be
ai*:ected by the expected exposures. An attera[)t has been trade tw
i ' , eatrate on quant i tat I:7--1 i-isk assessments, so this paper is not
representative ot the distribution of types ot risk assessment
activities in tht u. -.Lted States. however, it sometimes is not
clear whether a full quantitative risk analysis was conducted, and
several examples of more qualitative risk analyses are included.
*Assisted by Marie Zack
-------
-2-
The questionnaire and this paper exclude nuclear risks. In
the United States, the Nuclear Regulatory Commission conducts risk
assessments because of its responsibility to regulate nuclear
reactors and other sites. Other agencies also regulate radioactive
materials. Since this type of risk assessment is omitted from the
paper, its description or tht situation in the United States is
incomplete.
The term "public authorities" includes those at the national,
state, and local levels. In thi. Lnited States, there is much
cooperative action anon£ these levels. Most of the states have
their own environmental agency, public health agency, food safety
agency, etc. Larger municipalities often have similar independent
departments within their own government. For example, the U.S.
EnxTironmental Protection n^ei.ey (HPn; has initiated integratej
environmental management projects for Philadelphia, Baltimore, and
Santa Clara County. in Li.tfi. projects, the state and local goveru-
ments of Pennsylvania, Maryland, and California are examining the
risks from pollutants in each environmental medium (e.g., water,
air) to identify the n.o&^ invortant sources of risk and the opti-.cl
way of reducing these ris-:s. Because, state and local efforts term
to be limited by shortages of funds and expertise, this paper con-
centrates on the national level.
The depth of inf or" ••*•?'~ -'•- uneven across types of risk and
agencies covered in this paper. The focus is on risks from hazard-
ous substances, whether at the work place or in the general envi-
ronment. Some information is included for other types of risk, to
-------
provide persne.-.cive. A second tocus is on risks to people, kithi
the human health category, quantitative risk assessment (yielding
estimates of the number of cases) has been developed far more for
carcinogens than for noncarcinogenic toxicants. This is reflected
in the paper, even though an attempt has been made to include risk
assessmenLs for all health eftects.
The questionnaire defines risk assessment to include effects
on the environment as well as effects on health. Less information
is available for estimating ecosystem risks, although some eco-
system effects can be assessed by estimating changes in the produc
tivity of particular species. Since so many organisms with con^lc.-
ecological relationships are involved, it is not surprising that
knowledge in this area is less developed. Sometimes risks to othe*
significant aspects of society's well-being are evaluated. These
includp iirpact!- or, visibility, recreation, and material goods. A
feVk examples 01 these other welfare eftects are included.
"Use oi risk assessment" can have multiple interpretations.
At its most basic level, "use" can refer to methodological resn^rc
designed to improve any of the four steps in the risk estimation
process (hazard identification, dose-response assessment, exposure
assessment, and risk characterization; [National Academy of
Sciences (NAS), 19b3, p. 3]). Substantial efforts are under way by
several agencies to asses? a^*. extend the state of the art in .-is^.
assessment and to examine research priorities (U.S. Department of
Health and Human Services, 1985; Ehrlich, 1985; National Science
Foundation, 196^; i\isk Science Institute, 1985). At a more
-------
-4-
practical level, "use by public authorities" refers to using esti
mates from risk assessments as input for policy decisions.
RISK. ASSESSMENT AS PART OF THE POLICY DECISION PROCESS
A 1983 National Academy of Sciences report established a useful
distinction between estimating what the health (or ecosystem)
effects would be in alternative circumstances -- called risk
assessment -- and selecting an action to address the problem --
called risk management. Thus, risk assessment estimates the magni-
tude of health and ecosystem eftects; risk managers then decide how
serious those effects are, and what to do about them.
Public authorities must make two types of risk management
decisions. First, because of limited resources, they must set pri-
orities among the many substances over which they have jurisdic-
tion. To do this, they n,jst. first use some sort of risk assessment
to identify those substances with the greatest relative hazard, so
that they can be controlled first. Agencies have been criticized
because their risk assessments for setting priorities appear to
have been informal and not explained to the public. Some of this
fuzziness in the process of setting priorities depends on how much
control the agency has over the amount and quality of data avail-
able for the risk assessments used to screen substances (Field and
McCray, 1983). For example, Section 5 of the Toxic Substances
Control Act requires the Environmental Protection Agency to screen
new chemicals before production begins. In its premanufacturing
notification, the producer must provide any information on toxicity
-------
th-.t it can find, but the firm is not required to conduct any
toxicity tests. Since toxicity data often are not available, this
forces EPA to consider each new chemical on a judgmental case-by-
case basis. The burden of proving that risk may exist falls on
EPA. On the other hand, the pesticide program reflects a licensing
statute and requires the manufacturer to conduct acute and chronic
toxicity tests to obtain registration for a pesticide. In this
case, the manufacturer bears the burden of proving that a product
is "safe." If the test results exceed qualitative risk criteria
set for acute and chronic health and environmental effects and the
producer decides to pursue licensing, then the more elaborate
weighing of benefits against risks is undertaken in response to
EPA's issuing a "rebuttable presumption against registration."
Ihe second major use of risk assessment is in decisions about
what method and degree of control should be used for a particular
risk. These risk management decisions often are reguj.ai.ury ii.
nature; they restrict the use of a substance or process in order
to reduce exposures to hazardous situations. Other risk management
activities are nonregulatory but are designed to lower risk by dis-
seir.inating information, providing advice to other government
agencies, educating the public, conducting emergency am? retaedia"!
responses (to such things as industrial accidents, spills of
dangerous materials, and abandoned hazardous waste dumps), and
evaluating the effectiveness of specific risk management activities
Risk management is only one component in a complex decision
process. Ideally, the analytical tools used in risr-. management.
-------
-6-
will help determine the most efficienc way to reduce risk. In
addition to the number of cases reduced, the risk manager often
must consider whether the proposed action simply shifts the risk
elsewhere (e.g., to the lane or air instead of water), whether itt
overall benefits exceed its costs, and whether the distributions of
gains and losses are equitable. For the decisions discussed here,
it generally is impossible to tell how important the role of risk
assessment was, compared with the roles of other decision factors.
HOW MUCH RISK ASShSSllENT IS
Actions by Congress, the courts, and the Office of the
President have increased the use ot risk assessment. Many of the
laws passed by Congress to promote safety and protect public health
and the environment require some balancing ot benefits and costs.
Even though risk assessment is not required explicitly by these
laws, this balancing is iir^ossible without predicting how risks
will change unuf r the propo--- action. On the otfter hand, a tew
laws, such as the Toxic Substances Control Act (TSCA), the Federal
Insecticide, Fungicide ana Kodenticide Act, and the Consumer
Product Safety Act, have explicit requirements for risk assess-
ments .
Other laws call for mandatory controls once a hazard has been
discovered, such as the banning ot carcinogenic food additives
under the Delaney Clause, and sections of the Clean Air Act that
require "an ample margin of safety" in setting emissions standards.
The resulting policy decisions clearly require the hazard
-------
-7-
identification step, bjt the rest of tht risK assessment process
may or may not be used,
A tnira category of laws specifies that risks be controller
except whtn this is infeasible, perhaps because technology is not
good enough. Examples include sections oi the Clean Air and Clca''.
Water Acts that require using "best available technology" and tne
section of the Occupational Safety and health Act that requires
using the ir.ost stringent standards that are "feasible." Hazard
identification is essential for such policy decisions, and the
additional steps in risk assessment can be used to evaluate the
effectiveness of alternative control strategies.
The coarts als^ have promoted the use of risk assessment. toi
example, in 19bU the Suprenu Court struck down the Occupational
Safety and Health Administration's (UShA's) proposed standard tor
benzene, because OSHA hart not demonstrated that the standard wouia
significa.irly reduce risk. Even when thn legislation does not
require a risk assessment, this court action has led to an expecta-
tion that one x^'ill be essential for any court challenges. Simi-
larly, a U.S. District Court decision (City of New York v. EP^,
543 f. Supp. 10d4 (SDNY, mi)) has required EPA to compare risks
between ocean dumping of wastes and land-based alternatives, in
decisions about issuing ocean dumping permits.
Host federal agencies heve to conduct their own risk assess-
ments as needed for policy decisions, since few statutes authorize
agencies to require that the regulated parties do so. In general,
-------
-8-
for premarket approval programs -- such as those for pesticides and
for new drugs or f" • : -iMitives under the Federal Food, Drug, and
Cosmetics Act -- tht agency has the authority to require the sub-
mission of data for a risk assessment, while programs that must
show activities to be hazardous before they can be regulated typi-
cally are responsible "n*- *c. jiiring their own data (Clement
Associates, 1981).
The Office of trie President has had an important influence on
the use of risk as ;*•-..- .fit through Executive Order 122yl. Ihis
order requires beneiit-cost analyses of major regulations -- gen-
erally define' as thosr with an annual i'noact of at least $10U
million. Regulatory impact analyses (:-~'*-a attacked first, so the incremental benefits
of further risK reduction- nay be small. novdiices in technology
allow the detecti-> i ot xower and lower levels ut risk., and the
accumulation of dtu^ show- that an increasing number of substances
contain some danger. In practice, many RlAs have not included the
full risk assessments neeaeJ for a complete benefits analysis.
Administrative law often provides for public participation and
comment in regulatory decisions. In addition, some government
decisions are made bv t'w V.''-.i;;e house or Cabinet Council rather
-------
-g-
than by the regulatory agency. Although not required to conduct
risk assessments, the private sector or other government agencies
may do sc in order to influence the agency charged with the ris<-
management decision. The sources and types of information on risk
that are available to decision makers then are quite pluralistic,
with contending interests providing alternative estimates of at
least parts of the risk analysis. Risk assessments conducted out-
side the agency responsible for the management of that risk are not
discussed separately in this paper. Instead, it is assumed that
the agency incorporates such information in its decision process,
and more attention is given to the risk estimates stated to be tht
basis of the regulatory decision.
GUIDELINES FOR RISK ASSESSMENT
Interacc-ncy Coordination
There have been several attempts to upgrade the quality ot
risk assessments by including state-of-the-art developments in
methodology, ana to mak^ risk assessment approaches consistent
within and across government agencies. Effort? to achieve intfer-
agency consistency have been most prominent when the same chemicals
are under the jurisdiction of several agencies. For example,
asbestos can be regulated under four agencies' programs, vinyl
chloride under five, and DDT under two. Part of the reason for
this is because each agency has its own constituency. Of the
agencies regulating asbestos, for example, OSHA is responsible for
Protecting workers, the Consumer Product Safety Commission and the
-------
-10-
Food and Drug Administration are responsible for protecting consumers,
and EPA is responsible for protecting the populace from environmental
exposures as well as protecting the environment. Much of this over-
lap among agencies occurs for carcinogenic substances. When managing
risks such as those associated with automobiles, airplanes, or
industrial accidents is primarily the responsibilty of a single
agency, there has been less of a need for guidelines to ensure con-
sistency among assessments for a particular type of risk.
The Int'^ragencv Regulatory Liaison Group (IRLG) was formed in
1979 by five agencies: EPA, the Food and Drug Administration
(FDA), the Consumer Product Safety Commission (CPSC), OSHA, and the
Department of Agriculture (USDA). Their guidelines for identifying
and estimating the risks of carcinogens were used widely by the
IRLG agencies, although they never were adopted formally (IRLG,
1979) .
In response to a concessional mandate, the National Research
Council of the National Acadeny of Sciences (NAS) examined the use
of risk assessment in the federal government {NAS, 1983). NAS
recommended a clearer separation between the assessment of risks
and evaluation of risk management alternatives. It suggested
developing a set of guidelines that would indicate how agencies
should choose amonc alternative methods for inferring human risk
from limited data. NAS also recommended establishing an independ-
ent scientific Board on Risk Assessment Methods. The board would
be responsible for assessing the evolving scientific basis for risk
assessment, developing and periodically revising the guidelines for
-------
-11-
risk asstssrcent that would be used by federal regulatory agencies,
evaluating the usefulness of the guidelines, and identifying
research neeas. however, diversity in the laws for which agencie.*
are responsible, and the differences this implies for the scien-
tific needs across agencies, led NAS to recommend that risk assess
ments continue to be conducted within the agencies, rather than by
an independent risk assessment agency.
President Reagan's Office of Science and Technology Policy
(1984) prepared an update ot the principles and science of
assessing carcinogenic risk. Its proposal includes policy state-
ments about some scientific principles, but leaves most decisions
about the selection and use of risk assessment methods to the
agencies. During the same time period, then EPA Administrator
Iviilian Ruckelshaus set up a successor to the 1KLG, called the
lnterageTicy \isk Management Council. It was to draft its own
cancer guidelines and reproductive toxicology guidelines, and
coordinate several specific risk management activities across
agencies. However the council is inactive at present.
Intra-agency Guidelines
Some individual regulatory agencies have developed risk
assessment guidelines for their own use. For example, FDA (within
the Department of Health and Human Services (DHHS)) developed
acceptable daily intakes for noncarcinogenic subst^n^es, by applying
safety factors to no-observed-effect levels of exposures from
laboratory experiments (Lehman et al., 1959). Although FDA has not
-------
-12-
adopted formal guidelines for risk assessment, it continues to use
this approach for noncarcinogenic food additives, and other
agencies use the approach for pesticides and some exposures to
occupational and environmental contaminants. More recently, DhhS
prepared a document stating policy for the generic use of risk
assessment and establishing a boundary separating risk assessment
and risk management for all UHHS agencies (DHHS, 19b5).
EPA was the first agency to publish internal guidelines for
assessing carcinogenic ristc (EPA, 1976). It has published updates
as new information has become available. The most recent version
was proposed in 19&4, and was accompanied by proposed guidelines
for assessing exposure, mutagenicity risk, and risks from develop-
mental toxicants (EPA, 1984). Shortly thereafter, EPA proposed
guidelines for assessing the health risks of chemical mixtures
(EPA, 1985a). These five sets of guidelines have been reviewed by
EPA's Science Advisory Board, and currently are undergoing final
review by EPA and the Office of Management and Budget. EPA also is
developing guidelines for assessing risks to luale and female repro-
ductive systems, and for making and using exposure measurements.
Other agencies are moving toward using additional steps in the
full risk assessment process. For example, in response to recom-
mendations fror. the NAS and the General Accounting Office, the Food
Safety and Inspection Service (within the U.S. Department of
Agriculture) is developing an approach for using a hazard index
with an exposure index to assess the risks of food ingredients.
-------
-13-
DIFKICULTILS OF CONDUCTING RISK ASSESSMENTS
Regardless of whether guidelines are available, risk assess-
ment typically is a complex ana costly process. Data often are not
ideally suited for the needs of a particular risk assessment. It
may be difficult to predict the physical phenomena that will lead
to the release or creation of a particular toxicant, or to a par-
ticular typt. of accident. Once released, there often are gaps in
the scientific understanding of how a contaminant will move through
the air, water, and land, and how it will be attenuated by other
influences before it reaches people or the ecosystems of concern.
Typically, the dose-response function ir.ust be extrapolated far
below the responses observed for the exposures used in laboratory
animal tests. Since ethical considerations usually preclude exper-
iments with humans, some conversion must be made to find the huiuan
dose that is equivalent to the doses used in the experimental
animals. Lven when epidemiological data are available, it often is
difficult to draw inferences for the exposure levels being con-
sidered tor regulation (Fisher, 1982). Similar problems often
exist for assessing accidental risks, in terms of predicting their
probability, magnitude, and irnacts on people and the environment.
These gaps in our scientific understanding and data limita-
tions imply that it is difficult to conduct a good risk assessment.
It is no surprise that they vary in quality. The many stages where
judgment must be applied make it very eaty for the results to sub-
stantially overestimate or underestimate the unknown true risks.
because a government agency's mandate typically is to protect the
-------
-14-
public, or to be safe rather than sorry, the cumulative effect of
these conservative assumptions may be very large. The resulting
risk estimates often are treated a? plausible upper bounds. Unless
the uncertainty associated with each assumption is stated, risk
managers often view these risk estimates as actual risks.
Some gauge of the expected error can be developed by iden-
tifying those stage? ot tr.t analysis where the uncertainties are
greatest, and then conducting a sensitivity analysis to see how
important errors in tnose stages would be for the final estimates.
Of course, this also can pinpoint where better information woula
have the greatest pay-o-" ~ r- terms of improving the risk assessment.
USES OF RISK ASSESS^"-
This summary follows the broad outline of the OECD question-
naire, whic!". divides ris ;s- into five categories. These are acci-
dental risks, continuous risk^, risks occurring in managing natural
resources, risks from d: • - • *-ous installations, and risks associated
witn alternative stre* ' fjr managing hazardous wastes.
Attempts were made tc r--* c .- -rizc the activities properly, but there
is considerable overlap in some of these non-exclusive categories.
Accidental _Ri_sk_s
Some contamination occurs more or !•:•;<: continuously, while
other coi^an.ination result? from sudden incidents, or accidents.
This section describes activities using risk assessment to control
the risks associated with accidents.
-------
-15-
t.S.Environmental Protection Agency
At present, there are few risk assessments for manufacturing
accidents. huguenin, Fich, and Klkind (1985) fouric very few pub-
lished studies with quantitative risk assessments of chemical manu-
facturing and associated activities. Of the ones available, most
seerc to have been conducted to assist in the firm's decision
process, rather than for public authorities (e.g., Boytcin, treeuan,
and Levary, 1984). The studies generally consider the risks of
storing or transporting chemicals, rather than the risks of manu-
facturing or processing them. The risks examined usually are the
effects 01 fires and explosions, rather than potential toxic
effects. huguenin, Pich, and Elkind were able to find only a total
of eight studieb assessing the toxic risks from storage releases,
which include the chemicals ammonia, chlorine, acryionitrile,
sulfur trioxide, and hydrogen tluoriae. They found only one risk
analysis of chemical manufacturing, and that was for the
Kethfc*•]£-•'=. This indicate <, serious information ga^, since pre-
liminary analysis of historic release incidents suggests that
releases from processing operations are about as lar^e and occur
about as often as releases from storec* chemicals. It is relatively
difficult to conduct risk assessments of process operations,
because safety systems and operator procedures and errors tend to
ditiei. across plants yet will influence failure rates and quanti-
tieb rele-Mi.-. Much of this information may be t - •; ":>tary, which
would help in explaning why the literature is so sparse.
-------
-16-
tlPA is working to fill some of the gaps for accidental risks.
For example, the Acute Hazardous Events Data Base recently has been
assembled to providt information about dangers to the U.S. pualir
and workers from sudden accidental releases of toxic substances
(Industrial Economics et al., 198b). This data base was designed
to provide perspective on recent accident history in the United
States, but is not sufficient as a basis for nationwide estimates
of frequencies of events, amounts released, or their consequences.
Most of the 3,121 records in the data base are for 1983 and 198^.
Although the records report the number of injuries and deaths, rrost
do not have specific information on causation. However, releases
of chlorine, ammonia, hydrochloric acid, or sulfuric acid occurred
in over one-fourth of tht events leading to deaths, and the clast
of industrial organic chemicals had the most events with deaths.
Releases associated with deaths varied from small to large, and
inherent toxicity was not uniformly high, Toxicity seems to nave
caused rnost of the injuries reported, while fires and explosions
are associated with most of the deaths. One-fourth of the events
in the data base are Lransporcation releases, accounting tor one-
third of death or irjury events. These shares may change when the
data base is expanded to include more records.
An effort to respond to risks from accidental releases of
toxic chemicals is EPA's Chemical Emergency Preparedness Program
(CEPP). EPA's Science Advisory Board assisted the Office of
Pesticides and Toxic Substances (UPTS) in setting up criteria so
that companies and coir-rrunities could identify -what chemicals are
-------
-17-
of most concern anci how to prevent or respond to their accidental
release. The criteria rely primarily on acute toxicity effects,
with consideration of volume and chemical properties such as flam-
raability, reactivity, and the potential for explosion. EPA has
issued a list of 402 chemicals that are considered to be hazardous.
LP1.'s Superfund Office provides site-specific guidance to corntruni-
ties, witn UF1S examining the possibilities that the substances
will vaporize and how far the plume will travel and be of concern.
This does not constitute the tull risk assessment process, since
the focus is on hazard identification and the number of potential
cases is not estimated.
National Environmental Policy Act
The !,ational Environmental Policy Act (4U CFK 1i>00-1bUb)
requires U.S. federal agencies to prepare an environmental impact
statement (EIS) for aajo^ federal actions significantly affecting
the quality ot tlie human environment (section 102(2)(c)). Examples
include issuing federal permits tor construction, agency land man-
a^e~c--.t plans, ana ledera! construction projects. A larger siiare
of r-iSs have included risV assessments in the pasc few years,
reflecting the increased use of risk assessment in the regulatory
process. V^hen an EIS contains a risk assessment, it usually pro-
vides the federal decision maker with an estimate of broadly
defined environmental consequence0 of the proposed action and a
range of alternatives. For example, EPA prepares such an Lib
-------
-18-
Many of the ElSs that include some form of risk assessment arc-
concerned with accidental risks in energy production. For example,
the U.S. Army Corps of Engineers (19»4a) evaluated the probability
of an oil spill and effects on nearby sensitive ecosystems for the
Endicott hydrocarbon development project. For the Georgetown
refinery, the Corps (1984b) assessed the effects of a potential oii
spill from transportation and operating activities. The b.S.
Federal Energy Regulatory Commission's (1983) risk assessment for &
natural gas storage plant analyzed the potential for pipelines to
rupture. The b'.S. bureau of Land Management (1983a) examined the
probabilities of natural gas well blowout and pipeline rupture anr
the effects on people of h2b gas. In analyzing the sale of oil ana
ga& leasts, the U.S. Mineral Management Service (1983) assessed t,.t
indiviaaal and joint probabilities of oil spills and the resulting
impacts on target ecosysterns. Tne EISs tend to emphasize ecosyste;
effects, but they also consider human health effects.
Uej^artment of Energy
The U.S. Department of Energy (DOh) is completing a facility
aL the Nevada lest Site that will allow simulation of how winds ana
gravity affect large scale releases of cryogenic or pressurized
flararaables, toxic substances, and heavy gas (b.S. DUE, no date).
The facility is designea to provide information for risk assess-
ment, re^'ii^ior,, plant c'csijrr., plant siting, and hazard mitiga-
tion. An extensive sensor and data system is available to acquire
data on spill characteristics such as rate, volume, temperature and
-------
-IV-
pressure, J .> r ""•.'. gas concentrrst ion a = i • aerosol characteristics,
meteorological parameters, ana blast or nre effects.
T I,- i'rtO. iLity will provide H s Li.* f:>~ research and development
t'anded by the private sector although the government agencies"ralso
may use it, Trie user is t • oi<.»rdlu: hazard,) in facility siting anu
design. D-, ". requires SAKs for it;* n^-: ilacilities and major modi-
fications of i!--< existing facilities. Several existing facilities
have btrr- , .-aluated with SAKs, including its national laboratories.
DUE also is conducr.ii- enviroTr.er.':?.! surveys tor its tacilities.
"i.ic^- sjrvtys will allow prioviciei to be set for prulie:ns at Out
sites, based on relative risks to humans as modifiec by other con-
siderations such as ecosystem
Continuous Risks
EPA, the Consumer Product Safety Commission (CFSC), the Food
and Drug Administration (FDA), the Department of Agriculture's
Food Safety and Inspection Service, and the Occupat"1 ^~ •" Crfsty and
health Administration (OSHA) nre responsible tor controlling risks
caused by the continuous or repetitive release of pollutants in the
environment. Soce of them are reSiionsible for addressing the risks
-------
of natural cc-.tarinants (such as aflatoxin) or the risks of certain
manufacturing practices or additives in food. These responsibili-
ties anc sor.e recent atcis i-•-£ are summarized in this section. In
order to provide a context for risk estimates, it is helpful to
know that there are about 235 million people in the United States.
Many of the estimated risks are for cancer. About 472,000 people
in the Unitec States are expected to die from cancer in 1986.
Roughly 930,000 will have cancer diagnosed for the fi.rst time, and
about 2 million people have active cancer at any point in time.
U.S. Consumer' Product. Safety' Commission
CPSC regulates consumer products under the Consumer Product
Safety Act anc the Federal Hazardous Substances Act. The CPSC
staff submits risk assessments and other information to the
Commission for use in d-ciding whether to seek mandatory or volun-
tary standards or to deel with the issue in some other way. These
risk assessments becorr.c pa*--_ of the Commission's public record.
Only a fev; of the continuous risks summarized in Table 1 are dis-
cussed in this secLic-..
In 1974, CPSC estimated that 6,000 cancers per year would ,
result frorr. the continued use of 2CJ million spray cans produce-
annually between 1968 and 1973 with vinyl chloride as a propellaiiU,
Even though more cost-effective chlorofluorocarbon propellants were
being adopted, CPSC banned the use of vinyl chloride in household
aerosol products. When the ban became effective, only about one
-------
-21-
Table 1
Consurrc-r Product Safety CommissionDecisions
Substance
Vinyl chloride
(propellant}
Tris (llamt.
retardant)
Asbestos fibers
Type of Risk
Cancer
Canctr
Cancer
Action
Asbestos fibers
(other products)
Benzene (paint
removurt)
Benzidine dyes and
congeners in home
dying products
Nitrosamines in
rubber pacifiers
and nipples
Urea formaldehyde
foam insulation
Methylene
chloride
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
1974: banned use in household
aerosol products.
1977: banned sale of chil-
dren's garments treated with
Tris.
1977: banned patching and
joint compounds, and artifical
emberizing materials, containing
asbestos.
1985: aggregate risk estimate
of less than 10 cancers per
year. Recommendation has not
been determined.
1978: proposed ban on con-
sumer products containing more
than 0.1% benzene.
1981: withdrew proposal.
Manufacturers voluntarily
removed benzidine dyes, and
all but one have removed
benzidine congener dyes from
home dyes.
1984: banned rubber pacifiers
containing nitrosamine levels
higher than 60 ppb (which is
consistent with FDA action on
rubber nipples).
Ban prevents 13 cancers p«r year.
Now estimating risks from paint
strippers and aerosol paints.
Cancer
Industry reportedly has removed
DEHP from many children's pro-
ducts; staff is reviewing pro-
posed voluntary standard for
pacifiers and teethers.
-------
-22-
million cans per year remained to be subject to the action, for an
estimated saving of 26 cancers per year. (Most consumer uses of
chloroflurocaroon propellants were banned subsequently by EPA and
FDA, in cooperation with CPSC.)
In 1977, in response to its estimate that 540 cancers per yec.r
year would result from 25 million children's sleepwear garments,
the CPSC determined that Tris was banned under the Federal Hazardojs
Substances Act. Some manufacturers already had eliminated Tris fro"
their products in rt^-rise to information about the cancer hazard.
Also in 1977, CPSC banned patching and joint compounds con-
taining asbestos for use in wall repairs and dry wall construction
after estimating that 680 cancers per year could result from
exposure to 40 million five-pound packages. Since then, several
other products containing asbestos have been studied by CPSC.
Their aggregate risk is snail because the use of these products is
low and declining. CPSC is considering what additional action, if
any, it will take for these products.
CPSC estimated 615 cancers per year would result from benzene
in the 1974 formulations of paint rem^v°rs. By 1978, most pro-
ducers had switched to methylene chloride, reducing the risks from
benzene to 22 cancers per year. In 1981, producers no longer were
using benzene in paint removers and CPSC withdrew the proposed ban.
Methylene chlotiue is a less potent carcinogen than the benzene for
which it has been substituted, and CPSC is now estimating the risks
from its wide use in paint strippers and spray paints.
-------
-23-
Although urea formaldehyde foam insulation had been installed
in 550,000 homes, with more than a million people estimated to be
exposed, CPSC's bar, of ne* installation was estimated to prevent 13
cancers per year. This does not include other acute and chronic
health effects that would be prevented. (After a year, the ban was
vacated by the 5th Circuit Court of Appeals.)
The Commission1s staff also estimated the health effect poten-
tial fron pressed wood materials made with urea-formaldehyde (fiber-
board, particleboard unaer 1 ^y...*. r.t, ar.^ hardwood plywood wall panel-
ing) and used in the construction of single family detached houses.
About 24 percent of new houses include some of these pressed wood
materials. The 95 percent uppt:r confidence limit estimates indicate
an excess risk of 19 to 143 cases pe*- million people exposed.
Based on the number of houses b^ilt each year, up to 34 excess
cancers per year are estimated. The pressed woo^ industry appears
to be ta';inj action to reuuce e.;.i at ions. The CPSC is considering
what, if any, additional action is needed.
The plasticizer DEHP has been usea for a Ion:; time in many
polyvinyl chloride products. Studies demonstrated that this
carcinogen could leach out of the products and be absorbed readily
through ingestion. This led to CPSC concern about vinyl pacifiers,
with risk estimates of up to 7 cancers per million live births if
all babies use pacifiers containing DEHP. There are about 3.C ""'i-
lion births per year in tht United States. The industry reportedly
has removed DEHP from nearly all vinyl pacifiers and teethers and
is preparing a voluntary standard.
-------
-24-
U•S. E n v i ronme nt a1 Protect ion Ag e n cy
EPA is the U.S. agency with the broadest range of responsibil-
ities for controlling continuous risks. Its programs cover conta: -
inants in the air, water, and land, and include the production of
toxic substances, the use of pesticides, and the fate of hazardous
wastes. Table 2 lists EPA's regulatory impact analyses (RIAs).
Many of these RIAs include risk analysis. However, some of the
RIAs that have analyzed risk include only parts of the risk
assessment process. In addition, risk assessment sometimes is used
in decisions that do not require RIAs. Background documents often
are available for these risk assessments.
Perhaps the most extensive use of risk assessment within EPA
has been in its air program decisions. Its first use was in the
air topics program, for developing National Emission Standards for
Hazardous Air Pollutants (NESHAPs). Decision makers wanted to know
whether a particular level ot control would eliminate unreasonable
risks — in terms of both aggregate population risk and maximum
individual risk. Quantitative risk estimates were not used in
setting NESHAPs in the early 1970s for mercury, beryllium, and
asbestos. Quantitative risk estimates were used first in setting
the vinyl chloride NESHAP in 1976 and have been fundamental to
recent risk management decisions on NESHAP actions for benzene,
arsenic, and radionuclides.
The role of risk assessment has led to substantial changes i:<
regulations proposed for benzene. For example, a 1980 analysis of
emissions from maleic anhydride plants predicted a cancer about
-------
-25-
Table 2
EP/.' s Regulatory Impact Analyses
Air
Listing of Surface Coal Mines for New Source Review,
September 198E>, draft.
National Ar..oient Air Quality Standards for Particulate Matter,
March 1983.
National Ambient Air Quality Standards for Sulfur Dioxide,
forthcoming.
National Ambient Air Quality Standards for Carbon Monoxide,
July 1985.
National Ambient Air Quality Standards for Nitrogen Dioxide,
June 1985.
Cost? and Benefits of Reducing Lead in Gasoline, February 1985.
Control of Air Pollution from New Motor Vehicles and New Motor
Vehicle Engines, March 1985.
Radiation
i Final Environmental Standards for Uranium Mill Tailings at
Active Sites, September 1983.
Economic Impacts of 40 CFK 191: Environmental Standard for
Management and Disposal of Spent Nuclear Fuel and High Level,
Transuranic Radioactive Waste, August 1985.
Draft Economic analysis: Proposed Rules for Radon-222 Emissions
from Licensed Uranium Mills.*
Water
Effluent Limitation Guidelines Regulation for the Iron and Steel
Industry, March 1982.
Water Quality Benefits of the BCT Regulations for the Pulp and
Paper Industry, July 1985, draft.
An Economic Assessment of the Benefits of the Proposed Effluent
Limitation Guidelines for Organics and Plastics Manufacturers,
September 1984.
-------
-26-
Table2 (Continued)
Proposed Regulations to Control Volatile Synthetic Organic
Chemicals in Drinking Water, May 1985.
An Economic Assessment of Reducing Fluoride in Drinking Water,
November 1985.*
Solid Waste
Revisions to the National Oil and Hazardous Substances Pollution
Contingency Plan, February 1982.
Proposed Standards for the Management of Used Oil, November 1985.
Regulatory Analysis of Proposed Regulations under RCRA for Srrall
Quantity Generators of Ha^ardous Wastes, June 1985.
Reportable Quantity Adjustments under Sections 102 and 103 of
CERCLA, March 1985.
Risk and Cost Assessment of Hazardous Waste Incineration
Regulation, 1984, draft.
Regulatory Analysis of Proposed Restrictions on Land Disposal of
Hazardous Wastes, December 1985.
Regulatory Analysis of Proposed Restrictions on Land Disposal of
Certain Dioxi n-Containir.c Pastes, January 1986, draft.
Regulatory Analysis of Proposed Restrictions on Land Disposal of
Certain Solvent Kastt?, January 1986, draft.
Toxics
Proposed Regulation of Glycol Ethers, June 1985, draft.
Proposed Rule on the Use of Nitrites in Metalworking Fluids,
September 1985, draft.
Final Rule for Nonsubstation PCS Transformers, June 1985.
Final Rule Regulative Ir advertent PCB Generation for
Uncontrolled Sources, April 1984.
Final Rule for PCB-Contiiiiing Electrical Equipment, July 1982.
Controls on Asbestos and Asbestos Products, August 1985.
Indicates reports art, n-jt. formally designed as RIAs.
-------
r
-27-
Table 2 (Continued)
Pesticides
Economic Analysis of Regulations Implementing Certain Portions
of FIFRA, Section 3, Concerning Registration of Pesticides,
May 1982.
Proposeo Rules Governing RPAR Proceedings: Criteria for the
Initiation of RPAR Proceedings, May 1982.
RIA of 40 CFR, Part 166: Emergency Exemption Regulation,
February 1985.
Data Requirements for Registering Pesticides under the FIFRA,
August 1982.
-------
-28-
every two year?, leading to proposed controls that would reduce the
health effects to one cancer about every ten years. Significant
changes then occurred in the industry structure and the controls
that were in place. A reevaluation in 1984 yielded an estimate of
existing emission levels causing one cancer about every 35 years,
leading EPA to withdraw its 1980 proposal. Similar analyses for
ethylbenzene/styrene plants and benzene storage vessels led to the
conclusion that risks to public health are small from these
categories of emissions and that there are no significant health
benefits frorr, controlling them.
EPA takes several steps before promulgating a NESHAP, and
several chemicals are somewhere in this process. Table 3 lists 22
recent decisions relying on risk assessment in the air toxics pro-
gram, many of which rely on quantitative risk assessment. The one
that is closest to final action -- aside from those where the deci-
sion is not to regulate -- is the listing of coke oven emissions,
for which a NESHAP now is being prepared. (However, RIAs are not
prepared for NLSHAPs, since they are not classified as majc~ r•.•---
lat ions.)
Perhaps the most interesting NESHAP is the proposed rule for
inorganic arsenic emissions from copper smelters. On the cost-
effectiveness criterion, the proposal calls for using the best
available control technology on six of the 14 existing smelters
emitting inorganic arsenic, which would save 0.23 lives each year
(see Table 4). Since this proposal might mean no additional con-
trols on some smelters that could pose greater health risk, EPA
asked for comment on two alternatives.
-------
-29-
Table 3
Status of Decisons on Air Toxics
CHEMICAL
Acrylonitrile
1,3-butadiene
Cadmium
Carbon Tetrachloride
Chlorofluorocarbon 113
Chlorinated Benzenes
Chlorof orir.
Chloroprene
Chromiuir.
Coke Oven Err.issior.L
Epichloronydrin
Ethylene Dichloride
Ethylene Oxide
Hexachl?y cyclopentadiene
Manganese
Methyl Chloroform
Methylene Chloride
Perchloroethylene
Polycyclic Organic Matter
Toluene
Tr ichloroethylene
Vinylidene Chloride
ACTION
Referral
Intent to List
Intent to List
Intent to List
Not to Regulate
Not to Regulate
Intent to List
wot to Regulate
Intent to List
listing Notice
Not to Regulate
Intent to List
Intent to List
Net tr> Reaulate
Not to Regulate
Not to Regulate
Intent to List
Intent to List
Not to Regulate
Not to Regulate
'- '•f-nt to List
Not to Regulate
DATE
June 1985
October 1985
October 1985
August 1985
June 1985
August 1985
September 1985
September 1985
June 1985
September 1984
June 1985
October 1985
October 1985
October 1985
August 1985
June 1985
October 1985
December 1985
August 1984
May 1984
December 1985
August 1985
-------
1—1
<0
r-
6
4J
•D
•O
< tf
C
03 O
•c w
•rH -H
> U
O Q)
Lt Q
o, 1
D L,i
•n O
«* C JJ
1 (U (0
o a* (o — i
ro i-i Li 3
i -D < a
U-l
Ll
C re
C -H
1— 1 L:
Q)
jj ( i
C 'rl
y-H1 Lt
U
03
•r-*
03
>
1—!
fO
c
<
(D O
•H -r-l {/).
0 iJ ^
-r4\
-< JJ-D
r*- u"i
rO rH
e o >
M CJ
Q> Q) >•<
•H JO Li
.J CO >
X X
00 00
oi -H • •
T oo m
M
U3
4J
K
0
CJ
-i —
e J.
N r-l —
• i-l r-4 VC
r-! C 0 VD «*
(0 Ll r- 1 ' • •
3 4J <^- OC. CO
Z £ -^
C O
<: u ;
U_; ^
O C' 03
4J -U
LJ (TJ C'
O i— i ",
03 4J
0) -^
c to
Q) C
.5 2
4J
J>^
Li C
0 0
ro LI
3 -U
0) Ll
K U
O C
(!) O
n 1 t i
4-1 4J
H (0
1 r-l
4J 3
03 a
o o
U &<
O
a,
-r( O
"D OH
C
M i—l
m
o u
•u O
H
j^ T3
03 C
•-< (0
cr
03
3
0 Li
T3 flj
Ll 4J
^g -H
(0 CJ>
Li K
0 -H
y-i (Q
t/3 0)
13 -D
Li CJ
•n fc
C 00
+J
V3 i
•
C O
O -n
•rl C
1" c.
03 tl
•r-l L;
E <•
"" O
fO C
C (0
O c~
•r-l Jj
1 > rt
C C
2 M
£
£
• UJ
J>s,
L i:
C -D
0 Li
tr r.
^ *o
c
c c
C 4->
•H CO
Jj
u -o
(D Q;
•U 03
0 0
s- a
0. 0
Li
i—1 t,
(D
ij
C ••
O 03
E -u
c c
O nj
U 4J
•H 3
> r-l
C r-l
u o
04
u: LI
• -rl
^ <^
• »
0)
o
Li
3
O
Oj
•
m
03
^
rH
»
C
(N
>,
-------
-31-
Ths first alternative would require controls for plants in
areas with high population density and inorganic arsenic emission
rates greater than 25 kg/hr. Plants in lower density areas would
have to install the controls only if their emissions exceeded
35 kg/hr. Only three plants would have to install controls, and
0.22 lives would be saved each year. In the second alternative,
the only sources controlled would be those with unacceptable com-
binations of maximum individual risk and population risk. The
hypothetical corr.bination described in the regulatory package would
lead to control on five smelters, yielding the same maximum indi-
vidual risk at th<= proposal "nd the first alternative, but saving
0.39 lives per year. Because different plants would be regulated
under each cption, Table 4 shows the.I the cost per life saved
varies substantially across options, even though the maximum indi-
vidual risk stays the same. This shows how risk assessment can be
used to illustrate the real trade-ofts that must be made in such
regulatory decisions.
Six contaminants (carbon monox'd , lead, ozone, nitrogen di-
oxide, partic^lor matter, and sulfur oxides) have been designated
as criteria pollutants, and National Ambient Air Quality Standards
have b?r- set for then:. The Clean Air Act is relatively explicit
about how EPA must deal with criteria pollutants, but some form of
risk assessment is beginning to be used for the periodic reviews
of these ambient standards. For example, the risk assessment
methodology for the lead standard review recently was reviewed by
EPA'S Clean Air Scientific Advisory Committee, and the ozone risk
assessment is underway.
-------
-32-
EPA also regulates air pollution from mobile sources. The RI.
for the recently proposed regulations for particulate emissions
from heavy-duty vehicles includes health benefit estimates based r-
predictions of cancer and noncancer morbidity and mortality. It
turns out that benefits from improved visibility and reduced
soiling are higher than the benefits from decreased health damages
for this regulation. Thjs illustrates the importance of including
all categories of damages, rather than concentrating solely on
human health damages.
Another important risK assessment was conducted as part of
the RIA for reducing lead in gasoline. Lead has been a common
gasoline additive to reduce engine knock and valve wear. However,
evidence has accumulated that the level of lead in human blood is
closely related to the amount of lead used in gasoline. Lead has
been shown to cause anemia, behavioral disorders, mental retarda-
tion and nervo a&race i:. caildren; adult symptoms range fro... head-
aches and irritability to stupor, coma, and brain damage. Table 5
shows estimates for the health benefits for the first few years of
the lead phasedown. TV numbers are smaller for the first year
because the standard was lowered from 1.1 grams per leaded gallon
(gplg) to 0.5 gplg for only the second half of 1985, and because
the standard was tightened further to 0.1 gplg at the beginning of
1986.
EPA recently estimated that 75-80 cancers per year are caused
by the combined effects of benzene and gasoline vapors that are
released when vehicles are refueled. Although the analysis 0^:7-
-------
-33-
inally focused on the volatile organic compounds released in gaso-
line marketing, it now appears that the regulatory decision may
place more weight on the rather substantial ozone releases that
accompany gasoline refueling. At this time, EPA is considering a
nationwide on-board strategy to control refueling emissions.
Under the Toxic Substances Control Act (TSCA), EPA can pro-
hibit, restrict, or regulate the use of any substance that presents
"an unreasonable risk of injury to health or the environment." Risk
analyses recently w^rc usec. cu make decisions about priorities
under Section 4 of TSCA. The compounds 4,4-methylenedianiline and
1,3-butadiene were foun~3 to pose risks primarily in the work place,
and EPA has formally referred these to OSHA for action. Also on
the basis of (low) risk estimates, EPA informally referred formal-
dehyde to OSHA.
A comDination of risk and economic analyses for glycol ethers
showeo tiiu^ tr.cy shojlr! be banned for consumer use but regulated
for industrial use, because of their teratogenic and reproductive
risks. LPA's Office of Toxic Substances (OTS) is preparing a
formal referral to OSh« for these substances. OTS also has
examined the use of nitrites in metal-working fluids where they can
form carcinogenic nitrosamines. It may propose a ban, acccompanied
by a request for comments since the cost per case avoided is very
hign L... some industries. The language of TSCA has been inter-
preted to call for balancing the health benefits of restricting
chemicals with the costs imposed by the restrictions.
-------
-34-
Table 5
Impacts on Adverse Health Effects as Lead Is Reduced in Gasoline
Reduction in number of
children having more
than 25 micrograms per
deciliter in blood
Reductions in blood-
pressure-related effects
in males age 40-59*
Hypertension
Myocardial infarction
Strokes
Death
547,000
1,500
324
1,497
1,796,000
5,323
1,109
5,134
1,718,000
5,12f
1,068
4,942
*Since the study reporting adult health effects had not completed
the peer review process when the phasedown decision was made,
only child health effects were considered (along with reductions
in conventional pollutants, maintenance savings, and improved
fuel economy) on the benefit side of the analysis.
-------
-35-
OTS also us<=>d risk analyses for decisions about PCBs. in
addition to the ban on new uses of PCBs, OTS recently issued con-
trol regulations for existing high-voltage transformers that
could fail and become involved in fires.
EPA's risK assessment for asbestos considered only lung cancer
and mesothelioma, omitting asbestosis. Excellent epidemiological
information is available for these health effects, compared with
that for m^ny other dangerous substances. The asbestos analyses
recognize that OSH.'. n.ay change the occupational exposure standard
from two fibers per cubic centimeter (f/cc) to 0.5 or 0.2 f/cc, so
risks are estimated under all three potential standards. The
number of cases avoided was estimated for each of eight control
options and for banning 32 products. The asbestos ban and phase-
dovkTi rule proposed in January 1986 is estimated to prevent a total
of 926 cancers under the more stringent occupational exposure
standard. Without EPA's proposal, it is estimated that 1,325
cancers would occur. If OSHA does not tighten the occupational
exposure standard, then EPA's proposal is estimated to avoid 1,854
of 2,562 cancers.
EPA also uses risk assessment extensively in its pesticides
program. The Federal Insecticide, Fungicide and Rodenticide Act
(FIFRA) requires EP?. to examine whether the risks from using a
pesticide are worth the economic benefits (such as increased crop
yields or lower costs of achieving a given crop yield). Risks con-
sidered range from those affecting endangered species to human
cancer effects.
-------
-36-
A recent action concerned the use of diamonizide, commonly
known as Alar. Its primary uses were on apples, grapes, cherries,
and peanuts, mostly to make a crop ripen at a uniform time, improve
appearance, and increase storage life. The risks associated with
Alar's use are partly attributable to UDMH, the compound formed
when Alar breaks down. Initial risk estimates showed a one in a
thousand lifetime cancer risk, with an upper bound of one in a
hundred. On the basis of these very high risk numbers, EPA was
planning to restrict Alar. However, the Science Advisory Panel (of
EpA's Science Advisory Board) found serious flaws in the studies
behind the risk estimates, and concluded that there simply was not
enough information available to calculate the risk from Alar's use.
At this time, the only action by EPA has been to require additional
health risk testing by the manufacturer. The results are not yet
available.
Another recent decision under FIFRA restricted tht use of
ethylene dibromide (EDB). The major use of EDB has been as a fumi-
gant for application to the soil before planting crops. It also
has been used to fumigate crops after harvesting, for spot fumiga-
tion of grain milling machinery, and a few other minor uses. Table
6 summarizes the risk estimates for occupational exposures an. for
the average consumer. As required by FIFRA, these estimates were
weighed along with information about economic impacts in the deci-
sions about restricting EDr,. The risks that EDB would leach to
ground water and contaminate drinking water supplies were high
enough that EPA issued an emergency suspension order (followed by
cancellation) to discontinue the use of EDB to fumigate soil before
-------
-36a-
crops are planted, estimated to avoid 65 cancers per year. (As is
the case with most risk assessments, the risk estimates based on
laboratory animal studies could not be confirmed with
epideraiological data. The limited studies available for popula-
tions exposed to EDB during its manufacture did not show an excess
of cancer.) EPA also cancelled the use of EDB for fumigation of
stored grain and for spot fumigation of grain-milling machinery,
(avoiding as many as 800 cancers per year), as a quarantine fumi-
gant to eliminate fruit flies from citrus, tropical fruits and
vegetables (avoiding up to 325 cancers per year), and for fumi-
gating felled logs to control bark bettles. Registration was
continued for other minor uses but with requirements for more
restrictive labc? directions, protective clothing, submission of
monitorir.. and use data.
Under the Clean Water Act, EPA establishes effluent limitation
guideline-;- fcr iniustris;, and ir.onlcipal waste-water treatment
-------
-37-
Table 6
Cancer Risks from EDB
OccupationalInhalation of
Soil injection
Quarantine fumigation
Spot fumigation
applicator
mill worker
Felled log treatment
Lifetime Probability
Estimates
3.5x10-3 - 3.5x10-2
3.5x10-4 . 2.8x10-1
7.7x10-3 - 1.0x10-*
I.7x10-2 - 2.0x10-2
9.5x10-2
Number
Exposec
4b-60
2,400-6,C( f
16,000
uniii.own
Average U.S. Consuire-
Soil incorporation
Whe-3" grain (bulk fumigation)
Spot (wheat grain -
milling machinery
Fruit fuiiiigation
1.1x10-5
3,3x10-3
2.4x10-^
1.7vl?-5 _ 2.3x1O-
SOURCE: Tables 6 and 11 in £.thylene Dibromide (EDB) Position
Document 4, Offi^p of Pesticide Programs, U.S.
Environmental Protection Agency, September 27, 19fc3.
-------
-38-
facilities. Since these guidelines are technology-based, risk
assessment is not required.
Also under the Clean Water Act, EPA develops water quality
criteria and water quality standards, Risk assessment methods are
used in evaluating scientific information to find a concentration
level that will protect aquatic life and human health. The work is
coordinated wit^ other federal agencies such as FDA to develop
acceptable levels. The water quality criteria (based on risk
assessratnt) art, coupieu with thu water use objective (aquatic pro-
tection, drinking water source, navigation, etc.), and with com-
ments froni public hearings. This information leads to a regulatory
water quality standard that gives concentration levels that will
protect the designated uses of a particular body of water.
EPA also is developing risk assessment methods for the dis-
posal of sew a.. ~ sludge in landfills, by applying it to land, by
incine'*c-.J.o:i, anLi by oce-6n dumping. This is an innovative effort
to examine one pollution source -- sludge, which contains a variety
of pollutants — and assess its risk across media. In a closely
related program, EPA muFt consider the relative risks of ocean
dumping compared with land-based alternatives in its decisions
about issuing any kind of ocean dumping permits (Marine Protection,
Research and Sanctuaries Act). This requirement was reinforced by
a U.S. District Court decisior. (City of New York v. F^.".). FT.* is
responding to that court decision and is providing guidance on
comparative risk assessment as part of its revised ocean dumping
regulations.
-------
-39-
EPA establishes drinking water standards under the Safe
Drinking Water Act (SDWA). For each pollutant, a recommended rnax-
irnuir, contaminant level (RMCL) is based solely on health effects and
is set at a margin of safety to ensure that no known or anticipated
adverse health effects will occur. Each RMCL is a goal, not
standard. For conventional toxicants, each RMCL typically relies
on the same approach as described for FDA for noncarcinogens,
setting an acceptable daily intake. For probable human carcinogens
that are treated as nonthreshold toxicants (in the absence of data
to the contrary), a RMCL of zero is used. (Exceptions have been
made for several substances where the RMCL is larger than zerc
because the scientific ev'dence is inadequate to treat them ~s
carcinogenic by ingestion c'- because of other considerations.} For
-------
-4U-
enf orceo - j ^--ar^uses, a maximum coTt'aninant level (MCL) is set a?
close to thti R'iCL "as is feasible," considering technological
••-v--cit.y anc costs. for ex^njle, in ly?9 EPA set a MCL tor
t:rihalo~ethanes of 0,1 n^' /liter for water systems serving more than
lU,'1'1!) people. Under the worsi: case extrapolations this was esti-
mated to avoid u4< t'i '^20 cancers per year. Reducing the MCL to
O.Uit rag/liter tor ttiose systems would have avoided another 11U
cancers per year, however, this was determined to be too costly
and potentially h^/.^rdoas because it could compromise the disinfec-
tion of drinking water, which has well-known benetits of preventing
waterbur,.! i'iccctious diseases.
EP.-i .-> i-.acted a full risk analysis for the proposed htLs to"
nine volatile j'-^-v.lc compounds (VOCs), The contaminant levels
exanint-. t! -i-e closest, to the proposed HCLs are estimated to
avoid 1H-V cases of canctr per year; most of these are attrib.-'-
able to vinyl cU'im-ide, An additional analysis calculated aajubte.-
acceptahl ••* daily intakes (ALilS) for noncancer endpoints for all
nine Vf" = , V.-.--U oa a no-observed-adverMi-effect level divide.. 'iy a
safety tacL-'*- i-.-'^-re tote." exposure is assumed to be from drinking
wateT. Ve--y f•> >' public water systems exceed the adjusted ADI tor
any of the VOC; , so omitting the acute and chronic toxic effects
for the nonc,j.r..i io^ens does not bias the risk estimates (tor L^L-/'
cases ojr all effects) by very much (U.S. EPA, May 1 y85b). fciv <•>*
these V'JvJ- Vcr^ ri 0 .lated as probable carcinogens, one as having
equivocal evidence of carcinogenicity, and two as noncarcinogens.
So far, nu iii.^1 I'.HO is ion has been made on the last one. In
-------
-41-
addition to the nine VOCs described above, the Office of Drinking-
Water has proposed RHCLs for 43 other substances, including organic
and inorganic chemicals and biological contaminants. khCLs for
radionuclides also are being prepared.
Despite the mandate to protect human health and the environ-
ment, most of EPA's risK. assessments have been applied to human
health rather than to ecosystem and other welfare effects. There
are some exceptions (although some of the early ones may have
relied more on a finding of what concentration would kill half of
the test animals after relatively short-term exposures). Under
FIFRA, for example, endrir: was restricted on the basis of its risk-
to nontarget species (49 Federal Register 42792), sodium fluoro-
acetate was restricted on the basis of risk to endangered species
(41 Federal Register 52792), and toxaphene was restricted on the
basis of its risk to aquatic organisms. Risk of bird poisoning is
being considered in the evaluation of diazinon.
EPA is estimating the damages to material goods from SO^ and
its derivatives (e.g., when SU^ combines with fine particles and
forms acid precipitation) using a risk assessment that is less
rigorously structured than one would be for human health effects.
The results may be used in setting a secondary standard for S02«
(Primary standards are to protect human health; secondary standards
are to protto*- welfare and the enviroTiie-'t.) For the most rece-'-
review of the ozone ambient standard, EPA examined available data
with respect to its effects on crops. A more sophisticated anal-
ysis of crop damages is expected for EPA's upcoming review of the
-------
-42-
ozone ambient standard, for which it will conduct a risk assessment
of forest damages fron ozone.
As part of its review of existing chemicals, the Office of
Toxic Substance? has made substantial progress in assessing the
risks to aninal populations (with sont consideration of plants)
from chloroparaf f in*. . Ci.loroparaff i;.:. are used in about 200 com-
mercial products sucn as additives in lubricating oils, metal-
cutting oils, flame retardants, plastics softeners, sealants, and
printin- in*. EPV s Office of Research and Development is
developing guidelines on how to quantify ecosystem risks; interiu,
guidelines are expected to be available within the next few years.
Fooa and Drug Administration (FDA)
FDn, conducts research and develops standards on the composi-
tion, quality, and safety of human and animal drugs, human
vaccines, medical device^, food, food aJJLtives and colors, a;...
cosmetics. Its regulatory n.ithority was broadened in 1958 anu l^bU
with the passage of the rood and Color Additive (Delaney)
Amendments, which require manufacturer* to prove the safety ot food
and color additives and FDrt. to prohibit any additive found to
induce cancer in huraans or animals; in 1962 with enactment of the
Drug Effectiveness Amendments requiring that new drugs be proven
safe and effective prior t^ ir.arketing, and in 1976 with passage of
the Comprehensive Medical Device Amendment.
Food additives, color additives, and residues of animal drugs
in foods are evaluated by FDA primarily as a result of petitions
-------
-43-
filed by manufacturers, accompanied by quantitative toxicological
testing results. Normally, the company wanting to market the new
product will arran^.s. to hsve an independent lab conduct the risk
assessment. FDA evalxiatevS the petition for adequacy, and makes a
preliminary assessment of the toxicological data to determine
whether all of the potential health effects have been studied. An
acceptable daily inta':t for noncarcinogens is established, ana this
value is compared wit'i the estimated daily human exposure based on
the manufacturer's proposed use and predicted human consumption of
the foods in which the additive is to be used. For example, over
150 color additives ha\-o been permanently listed with levels estab-
lished for use in drugb and cosmetics, while over 50 have been
banned. These chemicals cro?? many chemical classes and their
changing uses often require new analyses.
At present, careinr>fens are treated separately. Some of the
substances e/aluateo l". c^rcinogenicity in the past by tht r,,..
Cancer Assessment Committee include:
Acrylonitri Lt<
Lead aretate
Vinyl chloriu.
Uioxane
p-Toluid ine
hydrazine
1,2-Dichloroetiiane
Diethyihexylphthaiate
Diethylhexyladipate
Furazolidone
Cinnamyl anthranilate
Triraethylphosphate
In 1982, FDA annou;:.i.d its "constituents policy." The policy
state= that if a carci:^^. .i :•-;• :••:'t;* ic found in a substance,
but the additive as a whole is not carcinogenic, then the entire
additive is not necessarily prohibited from use, provided that the
risk posed as estimated by risk asscss;..er.t procedures is found to
-------
-4*4-
be consistent w i • ' i'.-*- color additives in food, cosmetics, and drugs
that h•-..'< }> on on a temporary or "provisional" list.
r-L . also has tentatively invoked a de minitnis interpretation
of T}\ \aw's anticancer provisions in connection with the use oi
ij.etnylene calj. L . : I it decatteination ot cottee. In Deceir.bt^
IVhbj "tM-:i2i on risk assessnent, FL)n proposed to ban the use of
reethylene ci1 >-i
-------
.45-
Unavoidable food contamioM-il s tend to be addressed by .:»>rtin£
action levels based on risk, assessment, technical feasibility, and
econonic analyses. An icportn it. ^art of this process tor c'.ieMi-
cals, such as aflatoxin in cottonseed oil, is thp risk assessment
tor the chemical. Other residues, such as pesticide residues in
food, also contain action levels, based on polici^ influenced by
quantitative rl.-;1 •. -i_.-
*> and devices are evaluated on a more qualitative basis,
using ris'O'H'-i«'f j L analyses, These pr-.^Jares rely heavily on
clinical judgment and epider.ioloyical information. Since manufac-
turers have to pro-.'.- s.ifi-ty and efficacy, the an>unt of information
available to estimate risks for new drugs and devices usually tar
that available f->r contaminants.
Food Safety and Inspection Service
The U.b. Lepartment ot A^riCulture's tood Safety anc
Itispection Service (FSIS) inspects all meat and poultry ;jrouuets in
plants that ship in inte-state and foreign comm^r.-- . AC part of
i'ls activities to ensure th«i meat and poultry prod...t- ' • not con-
tain harmful chemical residues from either environi-it-Mit^l contamina-
tion or animal drus use, FSIS conducts formal risk assessments to
determine which chemicals to test for in its residue mo n torin^, anc
surveillance program. RISK assessment also is used to determine
acceptable limits for hazardous chemir-"1
-------
-46-
I •- rablishiny procedures for inspection of animals before
and after slaughter, tSIS uses a risk assessment approach to deter-
mine v;hich disease^ >r conditions to inspect for. The 1965
Nsi-'.>.Ml Acaden.y of Science-' report, Meat and Poultry Inspections:
the Scientific basis ot the Nation's Program, recommended a quanti-
lativ- risk assessment apt"* "-.oh be taken to support any changes in
inspection procedures. This would mean collecting data about why
i'^i-oction led to rejecting a carcass or product, so that the type
and rate of rejection i , different inspection procedares can be
rel -.'.-.I to public health effects. Through a NAS contract, FSIb' is
developing a quant 'L'-ettive model to evaluate different methods ot
t.---.1t-ry ^laughter inspection.
FSIS depends primarily on risk assessments performed by the
Food aiu Drug Aarcini*. r-nirion for making decision about the safety
ot" subsi-rices used in meat or poultry products. As for other
federal agencies, the Kblb proposals i- /uhlished in the federal
Register and its rules and regulations are published in the v'-*de of
Federal Regulations.
Occupational Safety and Health Administration (USHA)
USHA's primary ro--/^nfiibility under the 1971 Occupational
Safety and health Act is to protect the safety and health of
workers. To carry out this responsibilty OSHA regulates workplaces
anu in n^^onse to health hazards, OShA sets limits on the amount
of pollutant*. ->r dangerous substances to which employees may be
exposed.
-------
-47-
OSHA considers regulating organic chemicals and compounds fol-
lowing analysis based on discovery data or upon referral by other
government agencies including the Ivational Institute of
Occupational Safety and Health (NIOSH) and the Environmental
Protection Agency. In the past, NIOSH was responsible for pro-
viding OSHA with risk analyses necessary to support a proposed
occupational standard. This arrangement has changed in recent
years, largely as a result of the Supreme Court's decision on the
benzene standard.
In 1980, the Supreme Court decided that OSHA's benzene stand-
ard was invalid because the rule was not accompanied by sufficien*
analysis to document the fact that there was significant risk to
workers' health. The Court held that OSHA's policy -- to reduce
exposure to tin; lowest feasible level when there was qualitative
evidence of carcinogenic!ty -- was not sufficient alone to justify
regulation. In addition, tht Court required that a determination
be made tha:; a place of enjoyment was unsafe, and that significant
risks could be reduced or avoided by a change in practicp. The
Court ad-Jed, however, that the significant risk determination
should not be viewed as "a mathematical straight jacket," and that
OShA was nut required to support its findings that a significant
risk exists with complete scientific certainty. The 1980 Court
decision also directed OSHA to review all studies included in the
risk assessment supporting a proposed standard.
In response to the Court decision, OSHA's cancer policy was
modified in 1981 so that the significance of risk must be
-------
-48-
considered before setting a carcinogen standard. All recent OSHA
carcinogen standards have used risk assessments in their develop-
ment, (TaDle 7 shows the status of recent OSHA health standards
activity.) Because of its importance, OSHA has developed its own
in-house scientific capability to do risk assessments based on
animal and epidemiology studies. Risk assessments done by other
government agencies are routinely reviewed and if appropriate,
incorporated in an OSHA rule proposal. Outside experts are brought
in to revie.: risk assessment models,
Typically, the model will project worker deaths and illnesses
expected t i occur fror exposure to a dangerous substance. Becau0-'
the empirical conditions of animal and epidemiology studies differ
froc workplace exposure conditions, statistical formulas are relied
upon to e*.:. -ipclate results expected to occur in an industrial
setting. 0 !:•*• u';>- risk assessment model for a given substance has
been developer it is applied to the estimated distribution oi
workers exposed tu the substance at various concentrations and tirac
periods. While observed dose-response rates vary among substances,
the higher tht. exposure in terms of length of time and concentra-
tion of the substance, the greater the expected frequency of
resulting death or il
The result of the application of risk assessment models in
OSHA is r projection of benefits (de?th- ^nd illnesses avoidea;
fro... regulating and reducing worker exposures to dangerous sub-
stances. For example, in December 1985 OSHA proposed to limit
v exposures to an 8-hour time-weighted average of 1 part per
-------
Table 7
Sur.ir.ary of Ob tin btandards Development Projects
Standards
Completed
Vinyl chloride^3
Coke-oven enissionsb
Lead3
Cotton dust3
1 ,2 Dibroiao-3-
chloropropane^
Acrylonitrileb
Ethylene oxideq
Health conservation"
(noise)
Arsenic3
Hazard communication3
Standards Proposed
K o t. Cgmp 1 e t e d
L.ci/lene dibromidea
At Jfci tO£ a
benzene"
Standards
Being Uevelopeji
Methylene chloride
^t,4' -Methylene-
dianiline
Laboratories
Respiratory
protection
1,3-Butadiene
aThe developraenL of t'u <- . -ndard inclu:Jt"' « risx assessruert.
standard discusstt. ::J.sR, b-it a tur .al study was not done
-------
-50-
million (ppm) and to require certain other regulatory duties such
as measuring exposures and conducting medical surveillance. OSHA
developed a quantitative risk assessment based on epidemiological
studies, which was verified by outside experts. The assessment of
benefits was based on OSHA's risk assessment and predicted that
18.3 deaths per year would be avoided. OSHA estimated that the
cost per desth avoided would be $1.6 million for the 1.0 ppm pro-
posal. OSH^ has also proposed a revised standard for asbestos that
is expected to prevent about 75 cancer deaths and 30 cases of
asbestosis per year.
One example of OSHA action in response to the referral of a
new hazard Dy another agency is for 4,4'-methylenedianiline (MDA).
Based on available animal data, EPA has concluded that MDA is car-
cinogenic and pre^-nts an unreasonable risk to the health of
exposed workers. OSHA's risk assessment indicates that current
worxpiace exposures in the manufacturing and sec.:r.dary processing
sectors are expected to cause about 160 cancers over a 45-year
working lif «t ir>: . Because of this information, OSHA is drafting a
new regulatio', to protect workers from MDA.
Risks Occurring in Natural Resource Management
The National Environmental Policy Act (NEPA)
Under NEPA, federal land manage""0"1 agencies must prepare
environmental impact statements (EISs) for major federal natural
resource management activities that may have significant effects on
env:ronment. The Council or, Environmental Quality (CEQ)
-------
-51-
oversees the implementation of this act and has issued regulations
governing the environmental procedures of federal agencies.
Several agencies involved in managing natural resources have
conducted ristc assessments as part of ELSs. tor example, the
Bureau of Land Management (BLTi) , within the U.S. Department of the
Interior, manages about SOU million acres of land -- approximately
15. percent of all the lands in the United States. BLM is requirec
by law to authorize certain land uses and prepares EISs when
developing land use plans for grazing allotments, coal leasing,
forest cuttings, and activities involving minerals, oil, and gas.
BLM evaluates risks in its coal leasing program to decide what
sites are unsuitable for development, given criteria published in
its Programmic EIS Document in Coal Leasing. The kinds of risks
cons idered are dauages to surface water supplies, aquifers, and
endangered resources. Th - effects are quantified in many cases.
The U.S. Department of Agriculture (19b5) has assusbc^ the
risks to human health and to nontarget plants and animals from the
application of various pesticides to forests. Its 1>.S. Forest
Service manages 191 million acres of national forests and grass-
lands and regulates the use of forest resources. Management acid-
ities include outdoor recreation, timber harvest and reforestation,
protecting air and water quality, road construction, minerals
activities, wildlife ana fish habitat improvement, and livestock
grazing. In its planning process, the Forest Service follows both
NEPA procedures and those under the National Forest Management Act
of 1976. Risk assessments -nay be conducted at the regional,
-------
-52-
forest, or ranger district level for use in forest resource manage-
ment, Nat ional forest plans and other broad program direction pro-
vide general guidance, but most of the work is done in the foresl
and ranker districts where these programs are implemented. The
Forest Service is using risk assessment to evaluate such things as
the effects of pesticides on human health. The risk analysis
methodologies assess different strategies for managing natural
resources and their effects. For example, the Forest Service uses
a forr. of risk assessment to deternine how close to build a road
near a strear. or what type of forest cutting to perform, given the
expected draina^< and erosion impacts.
During the past two years, CEQ has paid particular attention
to one of the regulations (40 CFR 1502.22) which, ar.ong other
iuix , required federal agencies to include a "worst-c^.-u anal-
ysis" in an environnental impact statement if the information
abouL si^niL leant, adverse impacts is incomplete or not available.
A worst-cast analysis was supposed to identify the most severe
possible ettects of a", action o. humans and the natural environ-
ment, such as the possibility of cancer to those living anu working
in the region. CEQ became concerned that the requirement to pre-
pare a worst-case analysis, in certain circumstances, could require
federal agencies to go beyonJ the "rule of reason" in their anal-
ysis of potentially severe impacts.
For example, the Bureau of Land Management was ordered to pre-
pare a worst-case analysis assuming a causal effect between the use
of certain herbicides in federal forest lands and the development
-------
-53-
of cancer in hin^.n beings (in Sa ve our Ecosystems v. Clark, 747 F.
2d 1240 [9th Cir. 1984]). BLli contended that such an analysis
would ^ j» ire guesswork, becaa-; . . credible scientitic aat., .^up-
ported the contention that cancer could occur at any dose. Never-
theless, the ruling forced BUi i:.> stop its plan to encourage; timber
growth by treating 6,400 acres with herbicides.
In a U.S. Civ.irt of Appeals for the Ninth Circuit decision on
the same issue, the Interior Department was ordered to include an
analy«i-. of tlit? herbiciae 2,4-LJ in t'u- environmental assessment
prepared before s^caying federal timberlands (Southern Oregon
Citizens Against Toxic Sprays, Inc. v. Clark). The Interior
Department estimated that it would take nt least five years and I?25
million to perform the anal/^i.-.
After an intensive >-?view of this issue, including public
involvement, CbQ aiie i :• ' r'le regulation, etfective hay 27, 19bb.
Thn final amendment re;'i>-^s all federal agencies to disc". ~ce the
fact of incomplete o" ua-ivailable information w'jt;n evaluating
rnasonably foreseeable si.0iificant adverse impacts on the human
environment in an fclS, H^I ' r. -. obtain that information it the costs
of doing so are not exorbitant. If the agency is unable to obtain
the information, it must (1) affirmatively disclose that such
information is unavailable, (2) explain the relevance of the
unavailable i nciv-iiation to the decision at hand, (3) summer! ?:e the
existing credible scientific evidence relevant to the agency's
evaluation of significant adverse impacts on th.f huma^. environment,
and (4) evaluate the impacts based on research methods generally
-------
accepted in the scientific community. Impacts that have a Icn:
probability of occurring but would ha^e catastrophic consequ^ •
should bu evalua1:^! it the analysis is supported by credible scia-i
tific evidence rather than pure conjecture, and if it is within i' ablishes procedures for
assessing damages to natural resources froT. a discharge oi oil o-~
a relo-im- .>• ... U.izardous substance ir.^-r i h-2 Conprehensive
Environmental Response, Compeasation and Liability Act of 19bU
(CEKCL.a or Superfii'ii!) , or under the Clean VJ-3tor Act. (The
President delegated the responsibility for preparing this rult to
the Department of the In re" Lor in Executive Order 1231f>.) The ;-v •>-
po-sed rule is for use by federal and state officials vjho wana^t' the
public's natural resource-^. Such assessments wi] 1 he used in court
actions and adninistrativt proceedings when seeking compensation
for injuries to naturn". cebources. before an a =.--,'• sment begins, a*-
assess^ent plan is to be prepared and published, followed by a 30-
day period for comments from the public and perties responsible f^-
the da'.ia^e.
The L r^poseu rule describe? a?tei"n?=tive methods for conducting
assessments on a case-specific basis, but does not provide s^eiric
procedures tor implementing them. The rule is flexible because of
th,: variations among resourr.e-s, ecosystems, and hazardous
-------
-55-
substances, as well a1? t^ enable the incorporatioa of evolving
methods. Each damage assessment involves three major steps: (1)
establishing that a 11 injury has occurred and that the injury
resulted from the discharge or release, (2) quantifying the effects
of the discharge or release on the services provided by the injured
resource, and (3) deter. l.ii-ij the damage.
The proposed rule defines "injury" as a measurable adverse
long- or short-tera. change in the chemical or physical quality or
viability of a natural resource, resulting either directly or
indirectly fro a exposure to a discharge of oil or release of a
hazardous substance. The number of shore birds killed from oil
contamination is an exa-i.>le. The natural resources include surface
water, ground water, air, geological resources, and biological
resources. Biological resources are defined as fish and wildlite,
and inclxido shellfish, ter re^* - i al and aquatic plants, and other
living organisn-5.
Injury deterrainat.io - in this proposed rule is based on a
demonstrable adverse biological response from the oil or hazardous
substance. Both laboratory and field measurements are required to
demonstrate injury. For example, a resource is considered injurea:
0 if the concentrations and duration of substances in surface
water or ground water that was potable before the release
art: in excess of federal drinking water standards;
0 if the air emission concentrations are higher than federal
air pollutant standards;
-------
-56-
if the substance causes soil erosion by wind or surface
waters ; anu
if a biological resource or its offspring has undergone at
least one of the following adverse changes in viability:
death, disease, behavioral abnormalities, cancer, genetic
mutations, physiological malfunctions, or deformations.
second step in damage assessment is the quantification of
the effects on the injured resources. This phase requires charac-
terixi-i^' the base level for the uninjured resource and comparing
this base level LO the level existing or anticipated after com-
pleting any response actions. Although a formal risk assessment
can be used in this step, it is likely that most claims will rely
on tables anu formulas published to assist in assessing these
natural resource damaes.
The j« • >i)osea rule provides thtt the change in the resource
be quanti' '"'J in terras of the change in the level of "services"
that the r
-------
-57-
Dams
The Bureau of Reclamation (BOK), the boil Conservation
Service (SCS), and the U.h. Corps of Engineers use risk assessment
to assist then in making decisions concerning the construction,
modification, and repair of dams. The analysis considers the prob-
ability of dair: failure and what the effects would be to human lic'-
and the environment.
For example, the SCS (with otiier agencies) uses a form of risK
assessment in establishing criteria tor the design of dams. Their
approach establishes one set of criteria if the consequences of a
dam failure ar.. sTiall (such as for a small farm pond) and corres-
pondingly more stringent criteria as the consequences of failure
become greater (such as for a water supply reservoir for a commun-
ity). If the consequences of a dara failure are expected to be very
severe (such as loss of huiran life) , thun every known effort is
made to design the dare to prevent the most extreme events that are
known to be probable. The SCS develops watershed protection and
flood prevention projects for 90'J,OOU farmers, and is involved in
about 100 projects where it doeb soiae type of risk analysis.
The biggest problem with the use of more quantitative risk
assessment in designing dams is that flood probabilities for
extremely rare events are not well established. In addition, no
method is currently available for estimating the probability of dan
failure from various other causes. As the capability improves for
estimating the frequencies of such rare events, a more quantitative
approach to risk at>sessi.ient may bt_ adopted.
-------
-58-
i^uantitative risk assessments are used -i >r«-> often who", evalu-
ating repairs for existing dams. BOR evaluates the dam's condition
to develop probabilities of failure from being overtopped by
flooding, from earthquakes, or from static loading of the
reservoir. To the extent possible, the effects from railure are
estiir.are:! in terms of potentially lost lives, destruction or fis^
and wildlife habitat, property damage, soil erosion, ai.« crop losf- .
Within the past three years, bUR has conducted such analyses; tor
about 15 dams.
BUN supports the use of a decision analysis fraraework tor
evaluating existing or proposed dam projects (U.S. UUl, 1986). Itir
major components are a hazard assessment and a risk-cost analysis
of alternatives to identify what level of protection is econom-
ically justifiable. Tie i-ish and Wildlife Coordination Kct
requires BUk to coordinate with the U.S. Fish and Wildlife Service
ana tat currespondin^ state agency. lo address these rebponsiDiii-
tiVc, HOK has developed a monetary risk-cost methodology for fish
and wildlife habic-it changes caused by catastrophic Hood events.
Risks Associated with Dangerous Installations
In SOT*, crises, facilities may hav^ sudden releases, as dis-
cussed in the accidental risk section, as well as continuous
releases. A few examples are described in this section.
Mines
The Bureau of Land Management (1983b) has analyzed the risks
a gold mine poses to sa-'T^.A water quality, climate, air quality,
-------
-59-
and public health and safety. More broadly, the Mine Safety and
Health Administration (MSHA) (within the Department of Labor) per-
form^ risk assessments on ho*7 be.st. to reduce and prevent Mine acci-
dents and occupational diseases. MSHA develops and promulgates
mandatory safety and health standards, including those for airborne
contaminants.
MSHA recently used risk assessment in drafting a revision ot
the radiation standard. It estimated the risk of lung cancer to
miners, us ii,^ international epidenio Logical studies, information
from the scientific literature collected by NlUSh, hPA radiation
standard^, and national guidelines on safe industry level set by
the International Commission on Radiation. MSHA especially
balances the risks and costs to small mines.
Energy Installations
Within the Department of Lnergy, the heaiin and Environmental
Risk Progra-.i analyzes the potential health and environmental
effects of installing, operating, and decomnissioning induct - i^s
using emerging technologies capable ot providing a significant
fraction of the nation's energy (barr, 1983). Analyses of the
nature, magnituut a ,u sources o^ uncertainty regarding potencial
health and ecologicd impacts are -• :Mished and revised on the
basis of comments from interested parties in industry, government
and academia. These analyses provide input lor planning research
to reduce uncertainties that are likely to impede tne use of cost-
effective control strategies. Analyses have been published for
-------
-60-
seven energy technologies (refuse— derived fuel, geo thermal ,
fluidized bed combustion, coal liquet ication, liquid metal fast
breeder reactor, photovoltaics , and oil shale), two classes of pol-
lutants (nitrogen oxides and airborne particles) common to several
energy technologies, and two risk analysis methodologies (treatnenr
of uncertainty and hydrocarbon carcinogenic! ty) (see reterences
A-K) .
As in other agencies assessing ecosystem risks, the Department
of Energy hay been hampered by the lack of mature methodologies in
this area anu by the site-specific nature of potential ecological
effects. However, it estimated ecosysterr risks for mule deer,
Indian rice-grass, and plant communities for a one million barrel-
per-day (BPD) oil shale industry in the United States (1WG Corp.,
1984). Additional potential ecosystem effects were identified but
'Ntiirated. For risks to people, Table 8 shows far more acci-
i>
dents than cases of disease; the analysis allows the separation of
fatal injuries fron1 those that lead to no work loss or some work
loss, korkt ~ ht.-.lth effects also are separated frov those tor the
public.
Viork Pls.ce Risks
beside- analyzing continuous work place risks, OSHA also eval-
uates industrial processes, facilities, and i>rr.•'*• ices that may pose
a risk to workers b> causing injuries, fatalities, and accidents
from fires, explosions and falls. OSHA has proposed four safety
standards that, nave included risk assessments.
-------
-61-
Table 8
Risk Estimates tor a Million BFD OilShaleFuel Cycle
health or Environrci.tsi bttect
WORKERS (Population at risk: 41,000 persons)
Risk Per Year
(Uncertainty Rang^;
Cases Leaths
Injury, with days lost
j'iry, without dayc li.-t
24UU
(1700-3700)
150J
(1200-220U)
13
-2
N/.
Cancer
SiL icos is
Pneumoconiosis
Chronic Bronchitis
Airway Obstruction
high Frequency H» a^inj Loss
26 4
(0-300) (0-49)
232 7b
(0-1070) (0-387;
100
(33-310;
41
(13-130)
10
(3-3b)
3
(0-b)
17
(9-98)
15
(4-51)
(1-17)
PUBLIC (Population * * "i«"j 616,000 (Region), 313.000.00L ^L.b.
Premature Death fro., -.1 Pollution NA b
Internal Cancers
0.0002 O.OJU2
(0-0.001) (0-O.OUUb;
ECOSYSTEM (Area at risK: ^ ,Q(}(J square miles)
Mule Deer Decline (habitat Loss)
Indian Rice-grass Inj.'-j (Sulfur Dioxiae)
Plant Community Decline
(Solid Waste Disposal)
4% to 9%
(-50% to 200%)
0.87o
3 comraunities decrease
as much as 20%
Source; IWG Corp., 1984, p. ix,
-------
-b2-
A major focus is identifying the populations at ris1:, These
are the number and types of workers in those industries that would
be signit icantly affected by a safety stanaard. OSHA'.s standards
cover the hazards associated with concrete and masonry structures
in the construction industry (to reduce the risk that a structure
in progress would collapse), grain handling facilities (to reauot
the explosions ana fires in grain elevators, processing plants, ami
mills) , places of employment that use scaffolding or are in under-
ground In.-Ktions (such as tunnels, shafts, chambers, passageways
and covereJ excavations), and noisy work places.
OSKA's curren; nativities are concentrate^ o-i standard? for
walking and working surfaces, the handling and storing of tlararaable
an--' combc^ tible liquids, the communication of hazards, and electri-
cal safety work practices. Some of these categories are quite
broad. For example, OShA's standards for handling and storine c_~-
bus> Li-j-t ana nanunable liquids speciiically address fie to
indu-;t'-y "-notors and locations: petroleum refineries ana di^tribu-
t:>rs, chei: ical processing plants, gasoline service stations, bu1 ••
plants, marine terminals, and distilleries.
Transportation
The Department of Transportation is responsible for several
aspects of transportation safety. Its Office of Hazardous
Materials Regulation sets national safety standaras> iu. LC^J>-
porting all hazardous materials and hazardous wastes by air, water,
highway, and rail. The office designates substances as hazardous,
-------
-63-
and regulaf.es their transportation in interstate commerce. At
present some 2,400 specific materials are listed and labelled as
hazardous *r\^., transported. The oif^iee does not pertorm specitic
risk assessments for all of these materials, but regulates them
based on known hazardous properties (e.g., flaromability, corrosiv-
ity, and toxicity). The chemicals that cause the most injuries an
fatalities are gasoline (40'/o ot all fatalities, Department of
Energy's Battelle Research Lab Study), liquefied petroleum gas, an.i
sulfuric acid. The office sponsors risk assessments to be used in
the regulatory process and to identify areas that may need more
attention. For example, the office contracted for a risK assess-
ment of the transportation of liquefied natural gas out of a
terminal ia Boston, Massachusetts.
The Office ot hazardous 'iaterials Regulation follows guide-
lines for assessing safety risKs published by the Secretary of
Transportation. It nab aboaL 1 b ongoing rulemaking actions that.
are subject to benefit-cost/risk assessment ot varying complexity
depending on the scope and conte'p of the proposed rule. One cur-
rent regulatory analysis involves assessing hov: a package ot a
certain construction will withstand the rigors of normal transpor-
tation. The rpt-^latory action involves conversion from packaging
specifications to performance-related requirements. For example,
current requirements for carrying various types of hazardous
materials speei-C^ liiov a. drum must be manufactureu in accordance
with specified detailed design. The office proposes that drums
-------
-64-
pass various performance tests instead, such as drop tests and leak
tests.
In developing transportation regulations, tht office atte-upts
to estimate conditional probabilities that:
1. the vehicle will have an accident;
2. the accident will result in a spill;
3. the spill will be a certain sizej and
4. then, if for example, the substance is flammable, that it
will be ignited.
Historical statistics on accidents are used in tht estimation pro-
cedure.
V.'hen proposing new or revised transportation regulations, the
office assesses alternative strategies, usually in the form of di±-
ferent jodes of transportaLio . (truck, rail, tank, ca^). Tank
trucks nt.'t bee.i found to be the riskiest mode, especially in li^ht
oil safu ./ Improvements with rail tank cars over the last two years.
o., average for tht: past five years, there have been 14 fatalities
per year by all modes of transporting hazardous materials.
When chi: substances being transported are hazardous wastes,
then the Office of Hazardous Materials Regulation shares control
responsibilites with Lr*v. Ihe agencies promulgated closely coordi-
nated standards in 1980. The major change to earlier regulation.--
was the inclusion of intrastate shipments of hazardous wastes.
-------
-65-
Th e U.S. Coast Guani is a branch of the armed forces under the
jurisdiction of the Department of Transportation. It regulates
vessels, sets and enforces safety standards, prescribes license
requirements for merchant marine personnel, and has search and
rescue functions. The Coast Guard provides the federal on-scene
coordinator for responses to oil anr hazardous chemical releases
occurring in the coastal zon*., Great Lakes waters, and specified
inland ports and harbors. The on-scene coordinator is responsible
for coordinating federal response activities at the site of the
release. Risk assessment is used in tin.1 sense that the Coast Guard
uses quantified data from a hazardous materials incident to deter-
mine the appropriate response. The most appropriate action is
determined after assessing various £^rameters including location,
size, chemical type and character of the plume, and the environ-
mental impacts.
Other agencies i-i (:'>.• Department of Transportation administer
general safety regulations for various modes of transportation.
The National Highway Traffic Safety Administration's (NHTbA; pro-
grains ar<2 designed to increase motor vehicle safety and decrease
the threat of death or injury in traffic accidents. Agency
research determines whic'. parts of vehicles can be improved to
increase crash avoidance capabilities or to provide greater protec-
tion in the event of a eras'-!. NHTSA also investigates vehicle
ueiects and ca^ order manufacturers to repair flaws that affect t\c
safe performance of the vehicle.
-------
-6b-
NHTSA conducts regalntory evaluations for all new rules and
regulatory impact analyses tor those having a major economic impact
(usually ?1UU rr.illioa ur more). AF part of each K!A, the agency
performs a risk assessment, which
1, estimates hov nany people are killed or injured (probler
assess;ntnt) and
2. determines the effectiveness of countermeasures (i.e.,
reduction in injuries and fatalities).
This information is then compared to the counterrceasures' costs
such as higher prices, eFfects on tuel consur;t>tio"., and major
impacts on the regulated industry.
One recent risk assessment was conducted for the mandatory
inclusion of passive restraint syste > (i.e., automatic seat belts
or air bags) in nutomobiles. NHTSA officials estimated that pas-
sive restraints could prtvent as many as ^.UUC1 aeaths ana 15U,GOu
serious injuries a year and would pr:>7idt- annual savings of up to
$2.8 billion in insurance premium?. A benefit-cost analysis per-
formed by a consultant to the insure ice industry found th"
standards to be clearly cost-effective.
NhlSA also investigates reports of vehicle defects not covert^
by an agency standard, assesses the risK, a^d often forces the man-
ufacturer tu correct the defect a* a result of the assessment. tor
example, a major tire manufacturer's product was recalled du* to a
safety defect assessed by NHTSA. In performing its RIAs, NhTSA
follows in-housn policy, the guideline^ tor K*ecutive Order 12291
-------
-67-
and for f1. K.-^nlatory Flexibility Act, and the requirements in the
Motor Vehicle Information and Cost Saving Act of 197^ and in thf
National Iraftic ana \.i^ur Vehicle Safety Act of 19bb.
Tr. i-'oderal Aviaiio. ^Ouini stration (FAA) uses risk assessment
techniques to learn li •-. .<_, rtJuce the risk of catastrophic occur-
re:i • • -s to the extent possible- A recent risk assessment involveo
;«r itecting passengers * r v r^hin fires in commercial aircraft. The
FA-. studied safety proce r.;--«_•-,- and new equipment that will reduce
the risK o£ fires. o.i -ti.'.ly F/A.-V is developing a regulation o.;
prote.'.i;ive breathing ecuijv ^-:v for commercial aircraft. The FA«
••i-^tis methodology t!'.;jt I1- f^-i^atible with the general guidelines set
f -r-th in the Department of Transportation document, "Methods for
K:v-ioi.iic Assessment of Tra . -/ortation Industry Regulations." The
FAA1 s own inethodolo^., ~. • economic analysis, "Kconoraic Analysis of
Inv ••-! "-:>.Mit and Regulatory I?-^:'-ions -- A Guide," does not specify
risk assessment p^iK-.^iiarts per se but describe^ investment anal-
ysis procedures.
The Maritime .-v* ."i.stration subsidi/e:* /-"wate industry to
build ships and manage?- t~hfj sites for the national defense re-servt
fleet. BaseJ on its r^k assessment, the Administration chose a
maximum permissible exposure level for asbestos in its ships.
The Federal Rail^o^4 ^-irninistration (BRA) aspc* ;>s risks as
part of its «.„.. .ou.ic anox^^is of proposed major regulatory actions.
An example is FRA's r^^^.nL ruiemaking for trans^ >rting certain
hazardous materials in railroad tank cars. The accident record was
-------
-68-
analyzed for specific types »f tank cars carrying specific
materials. Continuation of the historical accident pattern was
considered to represent the risk of no regulatory acti > .. Pro,,ow
improvements in th*- safety features of tank cars were matched
against historical accidents to determine which adverse conse-
quences could have btc i prevented or reduced by individual safety
features or combinations o^" them. The differences bcjr*ven the ri
of no action and the risk remaining after variou> «• »ibinations of
saf-'-y feature5? forn,ed the b-r is for comparing the j •>•'. ••! ial bene
fits of the various options considered in the ec .uocic analysis.
Risks Associated with Alternative Strategies
Decisions about how to deal with hazardous wastes can be
viewed as bein;- in -t catf0 .>ry distinct fro'i t\«- previous four
catt;^.)ries. This is because hazardous waste regulations &** icrally
pay more attend I. . to a! f:ernat.ive ways oi: r i^rin^ these sub-
stances, rather than simply restricting the amount of air or ./ater
emissions from a particular source. EPA has primary responsibility
fnr the management and disposal of hazardous waste- -riJer the
Resource Conservatioi and Recovery Act (RLtv\" • •! ^^ recent araend-
T,ient5=. For hazardous waste regulation, the use of formal risk
assessment is somewhat new and controversial. Its use is made more
difficult since the emphasis in the legislati.c language is on pro-
tecting human health and the environment; EPA ha^ int.--1jreted this
e-nphasis to mean that a balancing of risks and costs is not permis-
sible.
-------
-69-
Land Disposal
A review of recent KlAs suggests the extent of risk assessment
used i; ' accompany regulatory analyses, even though a full compari-
son of benefits and costs was not legally feasible. The RlA for
the land disposal restrictions of solvent wastes (U.S. EPA, 19B6a)
considers cance^ vid other health risks, but provides estimates
only for average and maximum individual risk levels rather than
characterizing the risk by the number of cases that would occur
under each regulate-/ alternative.
In contrast, the RIA for restricting land disposal of dioxin
estimates that t'lere would be small chants in the number of cases
if the dioxin wastes were incinerated, rather than disposed of on
land under approved Waste Management Plans (U.S. KPA, 198bb). The
changes would be small mostly because dioxin is immobile in ta>'
environment, unless it i^ combined with mobilizing compounds suc'i
as solvents.
The RIA for the overall land disposal restrictions program
(U.S. EPA, 19ttbf) calculates individual risk for each cherrical anu
each environmental rcediun (air, water, land) as a measure of prob-
ability of harm weighted by a severity factor. Dose depends on
distance from the site, source of drinking water, and fish eating
habits. This information is combined with several population
scenarios to estimate the number of cases for that chemical anu
medium. Then cases are added over all the waste constituents and
all three media to determine the total number of cases. Estimates
of cases avoided by prohibiting each of three categories of land
-------
-70-
disposal >.rv shown in Table 9, by general waste classification
categories. It shows that prohibition of hazardous wastes in
treatment impoundments would be relatively ineftective in reducind
an already modest risk level. On the other hand, prohibition in
landfills, disposal impoundments, land farms, and waste piles would
be more effective (except for the K wastes) and would avoid more5
cases.
Risks also were estimated for the requirements imposed on
generators of small quantities 01 hazardous wastes (b.S. EPrt.,
IVb.1)^;.. On the basis of the assumptions used, there would be 2b5
fewer cases ot c-i'ioer since most of these wastes no longer would be
disposed of in ordinary municipal landfills. However, preliminary
estimates indicate an increase in noncancer cases of 2,369, which
nearly doubles the unregulated number. Most of these a^u expected
to be severe headaches fi-o^ methanol, which stop when exposure is
re-nove^. It should be noted tha-_ these estimate of change.^ in
health effects ttre tor a ^OU-year period, so thai the health bene-
fits would be very small on an anru:- "* V.-is.
Used Oil
Although the risk assessments for land disposal restrictions
and small qunntity generators yield relatively small estimates of
total cases, the RIA for managing used oil shows larger baseline
cancer estimates and correspondingly larger nu...^ers ot cases
avoided by the regulatory options considered (u.S EPA, 1985e) (see
Table 10). These calculations are fnr potential cancers rather
than expected cancers because conservative population exposure
-------
-71-
Table 9
ESTIMATED HEALTH BENEFITS
(AND 1 REDUCTION FROM CURRENT RISKS)
OF LAND DISPOSAL RESTRICTIONS a/
Currant Land Disposal Technology
Vaste Type
Landfills, Disposal
Impoundments, Land
Faros, and Vaste Piles
b/
Storage
Impoundnents
Tresto&zt
lapoundaerts
d/
Solvents
Other F
K
P aad U
D
X
(F001-FOOS)
84 (99.9)
902 (85.2)
282 (28.4)
27 (93.1)
(99.9)
(97.7)
Z81
28
Total
1,604 (64.8)
16
52
25
49
156
(99.8)
(59.2)
(99.9)
(99.9)
(66.8)
(64.9)
298 (72.3)
8 (76.I)
52 (29.7)
<1 (1.2)
<1 (0.6)
63 (38.7)
6 (58.8)
129 (31.2)
a/ Weighted cases over stvccty years.
b/ Thtri ar« a total of 16 disposed (landfill*, disposal ispoosdausts, land
farm* and wast* pile*) wast** by EPa coda (aad 3 X vastas), that potentially
pose increased risks in the treatment technologies initially selected. IE the
baseline, these wastes contribute 569 cases (over seventy years) to baselin*
risks, alaost 191.
c/ There are 3 wastes, by EPA coda, stored in surface iopoundpsnts that
potentially pose increased risks. They contribute a negligible portion of
baseline risks.
d/ There are 4 wastes, by EPA code, treated in surface iopoundaents that
potentially pose increased risks. They contribute a negligible portion of
baseline risks.
SOURCE: U.S. Environmental Protection Agency, 1985f.
-------
-72-
TabJb 10
WTOiTIAL CUMCOt RISK HIIX USED OIL IN MSEUNE
AM) «£QJUTOHY ALTERATIVES
(mabcr of health affoets1)
Regulatory AltarnatUa
Alt*cn»ti«e»
Baaalin*
Full SubUtla C
(Alternative 1—
100 otm Pb)
(Alternative J— (Alternative
100 PC* Pb) 30 otm »b]
10
Had oiling
Urtn burning
A»ph»lt pl*its
Spec* h»tUr»3
Incin«rttion
Lined l*ndrills
Unlintd l*ndfili«
Starag* l»»k»
Tottls
IA,»S2
4,01*
3,790
*,9I2
of »w«f 0*v t 70-r*« lifctla*.
•!! AMping in th« &••!!»• la by da.it.jreun«ir oil chmgir*, rfw tri 9gt»l* Ifw «cop« of tM aropaset
r»9ulttiane, risk r««ulL« r«fUct only inert wad Avpin; csu«Mi by rcaporaa to rafjifttiar,. fiUka fr» unte?
UUd da-U-yousMlf oil c*wn9»r« ox* «»U»»t*d «• 3,940 hMltn •ffoet* *»d «t» not orract^ by tn*
hvktam tlao =»u*- n Mtiaatcd 1,700 c«a«* of laad peiMnin^ (o*«r 70 y*tn) in trw b>Mlin*.
arc aliainatvd by tn» regulation.
SOURCE: U.S. EPA, 1985e
-------
-73-
assumptions in.<3.y a-u^z. upward bias in the estimates. Many of tho
cancers In the used oil Klri. result froir ingestion of arsenic (in
drinking wat^r). This substance is most closely linked with a fur .
of skin cancer that usually is treatable. In contrast, few data
are available about where cancer shows up in the body when other
contaminants are ingc^cec. Although cure rates tor other sites
will vary, the usual assuniptiori for other contaminants is fieti tuu
cancers they cause will be fatal. There are substances such as
lead in used oil their can cause noncancer health effect-, but these
are not included in t'l^ analysis because of the difficulty of
aggregating diverse health effects. This would tend to underesti-
mate the total health impact of regulatory action, ami these non-
cancer effects may nc L bt cht sane across regulatory alternatives..
The net importance o*- f.V considerations in a regulatory context
is not clear, but the estimates in the waste oil risk assessment
may have to be inter.jr^:. , oitferently from those in other risk
assessments.
Incineration
Incineration is r. t •;:'^-itially important alternative to land
disposal for managing hazydous wastes. EPA compared the risKs 01
incinerating PCB wastes and EDC wastes in land-based incinerators
with those for incinerators located at sea (U.S. EPA, 19b5z), The
analysis considered / T<-i':lt hur.ar. health effects and envir^i. aenta1
effects due to releases fro>: che incinerator, due to fugitive
releases from transfer ar.. storage equipment located at the land-
based incinerator or at tae ship docks, and due to the potential
-------
for spills v?-iilt- the ship is g-un^ from the dock to the ooe-:n burn
area.
Details ot the analysis include estimates that the total inci-
dence ot cancer at the two lano-oased sites in the study ranges
from O.Zy to u.43 cases over 7u years ot continuous enissions (p.
D-9). Although an explicit estimate is not given for cancers tro»r
the ocean-based incinerator, the lov: individual risk levels irply
substantially less than one case per million people exposed over a
70-year perioc.. Trie ho*]!':i risKs troiu potential marine spills wert
examined by comparing estimates of ambient pollutant levels with
tht ir-aximun concent---t^ ion limits recommended for occupational
expos-ires. Since both short-term anc long-terra recommended limits
could be exceeded if a spill occurred, EPA concluded that nearby
spills could pubt health risks -- but no estimates of cases are
^-ovided (t . H-7).
Thic malysis also c.--.eluded that incinerator emissions vjo-^^
have negligible ettects on marine ecosystems. Less information was
available tor their eftects on terrestrial ecosystems, but the
expected damages ace reported to be minor (p. 1-15). The analysis
of environmental effects from an ocean spill show larger impacts,
although they are reported mostly in qualitative terms (Table 11).
Other KisK Analyses
Aside from its recent use in KlAs, the tools of risk assess-
ment are coining to be a part of other decision^ about hazardous
waste management. Examples include: comparative risk analysis as
-------
-75-
Table 11
SUMMARY OF MARINE ECOSYSTEM EFFECTS FROM SPILLS OF HALF A TANK
Release Location
PCB Waste
Effect Bioconcen*
on tration
Biomase Levels
EDC Haste
Effect Bioconct;.-
on tration
Biomass Levels
Mobile Bay
Floating Case
Sinking Case
Small overallr
severe reduc-
tion for
benthos
Oncertain
3 to 5 orders
of magnitude
Uncertain
Not
Considered
Minor
Not
Consioered
Minor
Continental Shelf
Floating Case
Sinking Case
Uncertain
Small overall,
substantial
for benthos
Uncertain
2-3 orders
of magni-
tude
Not
Considered
Minor
Not
Considered]
Minor
Burn lone
Floating Case
Sinking Case
Uncertain
Minor overallr
substantial
for benthos
Uncertain
1-2 orders
of magni-
tude for
benthos and
demersal fish
Mot
Considered
Minor
Hot
Considered!
Minor
Source: U.S. EPA, lyBSc, Exhibit 1-8
-------
-76-
part of choosing effective treatment technologies when land dis-
posal is rest-.rictea, location guidance, decisions about site-
specific exe-iption ot hazardous wastes from regulatory controls,
combustion of hazardous wastes, and variances for facilities that
exceed background concentration levels in ground water. A final
example is the proposed risk model to set up constituent-specific
concentration levels to serve as a sort-en for land disposal of
hazardous wastes. If a particular waste exceeds those screening
levels, then it must be trc-ite^1 ,jrior to land disposal.
Supertunc
Under Supertund, EPA mas I designate substances as hazartiou.--
and set minima', quantities for reporting releases when they wouio
"present substantial danger tn the |>uM io health or welfare or the
environment." bPA is also responsible for updating the National
Contin-oncy Pl-i'i for remedial action in response to skills of oil
or hfir=.rd' . instances. Although neither ot these regulations has
been designsi HS major, R!AS have been prepared. RISKS were not
estimated in either klA,
On the other hand, EPA and other agencies (such as the Center
for Environmental hedith, in the Department of Health and Human
Services) examine risk to sot..- extent to learn about the relative
dangers of sites that were usec* for hazardous waste disposal but
are no longer subject to the Resource Conservation and Keeovtr/
Act. The most dangerous of these sites are put on the National
Priorities List (NPL) to be cleaned up under Superfuna. \>ith the
Ha/sard Ranking System, EPA's Superfund program uses information
-------
-77-
fror: si-tp liberations to compare the potential risks posed by dif-
ferent hazard jus waste sites. Three factors are used to score
site.;: (1) the possibility that hazardous substances will tnigraLt
offsite and reach populated areas; (2) the possibility that people
will come into direct contact with hazardous substances; ana (3)
the possibility of fire or explosion caused by hazardous sub-
stances. The firs>t factor is used to put sites on the NPL, and t!v
other factors are used to identify sites that need remedial
actions.
Sites are selected tor remedial actions based partly on risk
and partly on geographic distribution. Risk analysis is used to
develop target performance goals that will define an effective
clean 1,1 at- uncontrolled hazardous waste sites. However, thi
-------
-76-
KISK ASSESSMENT IN L'SK.: CONCLUDING COMMENTS
The previous sections have describe.! a multitude of uses of
risk as-st;.-s...ent by public author! t io.1- in tht United States,
Several caveats ne^is to be kept in mind, however.
Sor..t of tlie risk analyses described above do not include all
four steps i.i the ri.sk. assessment process (,i* definea on page 3).
This may be. because hazard indentitication is all that the statute
requires for a decision. Sometimes, exposures are estimated and
corrb'.'if-.i wita the "yet/' resulting f^or. the hazard identification
step. Other analy.si" have information about the risk, to the most
expose) individual but no exposure esrircates. An attempt has been
made to identify those risk studies that do not include all phases
o'-~ * full risk characterization, but that information has not
'i;. *'a.yj> been available.
In addition to differing statutory requirements, the extent ui
risk analysis may be influenced by availablr rec > i-c-i^t, ^nrt exper-
tise. Even before Lxecutwe Order 1/291 added explicit require-
ments for benetit-cost analyses, the cost to the a^iteies for pre-
paring regulatory packages rangeo frorr about S120,UUO to Si.b
million (U.S. General Accounting Ottice, 1982). Of course, only
part of this would be for risk assessment. In the regulatory anal-
yses performed in 1979-80, most of thr agencies (.liacusbea in this
paper spent at least 50 percent of thp?- analytical bxTi^e*- for work
by outside contractors; in some instances the amount approached 100
percent (U.S. General Accenting Office, 1982). Again, this does
not separate the portion of risk assessments that were performed by
-------
-79-
contractors, but it suggest*-; that outside consultants may havTi ;^ cr
an important su^, le*.;ent to agency expertise in this field.
Another limitation oT the paper is that it does not examine
the 4tiality of the risK assessments. Nearly all of the ones use_
to support regulatory decisions were made available tor expert or
publi.' T-eviev, but there is no indication in the paper whether
risks were re-t^st-v'ed on the basis of such feedback. This leaver
open the possibility that n detective risk assessment coulc have
been use.', for a couLro^ uecision.
There are a nun-be- i' •'esearch efforts aimed nt addressing th
deficiencies in quantitative assessment, for instance tn limit
•mcertainty resulting trot, the use of assumptions. Recent govern-
ment efforts include: t'i^ research recommendations in the
Department of Health and hu.aan Services document (DhHS, 1985); the
development of a proj-rfr at '••'".',, spearheaded by the rational Cente
for Toxieulocleal Researcn, t; adaress the critical assumptions
used in risk assessment a-;' t . examine modulators of: toxicity;
extensive epiaemiological studies relevant to cancer risK at the
Nations! Cancer Institute/Isatinn^.l Institutes of Health (MH) ; tht
National Science Foundatioo recommendations (NSF, 1y84); efforts a
t1! National Institute of Environmental Health Sciences/hlH and
evaluation of the eKtra^oiatio1. of chronic bioassays by tne
I\atiorial lexicology Program. I- c-Jr: ition, there are a number ol
efforts by private foundations and institutes. These are all part
of a growing effort to improve risk assessment and make it truly
quantitative.
-------
-8U-
As mentioned in the introduction, coverage is intended to be
comprehensive, but it. may not be o-iplete. The most important next
step, however, is to e\r!in.ine the role that risk assessment played,
relative to other factors, in tht- public authorities' decisions.
That will have to be the subject ot a separate paper.
-------
-81-
Keferences
10.
11.
Barr, Nathaniel F. , "The Kole of Environmental Risk Analysis
in the Cost-htf ective Development and Operation of Emerging
Energy Technologies," Ann. Kuc. Energy, 10:3/4, 1983,
187-194.
Boykin, Raymond F, , Raymond A. Freeman, and Reuven R. Levary,
"Risk A.seasment in a Chemical Storage Facility," Manager, crt
Science, 30:4, April 1984, 512-517.
Clement Associates, Inc., "Review and Analysis of Hazard
Exposure, and Risk Assessment as Practiced by EPa and Other
Federal Regulatory Agencies," Washington, D.C., mimeo, 1961.
Lhrlic1'!, Alan h. , "Initiatives in Risk Assessment Research,"
presented ci u the Conference on Environmental and health Kitr.
Assessment, 'Washington, D.C., October 28, 1985.
Fieia, Robert I., and Lawrence E. licCray, "Federal Risk
Assessments l:yr Potential Carcinogens: An Empirical View,"
in Working Papers prepared for the Committee on the
Institutional Means for Assessment of Risks to Public health,
National Research Council, Washington , D,C.: National
Academy Press, April 1983.
Fisher, Ann, "The Scientific Bases for Relating health Ltfects
to Exposure Levels," Environmental Impact Assessment Review,
3:1, March 1982, 27-^2.
Huguenin, hike, Kike Pich, and Jin El kind, "Review or
Probabilistic Kisk Assessment for Cheinical Manufacturing
Facilities," Draft Memorandum to Fred lalcott, U.S.
Environmental Protection Agency, July 25, 1985.
Industrial Economics, Inc., Management Technology and Data
Syste&s Inc., and PEl Associates, Inc., Acute hazardous
Events Data Base, Interim Final Report, Washington, D.C.,
EPA No. 500-5-85-029, December 1985.
Interagency Regulatory Liaison Group, "Scientific Bases for
Identification of Potential Carcinogens and Estimation of
Risks," Journal of the National Cancer Institute, 63:1, July
1979.
Lehman, *..J., c* si., Ajjpjraisal of the Safety of Chemicals ?'*
BoodSj Drugs , and Cosmetics , Association of Foo^: and Drug
Officials of the United States, 1959.
National Academy of Sciences, National Research Council, Risk
Assessment in the Federal Government: Managing the Procesr .
Washington, D.C. : National Academy Press, 1983.
-------
18.
19.
2U.
21,
22,
12. National Academy of Sciences, heat and Poultry Inspections :
Tne Scientific basis ot the Nat ion's Program, k'ashington,
D.C.: National Academy Press, 19bb.
13. National Science Foundation, Kisk Assessment and Risk
Assessment hethods: The State of the Art, V/ashington, D.C.,
December 19M*.
14. Risk Science Institute, Review of Current Research Activities
to Improve Risk Assessment and Identification ot Major baps,
V/ashington, D.C., November lyHb.
15. l!,S. Army Corps of Engineers, Kndicott Hydrocarbon Development
Project Environmental Impact Statement, 19843.
16. U.S. Army Corps of Engineers, Georgetown Refinery
Environmental Impact Statement, 1y«4b.
17. U.S. KurtirT.i ot Land Management, Riley Ridge Natural Gas
Development Project: Final Environmental Impact Statement,
19b3a.
U.S. Bureau of Land Management, McLaughlin Gold Mining
Proj ect: Environmental Impact Statement, 1983b.
U.S. Departim.----!- of Agriculture, Forest Service and Animal and
Plant health Inspection Service, Gypsy Moth Suppression ann
Eradication Projects: Final Environmental Impact
Statement, 1985.
L.S. Department of Lier^y, Assistant Secretary fu'
tuel Energy, Liquefied Gaseous Fuels, S^"!.1 Test Facility,
Was n i-),-/. ' , D.C., no date,
lj.y. D^jjartment of Health and Hunan Services, Risk Assessment
and Kisk Management of Toxic Substances, Washington, D.C.,
April 19db.
U.S. Department of the Interior, Bureau of K. .-lauation,
Guidelines to Decision Analysis, Denver, Colorado: ACER
Technical Memorandum ho. 7, iy6C.
23. U.S. Environmental Protection Agency, "Interim Procedures
and Guidt-1 i nes for Health Risk and Economic Impact
Assessments of Suspects1 C*ivi:;ogens , " ^1 Federal Register
24102, May 197b.
24. U.S. Environmental Protection Agency, "Proposed Guiflt. i ia._t
for Carcinogen Risk Assessment," "}'r.4nsed Guidelines for
Exposure Assessment," "Proposed Guidelines for luutagenicity
Risk Assessment," "Pr'>;- >sed Guidelines for the Health
Assessment of Suspect Development Toxicants," 49 Federal
Register 46293-40331, November 23, 1964.
-------
-83-
25.
2b.
27.
2b.
29.
3U.
31.
32.
33.
34.
35.
36.
U.S. Environmental Protection Agency, "Proposed Guidelines for
health RISK. Assessment of Chenical Mixtures , " bu Federal
Register 11 6y- 11 76, January 9, 198ba.
L'.S, Environmental Protection Agency, Economic Impact Analysis
of Proposed Regulations to Control Volatile Synthetic
Organic Chemicals (VOCs) in Drinking Water, Washington,
D.C., EP*. 570/9-85-004, May 1985b.
U.S. Environmental Protection Agency, Assessment of
Incineration As a Treatment Method for Liquid Organic
hazardous Wastes -- Background Report IV; Comparison oi
Risks from Land-Based and Ocean-Based Incineration,
Washington, D.C., March 1985c.
U.S. Environmental Protection Agency, Office of Solid Waste,
Regulatory Analysis for Proposed Regulations under RCRA for
Small Quantity Generators of hazardous Wastes, Washingtu .,
D.C. , June 1983d.
U.S. Environmental Protection Agency, Office of Solid Wast* ,
Regulatory Impact Analysis of Proposed Standard tor
Management of Used Oil, Washington, D.C., November 19b5e.
U.S. Environmental Protection Agency, Office of Solid
Regulatory Analysis of Proposed Restrictions on Land
Disposal of hazardous Pastes, Washington, D.C., 1985f.
U.S. Environ nental Protection Agency, Gu i d an c e on Fe a s i b_l 1 i t >
Studies Under CEKCLA. Washington, D.C.: EPA/540/G-85/UUJ,
June 1985g.
L'.S. Environmental Protection A£t»ncy, Superfund Public Health
Evaluation Manual, Draft, Washington, b.C., December lybbr.
U.S. Environmental Protection Agency, Office of Soiic Waste
Regulatory Analysis of Proposed Restrictions on Land
Disposal of Certain Solvent Wastes; braft, Washington,
D.C. , January 198ba.
U.S. Environmental Protection Agency, Office of Solid Waste,
Regulatory Analysis of Proposed Kestrictions on Land
Disposal of Certain Dioxin-Containing Wastes, Pratt Final
Report, Washington, D.C., January 198bb.
U.S. federal Energy Regulatory Commission, Natural Gas Storage
Plant, Ked Lake, Arizona; Final Environmental Impact
Statement. 1983.
U.S. General Accounting Office, "Improved Quality, A.-U'
Resources, and Consistent Oversight Needed if Regulatory
Analysis Is to Help Control Costs of Regulations," Report
Chairt.ian, Committee of Governmental Affairs, L.S. Senate,
November 2, 1982.
-------
-64-
'•jl. U.S. Mineral Manage \t-nt Service, Oil and Gas Lease Sale -
South Atlantic States: Final Environment Impact Statement,
1983.
38. U.S. Olticfc ot Science and Technology Policy, "Chemical
Carcinogens: A Keviev of the Science and Its Associated
Principles," 50 Federal Register, 1U372-10442, March 14,
1985.
(A) Ames Laboratory, health and Environmental Effects of Retusej-
Derivec Fuel (RUF) Production ana KLF/Coal Co-Firing
Technology. October 1983,
(H) Brookhaven National Laboratory, Nitrogen Oxides: health and
Environmental Effects Document III, September 1983.
(C) brookhavcn National Laboratory, health and Environmental
Effects Document tor Photovoltaic Energy Systems - 1983,
September 1983.
(D) Depart-ient of Engineering and Public Policy, Carnegie-Mellon
University, Improving Characterization and Treatment ot
Uncertainty in healt ana Environmental Risk Assessments,
December 1985.
(E) Harvard University, Assessment of Carcinogeniclty ot
Chemicals, Particularly hydrocarbons, July 1985.
(F) Energy and Environmental Policy Center, harvard University,
Analybib ot health Effects Resulting From Pollution
Exposures to Ambient Particulate Matter, September 1985.
((3) Inhalation Toxicology kesearcn. Institute, Potential health ano
Environmental Effects of the Fluidizeo bed Combustion of
Coal - Final Report LMF-IOb. October 1983.
(H) I\,w v,orp. , health and Environmental effects Document tor Oil
Shale - 198^. IwO-t^-Ubb-01, December 1984.
(I) Lawrence Livermore t.dtlonal Laboratory, Health and
Environments! Effects Document on Geothermal Energy - 19&2
Update, UCKL-533b3, Novetibtr 1983.
(J) Oak Ridge National Laboratory, Health and Environmental
Effects Document for the Liquid hetal Fast Breeder Reactor
Fuel Cycle - 1983. December 1983.
(K) Uak Kiage National Laboratory, Healtft and Environmental
Effects Document on Direct Coal Liquefaction - 1983,
OKNL/Tli-9287, October 1983.
-------
List of Acronyms
ADI
BLM
BOK
CEPP
CEQ
CEKCLA
CPSC
DHhi>
DOE
EDB
EIS
EPA
trtfi
FDA
F1FRA
FKA
FSIb
IRLG
MCL
ML) A
nsh.,
NAS
NEShAP
NHTSA
NIH
NIOS'-.
NPI,
NSF
OtCD
OPTS
OS HA
I.. T «
..x i . i
R» - »-i1
' I v 1 1
SAR
SCS
bD.v\
TSCA
US DA
VOC
acceptable daily intake
bureau of L-i 10 Management
Bureau ot Reclamation
barrel -per-d ay
Chemical Emergency Preparedness Progran.
Council m Environmental Quality
Cu-i^rehensive Environ cental Response, Compensaclt) *
Liability Act
Consumer Product Safety Commission
Department of health and hun^v Services
De^ r.•. Administration
Food and Drug Administration
Fi'il^ral -Insecticide, Fungicide and Rodenticide Act
tederal R . : • -oad Administration
Food Safety and Inspection Service
Interagt-ncy Regulatory Liaisoi, Group
UAX: ii'n contaminant level
me thy let i «M ianiline
Mine Safety anr1 health Administration
hazardo i- Ai'-
Administration
Lmissio
National
rational
National high.-;
National
National
national
National
of Science
i Standards for
./ Traffic Safety
T !-• i • ui-.c"-' of Health
Institute of Occupational Safety and heal-.
Priority List
be L>-' v Hjndation
Urganization fo^ Economic Cooperation aiu Development
Office ot Pesticides an,' Toxic Substances
Occup-itioi ."" ^i(rety and Health Admi i: s (.rntio .
Resource Conservation and Recovery Act
regulatory imp^ci analysis
recommended maxinu <: iritamination level
safety analysis report
Soil Conservation Serv/i:;e
f'.i^e Drinking Water Act
Toxic Subst^nt-pc Control Act
U.S. Department of Agriculture
volatile organic compound
L S. Envirorjacntal P-oteetloa
Libra, y Room 24C-i FH-27.1-A
40\ .1 Street, S.P,
Washington, DC
------- |