Office of Pesticides
and Toxic Substances
Washington DC 20460 February 1990
Office of Pesticide Programs
<>EPA Environmental
Fact Sheet
RISK/BENEFIT BALANCING UNDER THE FEDERAL
INSECTICIDE, FUNGICIDE, AND RODENTICIDE ACT
The Federal Insecticide, Fungicide, and Rodenticide Act
(FIFRA) governs EPAxs regulation of pesticides. FIFRA requires
that all pesticides intended for use in the United States be
registered (licensed) by EPA to ensure that they do not cause
"unreasonable adverse effects on man or the environment". FIFRA
defines unreasonable adverse effects as "any unreasonable risk to
man or the environment, taking into account the economic, social,
and environmental costs and benefits of the use of any pesticide."
Pesticides may pose some risk because they are meant to kill or
control insects, weeds, rodents, and other pests. But even though
pesticide use entails some risk, pesticides provide substantial
benefits to society. Hence, FIFRA, unlike other environmental
statutes administered by the Agency, requires EPA to balance these
risks and benefits in the pesticide decision making process.
In order to be able to balance risks and benefits, EPA
conducts complex risk and benefits assessments that employ the best
scientific and economic analyses currently available. FIFRA
authorizes EPA to conduct such full and public risk/benefit
analysis of pesticides through its Special Review process or under
circumstances where data show that a pesticide may pose an imminent
hazard to human health or the environment. These reviews are
initiated on the basis of adverse human or ecological effects data
which -ome to the Agency's attention subsequent to the initial
regis. ^tion of the pesticide. A Special Review may result in
cancellation of some or all uses of the pesticide, the imposition
of risk reduction measures such as protective clothing, or
continued registration with no changes.
This fact sheet describes both risk/benefit analysis, and how
it was applied to a recent EPA decision on whether or not to permit
the continued use of a problem pesticide.
RISK ASSESSMENT
EPA*s basic approach to evaluating pesticide risk is
conservative and protective of public health. These are four steps
in the risk assessment process: hazard identification,
-------
dose/response asessment, exposure assessment, and risk
characterization.
Hazard Identification The objective of hazard
identification is to assess the nature of any adverse health effect
that may be caused by the pesticide based primarily on the results
of laboratory studies conducted with animals. For example, is the
chemical an eye irritant, does it cause acute poisoning, does it
cause birth defects, does it cause cancer?
Exposure Assessment Once the type of hazard has been
identified, the exposure must be assessed. This involves
estimating the level, duration, frequency and route of exposure.
For example: Are people who often mix and apply large amounts of
a pesticide overexposed? Is there risk to people who consume foods
and drink water containing residues of the pesticide? Will use of
the pesticide affect non-target organisms, including endangered
animal and plant species?
Dose/Response Assessment The dose/response assessment
explores the relationship between a level of exposure and the
occurrence of an adverse effect. Since epidemiological studies
with reliable quantitative data on exposure are rarely available,
the Agency must rely on the results of experimental animal studies
to estimate dose/response relationships. EPA extrapolates from
animal studies to potential effects on humans, and from exposures
at high doses in those animal studies to the substantially lower
doses to which humans are typically exposed. Obviously, this
involves uncertainties; EPA is quite conservative in resolving
them.
For almost all chemically-related toxic effects (such as birth
defects, reproductive, or non-carcinogenic chronic effects), the
highest dose level at which the effect is not observed in the
animal study (No Observable Effect Level or NOEL) is determined and
an uncertainty factor (typically 100 or more) is applied to
establish a human exposure level that will pose essentially no
human risk. For example, an animal study may show that at 1000
milligrams of the substance in the daily diet of a 1 kilogram body
weight test a al (1000 mg/kg/day), the animal showed no adverse
effects. By allying a 100-fold uncertainty factor, daily exposure
of a human to 10 milligrams per kilogram of body weight would be
considered to be acceptable for a human. In conducting pesticide
risk assessments, EPA makes a practice of evaluating all
potentially toxic effects, but generally focuses its quantitative
risk/benefit balancing process on the effect observed at the lowest
dosage level.
Where animal studies indicate that the pesticide chemical has
induced cancer at relatively high dose levels, EPA ordinarily
presumes that there is no dose at which some level of carcinogenic
effect would not be observed. (This presumption sometimes could
-------
perhaps be overcome by extremely complex and costly studies, but
usually overcoming the presumption is practically impossible.)
Thus, EPA uses a mathematical model that assumes some effect could
be caused by any dose, even one at a very low level. Such a model
expresses the worst-case qualitative carcinogenic "potency" of the
pesticide, or the worst-case probability that a given low dose will
produce a response in the animal species.
As an index for regulatory decisions involving carcinogens,
EPA's stated policy is that lifetime incremental cancer risks from
exposure to a pesticide in the diet generally should not exceed
one-in-one million or 0.000001 (1 x 10-6) meaning one-in-one
million risk over and above the background cancer risk of 1 in 4.
This is the concept of "negligible" risk applied by EPA and the
Food and Drug Administration.
Risk Characterization Finally, the risk is characterized
by integrating the above factors. This usually involves
extrapolating exposure in animals to humans. It is scientifically,
legally, and ethically impracticable in most cases to do
experiments to show whether humans would have different responses
to chemical exposure than test animals. Thus, EPA ordinarily
assumes that human response is similar to that of test animals.
EPA seeks to express a risk qualitatively and, where
appropriate, quantitatively, and to give some perspective to the
risk by summarizing the strengths and weaknesses of the supporting
data and assumptions. Through risk analysis, the Agency is able
to determine areas where additional data may be needed or where
regulatory action may be warranted.
BENEFITS ASSESSMENT
In deciding whether to cancel a pesticide, EPA conducts
benefits assessment. A benefits assessment follows guidelines that
take into account both biological and economic factors. Biological
factors involve information about target pests on each site or
crop. Economic factorr ^volve the effects of various pest control
strategies on the vari*. _. businesses which produce and distribute
commodities to the consumer. Economic impacts are dependent on the
treated commodity or registered use sights of the chemical in
question. So that means that EPA must do a benefits assessment for
each registered use sight. One Special Review chemical can
therefore lead to dozens or even scores of benefits assessments.
Sometimes EPA must tailor its general approach to the particular
pesticide and market conditions encountered. A benefits analysis
includes:
Biological Analysis Benefits are specific to each use site
for example, corn, soybeans, or tomatoes, etc. The estimation
of benefits derived from the use of a given pesticide is a multi-
step process. The first step in this process is the identification
-------
of the pesticide's use (i.e., agricultural, industrial/commercial,
or home use), and the principal pest(s) that are controlled
through its application (e.g., insects, weeds, rodents).
Next, EPA takes a comprehensive look at the problem by
developing a thorough understanding of the pest populations, the
effectiveness of chemical and non-chemical alternatives, the
geographic distribution of the pest problem, potential crop yield
impacts, and alternative use practices which might mitigate the
risk. In other words, the process starts with gathering basic
usage and product performance data in order to determine really
how important the chemical is to the production of any crop from
corn to mint.
These data are obtained through comprehensive literature
searches, data bases created and maintained by EPA, and extensive
communications with pesticide users. The latter involves
requesting information from pesticide applicators, other government
agencies (the Departments of Agriculture and Interior), scientists
from state land grant universities, and organizations representing
various user groups. When little or no empirical data exist, the
Agency relies upon the experience and observations of pesticide
applicators, scientists directly involved in studying pesticide
effectiveness, and, to some extent, independent agricultural
consultants. These data are by no means easy to obtain and
sometimes require literally scores of telephone calls, etc. But
the Agency is committed to obtaining the most comprehensive,
accurate, and up-to-date information possible.
Economic Analysis The next step of a benefits assessment
is to express the biological impacts of yield loss, reduced crop
quality, and alternative pest control methods in economic terms.
The key question here is will it cost more to produce the crop?
EPA normally evaluates several types of economic impacts. The
first step is to determine the relative cost of alternative pest
control strategies. In some cases, EPA finds that there are
relatively minor economic impacts, because cost effective
substitutes (chemical and non emical) exist and are readily
available. However, in other ea_^s equally effective alternatives
do not exist. In this scenario users may have to modify their
current use practices, changes which often result in higher costs
of production for relevant agricultural and non-agricultural goods
and services.
An increase in production cost can be passed down to
intermediate processors and ultimately to consumers in the form of
higher food prices or more costly control of pests in the home.
In determining economic impacts, EPA takes into account such
influences such as agricultural subsidies, quotas, or allotments,
supply and demand factors, and international trade issues.
-------
The Agency estimates both short- and long-term impacts at the
user level. This analysis accounts for the net impacts to
producers/suppliers by examining economic effects on users and
nonusers of the pesticide under review. Based on this information
short- and long-term effects in down-stream markets (i.e.,
processors, distributors, consumers) can also be estimated.
Potential local/regional impacts are also identified as part of
this analysis. These local effects may result from movement of
economic activity to foreign markets, or to other areas of the
country where pest infestation is not a problem (and therefore
pesticide availability is relative unimportant).
AN ILLUSTRATION OF RISK/BENEFIT BALANCING DIN08EB
To illustrate this highly complex process, consider the case
of the herbicide, dinoseb. In October 1986, EPA issued a formal
notice of intent to cancel and deny all registrations for pesticide
products containing dinoseb. Simultaneously, EPA issued an
emergency suspension order to immediately stop dinoseb use during
the time to complete the cancellation proceedings. These
regulatory actions were based on the risk/benefit balancing
scenario described below.
Hazard Identification In assessing the risks of exposure
to dinoseb, the first step involved identifying the hazard, based
on the results of animal studies. The hazard evaluation of dinoseb
focused on animal test results that it may cause birth defects in
females exposed to dinoseb during pregnancy, and may also cause
sterility or decreased fertility in males, acute toxic poisoning,
and other potential adverse effects on health and the environment.
Dose/Response In assessing the test-animal dose response
for birth defects, EPA used an oral feeding study with rabbits to
provisionally set a NOEL at 3 milligrams per kilogram of body
weight per day (mg/kg/day) meaning that adverse effects in test
animal offspring were apparent in all oral exposure levels higher
than 3 mg/kg/day, the lowest dose admini red.
Exposure Assessment EPA estimated potential human exposure
based on either actual data or surrogate data. Surrogate exposure
data is simply data from studies of exposure to other pesticide
chemicals (surrogates) that are of similar formulation to, and are
applied in the same way as, the pesticide in question. Appropriate
adjustments in the calculations are made to reflect differences in
the rate of active ingredient applied per area, per hour, etc. In
the case of dinoseb, three exposure scenarios were identified:
- Possible dietary exposure to the public through
consumption of food or drinking water containing
residues of dinoseb.
-------
- Occupational exposure to workers who mix, load,
and apply dinoseb.
- Secondary or coincidental exposure to bystanders,
farmworkers, and others who could be exposed
to dinoseb through spray drift, contact with
residues in treated fields, or even contact with
contaminated clothing or farm equipment immediately
after dinoseb application.
Risk Characterization For dinoseb the main risk of concern
was the 3 mg/kg/day NOEL for birth defects in rabbits. EPA
compared this NOEL from laboratory studies with expected human
exposure levels to obtain numerical margins of exposure (MOE).
(NOEL divided by exposure equals margin of exposure).
Even when certain "worst case" assumptions regarding dietary
exposure levels were factored into these calculations, the MOE for
the risk of birth defects occurring from eating foods from crops
treated with dinoseb was found to be ample over 2700.
Similarly, from consumption of drinking water, the MOE was roughly
2430. However, for worker exposure, in many instances the
estimated worker exposure levels were equal to or greater than the
NOEL of 3 mg/kg/day leaving essentially no margin of exposure
against the hazard of birth defects in the offspring of pregnant
workers handling the pesticide.
The benefits assessment of dinoseb was driven by the risk
factors described above. Based on data from the U.S. Department of
Agriculture and other sources, EPA conducted an assessment of the
benefits of dinoseb by calculating the short-and long-term economic
impacts expected to occur if dinoseb were unavailable for certain
registered uses* Dinoseb use sites included soybeans, peanuts,
cotton, snap beans, potatoes, green peas, grapes, alfalfa, almonds
and walnuts, berries, and hops. For both short- and long-term
scenarios, the largest user impacts were projected for potato and
peanut growers. The overall impacts of removing noseb from the
marketplace were estimated at the user level in ...e range of $80
to $90 million.
The Agency quantifies both risks and foregone benefits for
several regulatory options. Economic impacts are typically
quantified in the monetary terms described above. In deriving
policy, annual costs are compared to the positive health or
environmental effects achieved through the various option. By
analyzing risk reduction and subsequent costs (i.e., foregone
benefits) for each policy option, the Agency can determine and
implement the most feasible and protective regulatory action.
In the case of dinoseb, EPA determined that the risks of birth
defects from occupational and secondary exposures to dinoseb
-------
presented significant risk to developing fetuses. There was also
convincing evidence exposure to dinoseb under both exposure
scenarios posed additional risks of adverse reproductive effects
in males and acute toxic poisoning. EPA considered a range of
measures to reduce the risk, which included additional protective
clothing, protective farm equipment, lower application rates, and
product reformulation. EPA determined that none of these measures
could reduce the risks to reasonable levels. Therefore, though the
benefits of dinoseb were substantial in terms of monetary value,
EPA concluded that risks of exposure to dinoseb clearly outweighed
its benefits, not only in the long-term but also during the
interval of time required to conduct usual cancellation
proceedings. Thus, in October 1986 the Agency opted for the most
drastic remedial option available under FIFRA emergency
suspension calling an immediate halt to the sale and use of a
pesticide while cancellation proceedings are conducted.
However, in reaching a decision to immediately suspend
dinoseb, EPA did not attempt to evaluate the economic benefits from
use of dinoseb for each affected crop in each geographic region.
Rather, the Agency focused on national impacts and on those crops
involving the largest amount of use and perdent of the crop
treated. A more extensive analysis would not have been consistent
with the Agency's concern about the imminent hazard to pregnant
women exposed to dinoseb.
As a result, growers of certain minor crops grown in the
Northwest sued EPA, asserting that the suspension was invalid
because EPA did not properly conduct its benefits analysis and that
there were modifications to use practices less drastic than
suspension that would have sufficiently reduced the risk. The
growers also complained that EPA could have delayed the suspension
of use of dinoseb on their crops, which would not have occurred
until several months after the suspension order was issued, while
EPA completed the benefits assessments for their uses. (The
suspension of the major uses of dinoseb, such as peanuts and
potatoes, was never challenged.) A district judge agreed with the
plaintiffs and issued an order allowing the use of dinosel- i the
minor crops if a series of use precautions were observed.
An appeals court later ruled that the district judge had no
authority to modify the suspension order by imposing conditions of
use, but by then the use had already occurred. Meanwhile, EPA
completed its assessment of the benefits of all dinoseb uses in
preparing for the cancellation hearing. However, all the
registrants of dinoseb eventually agreed to withdraw their
registrations, making both a cancellation proceeding and further
court proceedings moot.
Some of the other cases interpreting the suspension
provisions of FIFRA are not consistent with the results in the
Northwest dinoseb suspension case. However, if FIFRA is not
-------
8
amended or clarified, the need to fully assess economic benefits
prior to emergency action could prevent EPA from taking timely
action to protect human health. Under President Bush's food safety
plan announced in October 1989, EPA would have the flexibility to
order a suspension when there is substantial risk to human health
without consideration of economic benefits.
------- |