Page 786 - The Toxicology of Fishes

766 The Toxicology of Fishes

Hazard Identification or Problem Formulation

In human health risk assessment, the term hazard identification is more commonly used to represent
this phase of the assessment. Beyond the difference in terminology, the activities undertaken in this
component of the risk assessments also vary, primarily in scope. In human health risk assessments,
hazard identification is largely related to identification and characterization of the contaminants of
concern because the entity (i.e., Homo sapiens) and the attribute of the entity (i.e., health) to be protected
are largely predetermined. In contrast, in ecological risk assessments ecological risk assessors are often
charged with considering entire assemblages or communities of species in selecting and defining assess-
ment endpoints within the context of the contaminants of concern. Human health risk assessments are
likely to have less ambiguous management goals (e.g., protection of humans from increased cancer risk
–6
at a 10 probability) than those for many natural resources (e.g., water quality that provides for the
protection and propagation of fish, shellfish, and wildlife; Clean Water Act, 1972/1977). This lack of
singular ecological management goals reflects the vast diversity in ecological systems and generally
increases the complexity of ecological risk assessments and underscores the importance of the problem
formulation component. It should be noted that well-defined management goals are not always the result
of greater scientific certainty but may simply reflect more well-defined health or environmental policies
or societal values.

Assessment Endpoints
Perhaps one of the greatest differences between human health and ecological risk assessment is not
intrinsic but rather the result of practical implementation—this difference is in the nature of the risks to
be assessed. Although there are no fixed rules about what appropriate assessment endpoints are for
human health or ecological risk assessment, widely held societal values have led to a fairly uniform
perspective that acceptable risk to humans from involuntary exposure to chemicals in the environment
should be essentially zero for all individuals. Although truly zero risk may have relatively little meaning
whenever exposure is nonzero, this perspective has led to the need for human health risk assessments
–6
capable of addressing levels of risk such as “increased cancer risk less than 10 ,” meaning that the
acceptable frequency of cancer induced by the stressor is one individual in a million. From the toxicology
perspective, this approach has several implications. Not only is protection of single individuals empha-
sized, but risk levels of one in a million pose a substantial technical difficulty in appropriately extrap-
olating exposure–response curves generated from experiments conducted using tens to hundreds of
observations down to a frequency of one in a million. In contrast, ecological risk assessments generally
focus on expression of risk relative to populations or communities of organisms rather than individuals.
Thus, the emphasis is not generally on extrapolating to extremely low frequencies of effects on individuals
but rather on establishing the rates of mortality, reduced reproduction, or other demographic parameters
that will adversely affect the long-term size or structure of the population. The question therefore shifts
from “how low must exposure be to have essentially zero effects on individuals” to “how much effect
can this population endure before it decreases below an acceptable size?” This difference in perspective
greatly influences the science necessary to assess risk.

Extrapolation and Testing

In human health risk assessment, data from several species (e.g., mouse, rat, dog) are often used to
extrapolate to a single species (humans), whereas in ecological risk assessments, data for one or a few
species must be extrapolated to many species. On the other hand, an advantage in ecological risk
assessment is the ability to conduct direct experimentation and collect data directly on the resources
that are the subject of the assessment. Data from toxicity tests and field sampling are generally available
for many species of fish and wildlife of concern in risk assessments, whereas data for chemical effects
on humans are generally limited to epidemiological studies of accidental or inadvertent chemical
exposures.