Risk Assessment

views updated

RISK ASSESSMENT

Many decisions involve an intuitive assessment of risk; this subjective risk assessment is usually called risk perception. Risk assessment is also a formalized approach to evaluating risk, often defined as a function of the probability and magnitude of loss or harm from an event. Risk assessment is often thought of as ethically obligatory, but since it can be done in more than one way, it is itself subject to ethical assessment.


Risks are routinely assessed formally for a wide variety of human endeavors, from drinking tap water to operating nuclear power plants; for natural hazards such as earthquakes, hurricanes, and floods; and for the human use of and exposure to chemicals and other substances such as arsenic or phthalates. Risks may also be defined and assessed in terms of specific harms or losses to people, for example a person's lifetime risk of dying of heart disease, or aquatic ecosystem risks from anthropogenic eutrophication (that is, being overburdened with nutrients as a result of human action). While failing to assess risk can lead to Faustian bargains with the future, risk assessments for public policy can be risky in themselves, as illustrated by the effects of transnational debates about risk assessments of genetically modified organisms, vaccines, and terrorism.

Methods

As described in Risk Assessment in the Federal Government (known as the "Red Book," 1983), risk assessment consists of four steps: hazard identification, dose-response assessment, exposure assessment, and risk characterization. More broadly, risk assessment entails identifying and characterizing an underlying hazard—including its sources, pathways, effects for given exposures, and mitigating factors, and estimating the associated contingent probabilities. In effect, formal risk assessment requirements are intended to insure that human and even ecological health is considered in decisions with other primary objectives.

For example, product risk assessment may be required by law, as in the case of new pharmaceuticals in the United States. In the United States, the Food and Drug Administration (FDA) assesses the adequacy of new drug risk assessments, including how they are conducted. The FDA also determines what constitutes permissible risk for a licensed product. As risk assessments are generally conducted in the service of specific risk management objectives, the two are mutually dependent (Committee on Risk Assessment of Hazardous Air Pollutants 1994). In some venues separation of risk assessment and risk management is considered critical to protect the science of risk assessment from contamination by management or political pressures. However, many formal risk assessment processes now include participation by multiple stakeholders to deliberate about risk management objectives and values, in addition to experts' technical analyses (Stern and Fineburg 1996).

Human health risk assessments are of necessity carried out at a population or group level—that is, for a statistical person rather than an identified individual. They are based on extrapolations from animal studies; on experimental tests of human product use, which usually involve relatively small samples; or on epidemiological studies, which rely on statistical controls. Recent developments in risk assessment have included the ability to tailor risk assessment results interactively for subpopulations, as is illustrated by online risk calculators that determine an individual's risk based on a few personal characteristics. But individual differences can make the population health risk assessments applied in policy decisions more or less applicable, for which reason minority populations may be poorly served by general risk assessments. An example in point is airbags in cars, which when designed to optimally protect average adults may harm or kill children.

Environmental risk assessment, as required for example in environmental impact statements, has focused largely on risks to human health and the economy, but increasingly addresses ecological endpoints. Because selection of assessment endpoints can determine the structure and outcome of decisions, it is inherently controversial. Assessing risks from ozone only in terms of economic loss from damage to automobile tires paints a very different picture of the size of the risks than if the assessment also takes into account acute respiratory or cardiovascular events triggered by exposure to ozone, or possible ecological effects of ozone, such as reduced growth rates and plant deformation.

Basic Issues

By focusing on probabilistic loss, risk assessment frames management choices in terms of threat reduction and loss avoidance. Common criticisms of risk assessments have included that they are based on an overly narrow conceptualization of benefits, or that the dimensions of harm included are insufficient or inappropriate. It is difficult to incorporate into a risk assessment even proxy measures for intangibles—such as quality of life—or other poorly defined or understood endpoints. In part to take into account uncertainties, risk assessments are sometimes designed to produce estimates of risk that err on the high side, for example by using upper bounds of estimated risks, rather than averages. Those risk assessment procedures that have been codified by government entities incorporate scientific procedures, including requirements for representative empirical data, statistical analyses, and quality control in the form of peer review. Some also include ethical requirements, such as human subjects review, or the participation of parties who may have a substantive interest in the value at risk.

Four issues are key to risk assessment as currently practiced. The first is what is valued, how and by whom it is valued, and the distributive implications thereof. The selection of assessment endpoints can have far-from-obvious implications, as the airbag example illustrates. Assessing values remains a methodological and ethical challenge (Fischhoff 1991, Slovic 1995).

The second is the treatment and interpretation of uncertainty—both uncertainty stemming from limits to what is known, and uncertainty stemming from inherent variability (see Morgan and Henrion 1990). Especially in the case of extremely rare and catastrophic events, the selection of a distribution function or simulation procedure with which to analyze uncertainties can influence the outcome of the assessment considerably. Similarly, choosing how to represent the results of the risk assessment and the uncertainty therein can influence how recipients interpret and use the assessment.

The third key issue is the substitutability implied or assumed by risk assessment, as it often requires comparative values. As has been illustrated in discussions of protected values and irreversible effects, in reality trade-offs are sometimes impossible or unethical.

Fourth is that technically competent risk assessment requires significant resources, is both analytically and data-intensive, and can be difficult to interpret. Risk assessments that are carried out for new drugs, for example, require expertise in toxicology and epidemiology and investments in large studies, which still may not be large enough to discover devastating rare or long-term adverse effects.

Risk assessments may produce risk characterizations that are not readily used to compare or prioritize risks. For example, ecological risk assessments may conclude simply that a specific species is at some risk of extinction, while a human health risk assessment may produce an estimated probability of a specific health endpoint within a given timeframe, for example a five percent probability of being diagnosed with breast cancer within five years. Comparing the two is difficult.

For this reason it is desirable that risk assessment outcomes be translatable to a common measure, such as an abstract measure of utility, or monetary value. Summary endpoints like the probability of human mortality or morbidity, or economic loss, can be presented in a common metric that facilitates at least some comparisons, such as disability adjusted life years, or monetary value. But choice of a common metric itself can be problematic, both because individuals may not agree on the equivalence of different forms of bodily injury or harm and because not all endpoints can be equally well represented by all measures. In addition, some measures, such as dollars, carry their own meaning, which may or may not facilitate the risk assessment depending on how that meaning is construed.

However, no single metric or endpoint necessarily constrains environmental, technological, or human health risk assessments. Although many risk assessors with economic training might prefer to use dollars as a summary endpoint, doing so is not a requirement of risk assessment, but a methodological choice with ethical implications. The identification and definition of possible endpoints to consider, the valuation of these, and the estimation of their contingent probabilities all entail some degree of judgment and choice.

ANN BOSTROM

SEE ALSO Risk;Risk and Emotion; Risk-Cost-Benefit Analysis;Risk Ethics;Risk Perception;Risk Society.

BIBLIOGRAPHY

Committee on the Institutional Means for the Assessment of Risks to Public Health, Commission on Life Sciences, National Research Council. (1983). Risk Assessment in the Federal Government: Managing the Process. Washington, DC: National Academy Press. Known as the "Red Book," this is probably the most cited publication on risk assessment.

Committee on Risk Assessment of Hazardous Air Pollutants, Board on Environmental Studies and Toxicology, Commission on Life Sciences, National Research Council. (1994). Science and Judgment in Risk Assessment. Washington DC: National Academy Press. An update on the "Red Book."

Fischhoff, B. (1991). "Value Elicitation: Is There Anything in There?" American Psychologist 46(8): 835–847. Summarizes the literature on value elicitation succinctly and with eloquence.

Morgan, M. Granger, and Max Henrion, with Mitchell Small. (1990). Uncertainty: A Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis. Cambridge, UK: Cambridge University Press. A key publication on uncertainty.

Slovic, P. (1995). "The Construction of Preference." American Psychologist 50(5): 364–371. A must-read for anyone interested in values or preferences.

Stern, Paul C., and Harvey V. Fineberg, eds. (1996). Understanding Risk: Informing Decisions in a Democratic Society. Washington, DC: National Academy Press. Articulates the importance of deliberation as well as analysis in risk assessment.

UK Royal Society. (1992). Risk Analysis, Perception, and Management: Report of a Royal Society Study Group. London: The Royal Society. An important summary of EU perspectives on risk assessment and related issues.

More From encyclopedia.com