Use of decision analysis to include uncertainty in risk assessments and regulatory decisions

Conference · Sun Dec 31 04:00:00 EST 1995

OSTI ID:211978

Peterman, R ^[1]

---

1. Simon Fraser Univ., Burnaby, British Columbia (Canada). School of Resource and Environmental Management

Management of ecological risks is essentially a decision problem. Given a particular management objective, a regulatory decision must be made that fully takes into account known information, as well as uncertainties in that information. This paper will review the method of decision analysis and illustrate its applicability to toxic substances. Decision analysis, originally developed in the field of business, is now being applied in other areas of resource management including fish and wildlife management because there, many examples have shown that it is inappropriate to use only the best ``point estimates`` of parameters in an analysis rather than taking uncertainty into account. There are 8 elements to decision analysis. (1) A management objective, such as minimizing the expected loss, provides a means of ranking actions. (2) The analysis will help choose among a set of alternative actions such as different quantitative values for some guideline for a toxic substance. (3) Uncertain states of nature may be included for a component such as a parameter reflecting the steepness of the dose-response relationship. (4) Probabilities must be stated for each of the alternative states of nature. These can be derived from the data through Bayesian statistical analysis. (5) A model is used to determine the consequences of each combination of states of nature and action. (6) A decision tree or decision table lays out the calculations in a systematic framework. Expected values of the measure of how well the management objective is met are calculated by weighting each state of nature by its respective probability of occurrence. (7) Alternative actions are then ranked based on these expected values. (8) A sensitivity analysis is done on the rank order of management actions to determine how robust the ranking is to the assumptions made and parameter values used.

• OSTI ID: 211978
• Report Number(s): CONF-9511137--; ISBN 1-880611-03-1
• Country of Publication: United States
• Language: English

Related Subjects

29 ENERGY PLANNING, POLICY, AND ECONOMY
DATA COVARIANCES
DECISION TREE ANALYSIS
MATHEMATICAL MODELS
POLLUTION REGULATIONS
RESOURCE MANAGEMENT
RISK ASSESSMENT
TOXIC MATERIALS