definition and risk-management goals. Similarly, selection of a management option or the evaluation of risk-management results can lead to a reformulation of the problem and might require additional data collection and analysis, such as congener-specific measurements of PCBs in key receptors or media. The extent to which additional site-specific information is collected should be balanced with the costs of conducting the ERA and the costs of risk management. The committee stresses that references to cost refer not only to monetary costs or risks but also to social and political costs of actions or lack of actions in a timely manner.
The problem-formulation stage for PCB-contaminated-sediment sites involves discussions among the various affected parties to identify the specific geographic areas of concern, all possible risks to humans and wildlife from immediate and long-term exposure to PCBs and from remedial activities, the identification and size of the populations potentially at risk, and the possible presence of co-contaminants at the site. This information is used to identify clearly the assessment endpoints, select measurement endpoints, and develop a conceptual model for the site. At most sites with PCB-contaminated sediments, human health assessment endpoints include both carcinogenic and noncarcinogenic effects (e.g., children born with learning dysfunctions). Special consideration should be given to certain sensitive subpopulations, such as women of child-bearing age, pregnant women, and young children. Other populations who eat fish from contaminated water ecosystems on a regular basis might also be at increased risk. Such populations include many American Indian tribes, immigrants from fishing cultures, such as Southeast Asia, and subsistence fishers who rely upon fish as a major source of protein. For ecological assessment endpoints, reproductive success and population sustainability of resident fish, piscivorous and other predatory birds, and marine mammals are often considered.
Assessment endpoints are used to select measurement endpoints, for which indirect effects, sensitivity and response time, diagnostic ability, and practicality issues are considered. Measurement endpoints are responses (e.g., litter size in mink) that can be measured more easily than assessment endpoints (e.g., reproductive success in mink) but are related quantitatively or qualitatively to the assessment endpoints. Whenever practical, multiple measurement endpoints should be chosen to provide additional lines of evidence for each assessment endpoint. For example, for humans, it might be possible to measure PCB concentrations in food and in human tissues. For