mechanisms of action are well understood, it will be still be very difficult to determine its dose-response relationship accurately enough to predict doses that correspond to risk as low as one in a million.
The strength of the two-stage model for application in cancer risk assessment is its ability to use information about intermediate steps in carcinogenesis; however, it is difficult to characterize these steps. Few experimental data sets now available provide all the types of data required. Before the two-stage model can be adopted for routine health risk assessments, it will be necessary to expand current rodent bioassay methods so that the necessary data are generated. The two-stage model can be used now to gain insights into induced carcinogenesis, such as identifying and characterizing the critical events, as well as to examine the ranges of assumptions. The committee strongly encourages further development and continued applications of the two-stage model to gain insight into its usefulness.
Two general approaches have been used for fitting two-stage models to data. One involves specifying trial values of parameters and simulating the subsequent tumor response. Values are then varied until the realizations conform to the data. The second approach involves applying standard statistical data-fitting methods (e.g., the methods of maximum likelihood). The former approach can be quite useful in some circumstances (such as exploratory data analysis). However, the committee encourages the use of formal statistical methods, whenever possible, to estimate values of parameters, assess goodness of fit, calculate statistical confidence-intervals for values of parameter sand risk estimates, and determine the extent to which the model is consistent with other mathematical representations and ranges of risk.
Two-stage models can be used as a basis for decision-making if there is sufficient mechanistic understanding and a sufficient data base for the chemical in question. At present, it is recommended that the two-stage model be used primarily to increase understanding. For health risk assessments, the two-stage model can be used with other models to add perspective and scope to the evaluation.
A Paradigm for Ecologic Risk Assessment
The third issue addressed by the committee and the subject of the last
report in this volume is a conceptual framework for ecologic risk assessment, defined as the characterization of the adverse ecologic effects of environmental exposures to hazards imposed by human activities. The workshop held on this subject had three principal goals: (1) to survey existing approaches to ecologic risk assessment through case studies representing various types of environmental stresses, (2) to consider the feasibility of developing a consistent framework for ecologic risk assessment analogous to the four-part health risk assessment framework proposed in the 1983 NRC report, and (3) to identify major scientific uncertainties and research needs common to many types of ecologic risk assessments.
The committee's principal conclusion is that, despite the diversity of subject matter and approaches taken in many different studies of ecologic stresses, a conceptual framework similar in form to that of the 1983 framework is applicable to ecologic risk assessments. However, for general applicability to ecologic assessments, the 1983 scheme requires augmentation to address some common grounds between science and management, primarily because of the need to focus on appropriate questions relevant to applicable environmental law and policy under different circumstances. Specifically, the scheme needs to address the influence of legal and regulatory considerations on the initial stages of ecologic risk assessment and the importance of characterizing ecologic risks in terms that are intelligible to risk managers. The committee's opinion is that such augmentation is as important for human health risk assessment as it is for ecologic risk assessment.
Although ecologic risk assessment and human health risk assessment differ substantially in their scientific disciplines and technical problems, the committee believes that the underlying decision process is the same for both. Therefore, the committee recommends that a uniform framework be adopted and applied to ecologic and human health risk assessment—a framework that is flexible and able to facilitate communication between scientists and risk managers. The committee extends the 1983 framework to satisfy those requirements.
The committee recommends that the use of risk assessment in strategic planning and priority-setting be expanded so that financial resources of state and federal environmental agencies can be focused on critical environmental problems and uncertainties.
The committee also recommends that research programs be estab-
lished and maintained to improve the credibility of ecologic risk assessments and that ecologic risk assessments be followed by systematic research and monitoring to determine the accuracy of their predictions and to resolve remaining uncertainties.
The committee identified the following kinds of research as likely to provide major opportunities for advancement of ecologic risk assessment:
Extrapolation across scales of time, space, and ecologic organization.
Quantification of uncertainty.
Validation of predictive tools.
Valuation, especially quantification of ''nonuse" values (values for environmental attributes that cannot be bought or sold).
Finally, the committee recommends that expert committees drawn from the academic community, the private sector, and regulatory agencies develop technical guidance on the scientific conduct of ecologic risk assessments.