摘要:How "default options" should be used in health risk assessment divides the risk analysisprofession. Some argue that these should be "hard": set by policy, generally biased to be "healthprotective" and requiring a substantial body of evidence to replace by decision-specific alternatives.Others argue that they should be science-based, identified by consensus of the professionalcommunity, replaced by whatever decision-specific information may be available to the analyst. Thispaper shows that both positions have validity, and that both are incomplete. Each kind of constructhas a useful role to play, but in different kinds of decisions. Because the two are different, we suggestgiving them different names, "default option" being assigned to the policy-based construct, "inferenceoption" (NRC, 1983) assigned to science-based assumptions, etc. We develop a theory that explainswhy these two different kinds of construct exist, and comment on some of the implications."Inference options" constitute an integral part of human health risk assessment, providingpractitioners with consensus theories, models, or parameters that can be used to bridge knowledgegaps in specific analyses. Because human health risk assessment is both considered "scientific" andemploys scientific reasoning, inference options must be treated as priors in an empirical-Bayesianinference process. Decision- or case-specific information modifies each prior according to thereliability of this information, with conflicts resolved by a scientific, weight-of-evidence process.Inference options are science-based "best estimates" and evolve through consensus within theprofessional community."Default options" constitute policy-derived components of particular kinds of decisions, servingas instructions to analysts. Such use is appropriate when many very similar, nontrivial decisions are tobe made by a particular agency. In these decisions, which we suggest calling "recurrent," value-judgments are prescribed in advance, usually by legislation; generally only two decision options existand the decisions usually turn on judgments made by experts. Authority to make these decisions isoften delegated (sometimes tacitly) to permanent staff members who have the requisite expertise.Policy-based default options exist in part because delegation of decision authority carries risksfor organizations; those to whom it is delegated may unwittingly make decisions differently fromsenior officials, and may thus in some way harm the organization. Thus, delegation of authority isalways conditioned by various forms of controls, including limits on the authority. We postulate thatdefaults serve as one means to control delegation risk. (They also simplify decision-making and makeit more consistent.) These default options need to be tailored to the policy ends served, whichgenerally means that some or all will be biased. They must be "hard," with "departure from" themrequiring a high standard of evidence and also assurance that choosing an alternative will still satisfythe policy ends of the decision process in which these are a part. Default options need to be developedin the same way as any other policy-implementation practices, including deliberations that engagethose who will be affected by the decisions. Such deliberations do not always take place.