28 THE FIVE-YEAR OUTLOOK 



rather than the products of those activities. Again, the 

 reason why biomedical research is a candidate for regula- 

 tion is obvious: to protect individuals from excessive 

 physical or psychological harm resulting from being sub- 

 jects of that research. However, without research on 

 human subjects, no new drugs could ever be tested and 

 marketed, nor could new medical or surgical procedures 

 be developed. Recognition of the conflict between the 

 potential benefits to be derived from research using 

 human subjects and the potential harm to the subjects 

 themselves makes the form and extent of regulation of 

 research on human subjects a particularly difficult policy 

 problem. 



Science can make two kinds of direct contributions to 

 framing regulatory policies for the assessment and man- 

 agement of risk. First, it can supply empirical data about 

 the prevalence and effects of a given set of potential 

 hazards. Second, it can provide a battery of analytical 

 techniques for assessing invariably complex situations 

 and for evaluating the effects of alternative regulatory 

 strategies (SSRC-1). 



Science can also contribute indirectly to other aspects 

 of policy decisions about the management of risk. For 

 example, a major question concerning risk management 

 is: What levels of damage to health or to safety or to the 

 environment from a given technological advance are ac- 

 ceptable? It is now widely recognized that achieving zero 

 risk while still reaping the benefits of scientific and tech- 

 nological advance is impossible. Therefore, a question 

 that remains is: What level of risk is tolerable or accept- 

 able? (NRC-Obs.; AAAS-5). Answering that question 

 requires an understanding of public values about health 

 and environmental quality, as well as an understanding of 

 the values placed on the benefits that might also be derived 

 from the technology. Scientific methods can be used to 

 gather information about public perceptions and values 

 and. thereby, can contribute indirectly to the nonscientific 

 aspects of risk management decisions (SSRC-1). 



SCIENTIFIC ASSESSMENTS OF RISK 



Scientific contributions to risk assessment often involve 

 two different types of research and analysis. First, meas- 

 urements have to be made both to establish, quantitatively. 

 the levels at which the hazard (a suspected toxic chemical , 

 for example) occurs in various settings (such as a factory 

 or dump), and to relate different levels of occurrence to 

 different classes or degrees of damage to health, safety, or 

 the environment. Other types of measurements may also 

 be necessary — for example, atmospheric transport and 

 absorption rates in the case of air pollution from coal-fired 

 power plants. 



Makmg those types of measurements and establishing 

 precisely the necessary correlations are often very diffi- 

 cult. Estimating specific levels of possible damage can be 

 greatly facilitated by a detailed understanding of funda- 



mental physical, chemical, psychological, and biological 

 processes, but even in those cases some residual uncer- 

 tainties are common. The situation is further complicated 

 in some cases since the sensitivity of instruments capable 

 of detecting such things as contaminants in food, water, or 

 the atmosphere is continuing to increase, so that minute 

 traces may be found in previously unsuspected situations. 

 Foirthermore, there are cases where there may be no way to 

 quantify risks, such as the potential psychological damage 

 to a subject of a behavioral science experiment. In short, 

 there is almost always a certain amount of residual uncer- 

 tainty associated with scientific assessments of risk 

 (NRC-Obs.; AAAS-5). 



The second type of research and analysis required to 

 assess risks involves extrapolation of present knou ledge 

 about hazard levels and their correlations with damage to 

 health, safety, or the environment to future situations. In 

 some cases, the extrapolations are relatively straightfor- 

 ward. In others, they are at best educated guesses and, as 

 such, are open to debate. It is also particularly difficult to 

 evaluate what might have been the risk of a particular drug 

 that was never marketed or a power plant that was never 

 built. Therefore, although high levels of precision are 

 sometimes approached, there frequently remains consid- 

 erable uncertainty in. the analysis and measurement of 

 risk. Risk assessment is not yet a sufficiently precise 

 activity to cover all cases equally well or with equal levels 

 of certainty, and further methodological refinements are 

 needed. Dealing with the uncertainty increases the judg- 

 mental burden on policymakers. 



WEIGHING RISKS, COSTS, AND BENEFITS 



The issue of risk can never be resolved in its own right. 

 That is, individuals and populations are never asked to 

 accept a risk for its own sake. Rather, risks are acceptable 

 only to the degree that they are a necessary price to be paid 

 for anticipated benefits. It follows that any creditable 

 regulatory policy has to be based on a comparison of risks 

 and benefits, and on a broadly accepted consensus that the 

 anticipated benefits outweigh the anticipated risks 

 (AAAS-5). This need to weigh both the costs and the 

 benefits of potentially hazardous technologies, as well as 

 the costs and benefits of regulating those technologies, 

 was recognized in President Reagan's Executive Order on 

 Federal regulation.' 



Thus, yet another type of research and analysis consist- 

 ing of formal techniques for weighing sets of risks, costs, 

 and benefits is sometimes brought to bear on risk assess- 

 ment and management. The use of such cost-benefit 

 analyses, however, it not without difficulties. During the 

 past decade, there has been increasing interest in expand- 

 ing formal methods of cost-benefit analysis to the realm of 

 risk-benefit comparison. As originally and narrowly con- 

 strued, a cost-benefit analysis of an intended policy alter- 

 native simply totaled the anticipated monetary costs of a 



