. "7. Concerns Related to Scientific Uncertainty, Policy Context, Institutional Capacity, and Social Implications." Animal Biotechnology: Science Based Concerns. Washington, DC: The National Academies Press, 2002.
The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Animal Biotechnology: Science-Based Concerns
Indeed, the essence of regulatory decision-making on health and environmental issues is to make judgments, in the face of uncertainty, about whether established standards have been met. Although it is impossible to prove the safety of a product or technologic application with complete certainty, regulatory scientists (scientists who are responsible for scientific evaluations for a regulatory agency) usually operate within established protocols for evaluating the safety of products or technologies and manage uncertainty by applying safety factors when estimating risks and by identifying additional studies that can provide data to reduce uncertainties. In the case of at least some applications of biotechnology to animals, however, scientific uncertainty will be a particular concern, due to the novelty of the health and environmental questions posed, and due to the lack of established scientific methods for answering them.
Uncertainties can be placed in three categories—statistical, model, and fundamental. These categories generally correspond to technical, methodologic, and epistemologic considerations, respectively, which also can be described as inexactness, unreliability, and insufficient knowledge (Funtowicz and Ravetz, 1992).
Statistical uncertainty—usually centered around the value of a single variable—is reduced most easily by additional data collection, leaving residual uncertainty that can be quantified. For example, the impact of bovine somatotropin (BST) use on milk production, IGF-1 levels in milk, or the incidence of mastitis in treated animals can be studied rather easily, and the probability distribution of values for each of these variables can be determined.
Model uncertainty results from not fully understanding interactions among variables in models used to predict the behavior of multivariate systems when one or more variables are changed. Model uncertainty inherently is more difficult to reduce and to quantify than statistical uncertainty. For example, the potential of transgenic fish to enter the natural environment and alter the marine ecology is a new concern for regulators and scientists that brings into play multiple variables and interactions; this issue poses novel scientific questions, and requires new data collection protocols and methods of analysis. Similarly, a transgene might have pleiotropic effects on multiple fitness traits, making the net effect difficult to predict. In Japanese rice fish engineered with a growth hormone transgene, for example, the disadvantage of a reduction in juvenile viability might be more than offset by the advantages of earlier sexual maturity and an increase in female fecundity relative to wild type (Muir and Howard, 2001). The Trojan gene example in Chapter 5 also shows how, as model uncertainty increases, an even more fundamental kind of uncertainty begins to appear.
Fundamental uncertainty results from indeterminacy, ignorance, or ignorance-of-ignorance. In the case of novel technologies, existing models might not apply. Moreover, if we are ignorant of the potential existence of a particular hazard, we might fail to consider it at all when attempting to estimate