A systems-level approach to dealing with bioterrorism threats, especially those involving communicable diseases, is needed. This approach must consider the integration of multiple modes of management, risk analysis in the face of inherent uncertainties concerning what agents will be introduced, and potential interactions among multiple biological agents. Such research is likely to rely heavily on the techniques of operations research, especially models that can be used for scenario development and training, for rapid response following detection of infected individuals, and for redesigning current systems (including possible patterns of movement) in order to make societies less susceptible to catastrophic outbreaks. Indeed, all of this argues for major development of modeling capabilities.

UNCERTAIN UNDERSTANDING OF THE EFFECTS OF BIOLOGICAL WEAPONS

Modeling the likely outcomes of different bioterrorism attacks is important for two reasons. It provides insight into the severity of the threat posed by the proliferation of biological weapons, and it allows one to estimate the effectiveness of different defensive responses (and hence the priority one should assign to each). Modeling efforts over the past decade, at least those publicly available, tend to emphasize worst-case scenarios—broadscale attacks involving millions of human casualties, if not fatalities. While such scenarios may be possible under the right circumstances, they probably are less likely than localized threats. In any case, a wider range of simulations is required to capture the range of possible outcomes. Here there is a major need for training; a critical mass of competent scientific expertise in epidemiological modeling has not to date been adequately supported. Such efforts should become major responsibilities of NIH, CDC, and DOD.

Constructing models may be easier, however, than supplying them with meaningful data. There are gaps in our understanding of the factors that affect biological agents’ dispersal and uptake by humans, animals, and plants. For example, uncertainties of a factor of 10 or more in the LD50 values and a factor of 2 or more in the probit slopes (i.e., the dose-response curves) for different agents are common. These uncertainties are even greater if strain type is not known or the mechanism and magnitude of environmental decay rates for different agents are not well understood. Moreover, the incubation period (and its dose dependence) for different agents can vary by factors of 2 or more; and diurnal and weather variations can easily affect the contaminated area by an order of magnitude or more for open-air releases (typically the highest-casualty scenarios). Finally, uncertainties surrounding the amount and purity of the agent, the aerosolization efficiency for 1- to 5-micron particles, reaerosolization for agents that have settled onto the ground versus other surfaces, protection factors associated with buildings, and breathing rates can easily affect the inhaled dose by an order of magnitude or more.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement