For a number of biological terrorism agents, we found evidence that animals could provide early warning of an acute attack. For the agents for which we found evidence of sentinel potential, a key factor was the relative exposure risk of an animal compared to that of a nearby human population. However, in an actual event involving both humans and animals, the fact that disease was detected sooner in animals could be due to an interplay of a number of factors, including local infrastructure of animal and human health services, public awareness, and laboratory capacity. For other agents, however, humans would demonstrate symptoms at the same time as nearby animals or before. Therefore, the strength of evidence regarding animals serving as early indicators of an attack depends strongly on the agent and species involved. For some agents for which animals would not provide early warning, however, animals could help detect pockets of ongoing exposure risk. For the remainder of agents, evidence regarding the value of animals as sentinels is insufficient at this time.
Overall, according to our classification taxonomy, the strength of the recommendation that animals could provide early warning of an acute bioterrorism attack seems to be, at best, "fair" because of the inconsistency of the evidence. A somewhat more consistent level of evidence appeared to support the recommendation that animals could be markers for ongoing exposure risk and also that animals could play a strong role in propagating outbreaks caused by particular agents. At the same time, our ability to assess the overall strength of evidence for such recommendations was hampered by large gaps in current knowledge.
These findings suggest the need for certain steps related to preparedness for biological agent attacks. First, improved communication is needed between animal health and human health professionals, so that sentinel events could be rapidly detected. Such improvement would mean overcoming existing barriers to communication; a recent survey found that physicians and veterinarians communicate little about zoonotic issues (38
). Also, an adequate surveillance network should be developed to detect unusual health events in animal populations. Data on usual trends is missing for most animal species that could be potential sentinels. Whether public health resources can be committed to gathering such baseline data remains an open question.
Second, the results of this review indicate that active surveillance of animal populations, including wildlife and companion animals, could fill a critical need in the aftermath of an attack involving certain bioterrorism agents by helping identify persistent sources of infection in the environment. Third, better approaches for intervention are needed to be able to stem the propagation and amplification of an introduced biological warfare agent into a wild or domestic animal population. The US experience with West Nile virus reflects the difficulties of controlling an emerging zoonotic threat as it spreads through animal populations (39
Finally, the results of this review point out the need for additional research to fill knowledge gaps about animals as sentinels of human disease threats, including data on relative susceptibilities and exposure pathways for animal species living near human populations. Concrete steps could include establishment of surveillance veterinary clinics in strategic areas with incentives for practitioners to report unusual events. Another approach would be to make greater use of electronic databases of animal diseases such as those used by the Banfield Clinics, a nationwide chain of veterinary practices. Similar efforts could be useful with wildlife populations.
Such steps would foster ongoing communication between community practitioners and regional public and private veterinary diagnostic laboratories to establish baseline disease incidence trends and algorithms to identify outbreaks. Common links or web-based interfaces should be developed to integrate human and animal disease surveillance information. Reporting systems for wildlife professionals and the public should be created, and their use should be encouraged to document unusual disease events and die-offs. Another constructive step would be to improve the capacity of existing veterinary rapid-response teams, which exist in many states, to carry out active surveillance with animal populations as well as to improve the coordination of veterinary diagnostic laboratories. Again, barriers to funding and cooperation between animal and human health agencies need to be addressed. In the past, these have hampered efforts to have a coordinated approach to collection of animal surveillance data). In addition, state-based efforts would need to be coordinated on a regional and national scale. The growing awareness that animal health and human health are inextricably linked, however, makes cooperation between human and animal health professionals imperative to strengthen the evidence base that will allow for rational use of animal data in public health decision-making.