Scientific and Policy Tools for Countering Bioterrorism
Several scientific, scientific policy, and legal tools that were presented and discussed during the workshop but have not been addressed elsewhere in this report summary are included here. These include innovative surveillance; detection and diagnostic tools and technology; scientific policy issues unique to bioterrorism response preparedness; and, bioterrorism-related legal needs and obstacles. These ideas include multiple components of the individual sessions and have cross-cutting implications for an overall response to biological threats.
Surveillance and rapid detection are crucial to an effective response to a bioterrorist attack. Delayed detection results in delayed prophylaxis and aggressive treatment measures. Because it is practically impossible to predict when or where a bioterrorist attack is going to happen, there are limitations to the “dropin” terrorism surveillance systems that have been used to monitor specific places or events such as the Super Bowl or Democratic National Convention. Nor can bioterrorism surveillance be solved simply with pentium chips. Comprehensive bioterrorism surveillance will require integrating human resources, laboratory resources, and information management in innovative, legal, and acceptable ways that allow for early detection and characterization of threats.
There are several innovative surveillance systems in use or being developed. ESSENCE, for example, is an automated syndromic surveillance system that initially relied on the already extant automated health care information system across D.C., Maryland, and Virginia. Since September 11, ESSENCE has
been extended to over 300 installations around the world. Every 24 hours, 30,000 ambulatory diagnoses from these various installations are downloaded, automatically analyzed, prioritized based on expected values from historic data, and visualized using geographic information systems. However, none of the systems currently being developed are likely to be adequate in and of themselves. The best solution will probably be a system of systems that is sensitive enough to detect specific conditions and even small outbreaks.
Detection and Diagnosis
In terms of environmental and clinical detection and diagnosis, although reasonably good assays are available for a limited range of specific agents, the immense diversity of microorganisms, including bioengineered pathogens, presents a major challenge. There can be considerable variability even within a strain, let alone a species. Pathogens are a natural part of the environment and can confuse detection efforts. For both environmental and clinical settings, we need rapid, standardized methods that allow for the detection of a broad spectrum of potential biological weapons in a quantitative fashion.
Rapid detection and diagnosis requires access to an extensive sequence database and high throughput laboratories. Biotechnological barriers in the public health infrastructure must be identified so that the proper tools can be appropriately distributed or accessed. Academia, industry, and government laboratories must all be brought in at appropriate levels and in appropriate ways to help build new capabilities.
Specimen collection needs to standardized and automated. For example, there is no standardized collection method for samples from the inside of a computer. Indeed, specimen collection is often the major obstacle to rapidly processing a large number of samples and the weak link in what seems to be an otherwise very promising detection and diagnosis technology.
The capability to use molecular sequences to rapidly detect and identify bioterrorist agents could serve as an important form of deterrence and might possibly prevent bioterrorist attacks from occurring in the first place. One vision is an international molecular forensics lab that would rely on a molecular fingerprint global database to identify the source of the bioterrorist agent. This capability could provide the biological equivalent of the threat of nuclear retaliation. Again, it must be emphasized that bioterrorism is a national security issue and bioterrorism preparedness efforts are a strategic defense.
Scientific Policy Issues
The fact that bioterrorism preparedness is a national security imperative raises many important and new scientific policy issues:
It was suggested that we need a new peer review system for screening new bioterrorism defense research ideas.
There needs to be improved communication between scientists in the clinical response laboratories and law enforcement personnel, for example with regards to resolving crime scene versus public health needs and ensuring that the physicians or other individuals who have provided samples can receive results in a timely manner.
Law enforcement investigators need to be educated about relevant microbiological issues, and the scientific and public health communities need to be informed of how criminal investigations proceed.
There is concern about who should have access to certain scientific materials, equipment, and information and whether access to select agents should be restricted. At the same time, it is crucial that as much information as possible be in the hands of the biomedical community so that scientists can conduct the type of research that is necessary to build a strong biodefense arsenal.
Finally, computational modeling is an important but undervalued scientific component of bioterrorism defense preparedness. New computational capabilities can be used to model interactions between digital microbes (as opposed to actual, biological microbes) and digital immune systems. This kind of simulation approach could be used in guiding decisions about experimental design as well as in testing various policy and response scenarios.
Bioterrorism preparedness as a national security imperative also raises many important legal issues. The first step toward evaluating the necessity of a legal strategy for bioterrorism is to assess the adequacy of the existing legal infrastructure for dealing with bioterrorism issues. Do provisions in the law exist that enable authorities to do what needs to be done in the context of a bioterrorism event, for example decontaminating a building or quarantining individuals? Are there any legal obstacles that would interfere with a public health response to bioterrorism?
The primary legal authority for bioterrorism preparedness and response is at the state level. Recently, the Center for the Study of Law and the Public Health at the Georgetown Law Center and Johns Hopkins University prepared a model state emergency health powers act (see Appendix G) in an effort to facilitate the analysis of public health law at the state level. The proposal is being given to states for their consideration either for adoption or simply as a tool for review of their own public health statutes in the context of bioterrorism. The proposal has stimulated much controversy. Indeed, this controversy may be a reflection of the importance of the legal infrastructure for an effective public health response to bioterrorism.
INNOVATIVE SURVEILLANCE METHODS FOR MONITORING DANGEROUS PATHOGENS
Julie A. Pavlin, M.D., M.P.H.,1 Patrick Kelley, M.D., Dr.P.H.,1 Farzad Mostashari, M.D., M.S.P.H.,2 Mark G. Kortepeter, M.D., M.P.H.,3 Noreen A. Hynes, M.D., M.P.H.,4 Rashid A. Chotani, M.D., M.P.H.,5 Yves B. Mikol, Ph.D.,6 Margaret A. K. Ryan, M.D., M.P.H.,7 James S. Neville, M.D., M.P.H.,8 Donald T. Gantz, Ph.D.,9 James V. Writer, M.P.H.,1 Jared E. Florance, M.D., M.S.,10 Randall C. Culpepper, M.D., M.P.H.,3 Fred M. Henretig, M.D.11
Historically, emergent public health problems have been recognized by astute health care providers who then report their suspicions to public health authorities, rather than the reverse (Thacker, 1994; Thacker and Berkelman, 1998). Even with luck, this approach usually falls short of optimal public health care. Outbreak surveillance seeks to reduce reliance on the epidemiological insights of individual practitioners or facilities and to significantly decrease the time needed to collect and assess data, thereby allowing officials to be alerted more rapidly of an emerging threat.
A review of some recent disease outbreaks will help define the requirement for more timely, high-quality systems. Past epidemics have characteristics that can help identify the epidemiological, behavioral, and political factors that affect the detection of and response to an emerging infectious disease epidemic. Many emerging infections present as syndromes that initially do not point to a specific underlying pathogen. Potential bioterrorism events could also pose similar challenges of delayed recognition (see Table 6-1).
In addition to earlier detection of events, surveillance systems for emerging infections, including bioterrorism, are essential for focusing limited response assets and for providing evidence-based information for governmental risk communicators attempting to manage community concerns. Plans to improve
public health capabilities to identify and address such disease emergencies must include determining how surveillance systems can be made more timely, flexible, and sensitive without overly compromising other aspects of quality.
A recent meeting addressing innovative, responsive surveillance systems focused on three areas: 1) defining the existing functional capabilities in need of improvement, 2) examining existing prototype systems attempting to meet these needs, and 3) identifying the ideal features of a “system of surveillance systems” that would meet the need for more timely, sensitive, and flexible detection and response (Pavlin et al., 2001).
Examples of Recently Developed Innovative Surveillance Systems
In recent years, agencies and municipalities have attempted to improve public health capabilities with novel and innovative approaches to surveillance (see Table 6-2).
New York City—911 Calls
Beginning in March 1998, the New York City Department of Health, in collaboration with the Mayor’s Office of Emergency Management and the Fire Department’s Emergency Medical Services, began monitoring the chief clinical complaints noted in daily 911 calls as a citywide health indicator. The intent was timely detection of public health events, with particular emphasis on influenzalike illness. Several complaints, such as “difficulty breathing” and “sick”— thought to represent influenza-like illness—were selected. A review of data from 1991 to 1998 found a temporal association between the onset of annual influenza epidemics and a rise in the volume of the selected call-types. Thresholds were developed that “detected” all four annual influenza epidemics from 1994 to 1998. In addition, the system generated very few false-negative alarms—times when the expected threshold was exceeded outside of periods of peak influenza activity. This system indicated the 1999–2000 influenza epidemic approximately two weeks before recognition by traditional influenza surveillance systems.
New York City—Diarrheal Diseases Surveillance
New York City also developed three independent and complementary systems to monitor for community-wide gastrointestinal outbreaks: 1) sales of antidiarrheal medications, 2) submission of stool samples for laboratory tests, and 3) incidence of gastroenteritis in nursing home populations.
Monitoring sales of antidiarrheal medication approximates the incidence of diarrheal illness in a community. Large increases in sales of anti-diarrheal medicines have been reported during outbreaks of gastrointestinal diseases (Proctor et
al., 1998; Collin et al., 1981; Rodman et al., 1997; Miller and Mikol, 1999). Volume of over-the-counter medication sales is obtained from a regional distributor and a chain of drugstores (see Figure 6-1). Sales data are received weekly and analyzed for unexpected variation.
The number of stool specimens examined for 1) bacterial culture and sensitivity (three laboratories); 2) ova and parasites (three laboratories) (see Figure 6-2); and 3) Cryptosporidium parvum (one laboratory) is collected daily and indicates the incidence of gastrointestinal illness in the population. This information is provided in addition to test results for giardiasis and cryptosporidiosis from active disease surveillance.
The number of new cases of gastrointestinal disease in nursing homes provides information on the incidence of illness in a population with limited exposure to the wider community. Twelve nursing homes participate in the program, with 1,850 residents. Numbers of new cases of gastrointestinal disease are provided daily.
DoD-GEIS Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE)
The DoD-GEIS monitors patient data from medical facilities for changes in disease incidence in the National Capital Area (see Figure 6-3). For every outpatient encounter within the DoD medical system in the US, the provider electronically enters a code describing the reason for the patient visit. All encounters are coded at or near the time of service, even if the definitive cause of illness is not established during the visit. Most codes chosen by providers reflect this prompt diagnosis and may include syndrome-based codes such as cough and fever in addition to empiric diagnoses such as pneumonia.
All personal identifiers are removed from the data when received and the diagnoses are categorized, if applicable, into one of nine syndromic clusters. The frequency of outpatient visits in the different syndromic categories is then plotted and compared to previous years’ experience. It can also be depicted geographically through geographic information systems (GIS) software.
Sandia National Laboratory—Rapid Syndrome Validation Project (RSVP)
The RSVP is an Internet-based reporting system, intended for daily routine use by physicians and epidemiologists. The system features rapid data input of clinical and demographic information via a touch sensitive monitor, automated screening of reports for signs and symptoms correlated with reportable diseases and subsequent instantaneous notification of public health officials if indicated, and rapid feedback to clinicians of the geographic and temporal distribution of recent similar syndromes in their community in recent weeks. Public health offi-
cials may use RSVP to inform users of current disease outbreaks of public health importance, specific to each of six syndromes. RSVP is easily expandable to multiple sites, requires no specialized client software other than a web browser, and may be operated as an intranet or for general data sharing.
Expectations for a Health Indicator Surveillance System
The variety of information used to monitor the health of a community can be called “health indicator” surveillance. The requirements of different users should be documented while these new systems are still in development.
Surveillance systems can assist federal public health management of epidemics in many ways, not only in the traditional detection of outbreaks and monitoring of the effectiveness of preventive measures, but in determining how to assist and augment local health and emergency response activities. The systems should be capable of being integrated with other surveillance systems at all levels of government.
The ability of the local health department to identify, evaluate, and contain the effects of disease outbreaks is dependent on the timeliness and accuracy of reporting. Optimally, a surveillance system should contrast local data with data from other sources (e.g., CDC, nearby states, other jurisdictions, and military installations). Surveillance systems with greater sensitivity, completeness, and timeliness than existing systems are needed to allow the most effective identification of unexpected events across jurisdiction boundaries.
The military services maintain centralized health-related surveillance systems for routine public health policy and disease control. However, more rapid and sensitive recognition of a disease outbreak wherever U.S. forces are located requires enhancements to both the level of detail available and the speed of data transmission and analysis. New health indicator surveillance systems for military communities should cooperate with civilian public health personnel since disease outbreaks do not respect military installation boundaries.
Specific Needs for Bioterrorism Surveillance
To maximize patient survival after a bioterrorist attack, we need surveillance systems that: 1) facilitate the rapid recognition of a bioterrorist event, 2) assist in determining the site of exposure, 3) maximize efficient delivery of limited medical countermeasures to the infected population, and 4) assess containment and mitigation. Expansion and improvement of surveillance systems for bioterrorism will likely have a dual benefit of strengthening the public health infrastructure for detection of naturally occurring infectious disease outbreaks and emerging diseases.
Needs for Animal and Plant Surveillance
There are many points along the farm-to-table continuum at which infectious agents can arise from or be introduced into the food supply. The increasingly centralized nature of food production in the United States, with subsequent widespread distribution, means that the impact of contaminated products would be national (and potentially international) in scope rather than regional. This argues for an integrated human-animal pathogen and antimicrobial resistance surveillance system providing rapid feedback to public health and foodregulatory agencies. Food monitoring data for microbial pathogens and food recall information are already collected by federal and state regulatory agencies and could be integrated into existing food-borne disease and outbreak surveillance systems. This system could 1) enhance the speed with which outbreaks are identified and control measures implemented, 2) identify patterns of product contamination that would lead to more rapid intervention, and 3) identify unusual illness patterns and pathogens that are dispersed but possibly related.
The use of information on disease in animals, both wild and domestic, can prove a useful tool in monitoring the health of human populations (Jaroff, 1999; Steele et al., 2000; CDC, 1999). A surveillance system will include data on animal morbidity and mortality to achieve the greatest sensitivity.
Key Issues for Developing a Surveillance System
Health indicator surveillance is the foundation for early recognition of an emerging infectious disease. There is a critical need for a system of systems, with the flexibility to fit the needs of each level and locality. Measurable alterations in personal behaviors within the first hours or days of illness can assist early detection of an event or epidemic. These include work or school absenteeism, changes in usage of public transportation or toll roads, and the purchase of over-the-counter remedies. Data about delivery of medical care have value not only for outbreak detection but also for the ongoing management of an epidemic
(Rodman et al., 1998). These include emergency response calls, required disease reporting, outpatient clinic and emergency room activity, inpatient and intensive care unit records, and laboratory and prescription drug requests.
An infrastructure for pre-clinical or many types of clinical data is not readily available but might use data already collected for other purposes such as billing, inventory control, or resource management. Concerns over ownership may block access to existing data deemed valuable for surveillance. Resolving these issues will require high-level leadership, commitment, and prioritization.
The following questions addressed the usefulness of health surveillance data.
Are the data sufficiently representative of the entire population of interest? Are there important sub-populations that will not be captured by this surveillance system?
Are the data timely? How much time will elapse between the onset of symptoms and detectable change in the data?
Are the data available electronically? Electronic data can be transferred and analyzed more quickly than paper reports.
Can the data be categorized as symptom clusters or syndromes? Summary data (e.g., total number of admissions or transit ridership) may indicate an event is occurring, but interpretation of the cause will be difficult in the absence of more specific information.
Are retrospective data available? Without baseline data, it will be difficult to assess whether the data can detect new events. Furthermore, alarm thresholds using historic data obviate the need for a lengthy “run-in” period.
Improvement in Active Patient Data Collection
Although use of existing data sources can help, some situations, geographic areas, or types of medical practices may require additional data for an effective surveillance system. If a system requires new data collection, it is imperative to work closely with medical practitioners to achieve a workable solution and to provide feedback so they can benefit from their participation.
Possible features of an active data-entered, real-time surveillance system include:
Syndrome-based reporting from a pre-determined list of signs and symptoms;
Touch screen or personal digital assistant (PDA)-based electronic data reporting, collection, and submission;
Graphical presentation of data based on GIS and temporal information;
Automatic alerts to public health officials of specific signs and symptoms (e.g., fever with skin rash in young adults) suggestive of serious communicable disease; and,
Alerts from public health officials to health care providers that can be easily updated.
The Need for a System of Systems
Ideally, a surveillance system will be sensitive enough to identify the emergence of an outbreak, categorize its nature, and identify those affected so that the outbreak can be quickly and effectively contained. Bringing together information from various health indicator data sets can allow the public health practitioner to 1) evaluate many indicators simultaneously, 2) compare variations and identify common trends, and 3) track confounding factors and reduce noise.
The compilation of information provided by independent and complementary data sources allows inter-system comparisons. Comparing the data from several indicators, some of which are more sensitive than others for different scenarios, can enable a trend observed in any single system to be confirmed by the systems. Simultaneous unexpected but concordant variations in multiple data sets may suggest actual emergence of an outbreak. Clinical reports are needed for confirmation.
The importance of collecting data through an intricate surveillance system is to use it to quickly identify and respond to an adverse event rather than to develop an archive. Ideally one organization would collect, compile, integrate, and analyze all data. Moreover, this data would need to be shared effectively and efficiently at different levels of the existing health systems. Of utmost importance, the fundamental issue of personal and organizational privacy needs to be addressed when setting up such a system.
TABLE 6-1 Selected infectious disease outbreaks characterized by delayed recognition, characterization, or response
Disease Outbreak Characteristics
Rapid spread over large geographic area
Overwhelms health care system
Overwhelms essential services (e.g., burial of dead)
No available treatment
Legionellosis (Legionnaires’ Disease)
Exposed population disperses from point of exposure throughout the state of Pennsylvania
Rapid spread and demise
Mimics a biological terrorist attack
Acute Respiratory Distress Syndrome
Affects small population spread over a large geographic area
Oregon, 1984, US, 19937
Bioterrorism attack that mimics naturallyoccurring outbreak
Unrecognized as bioterrorism at time
Encephalitis (West Nile virus)
New York City, 19998,9
Zoonotic (birds are first victims)
Initial diagnosis wrong
Limited geographic area affected (humans)
Specific population group affected (elderly humans)
New agent to New York City
Suspicion of bioterrorism
TABLE 6-2 Possible sources of health indicator surveillance data
Cons and Confounders
Outpatient and ER visits
Reflects incidence of disease in general population
Nonspecific- May be difficult to document definitive information
Best indicator of rare events like West Nile virus or Hantavirus pulmonary syndrome
Will not capture milder cases
OTC pharmacy sales
Reflects symptomatology most broadly
Subject to promotions/sales
Clinical lab submissions
Ordered by clinicians
May not be ordered for all (most) patients
Medicare or Medicaid claims
Ease of data capture
Problems with timeliness and accuracy
Not broadly representative
Reported by medical personnel
Immobile population with limited exposure possibilities
Immobility reduces exposure potential
Not broadly representative
Systematic testing for specific disease agents of specimens submitted to public health lab
Specificity of diagnoses
Broad screening not likely to capture meaningful data
Difficulty getting information on positive samples
School and work absenteeism
May occur earlier than visits to clinician
May be absent for nonmedical reasons
Delays in obtaining data
Ambulance call chief complaints
Many communities with timely access to data
Poison info calls
Ability to access real-time
May not be related to infectious diseases
HMO/Nurse hotline calls
Occur very early in disease outbreak
May be difficult to categorize
THE USE OF COMPUTATIONAL MODELING IN RESPONDING TO BIOLOGICAL THREATS
Donald S. Burke, M.D.
Professor, Department of International Health
Director, Center for Immunization Research, Johns Hopkins
Bloomberg School of Public Health
Computational modeling and simulation is an important but undervalued component of the scientific strategy to defend against emerging infectious diseases and bioterrorism. The extraordinary sustained growth in computational power has already given rise to ambitious modeling efforts in a wide variety of other scientific disciplines including nuclear physics, astronomy, economics, and other fields. My own experience in collaborations with computer scientists at the Navy Center for Applied Research in Artificial Intelligence has lead me to realize that computational approaches can productively be used to analyze and conceptualize the behavior of epidemic microbes. I suggest that we are fast approaching an era in which we will use computational modeling and simulation to guide public health policy decisions with regards to emerging infectious diseases and bioterrorism. We may be able to predict and prevent, rather than just respond to, epidemic infectious diseases.
A variety of computational approaches can be used. Using a “top/down” approach analytic techniques borrowed from physics can be used to study epidemiologic data. Fourier transforms and wavelet decomposition—standard tools in electrical engineering—can be used to decompose a temporal or spatiotemporal epidemiological data set into its various components. Decomposition in effect reduces noise and permits careful analysis of individual temporal and spatial components. Such techniques will probably prove useful in understanding environmental forcing of epidemics, such as the cyclical influence of weather and climate on disease transmission. Evolutionary and genetic algorithms, used routinely by computer scientists, can be used to model rapidly evolving organisms such as RNA viruses. All RNA viruses (HIV, influenza, Ebola) have very small genomes (total length approximately 10,000 nucleotides), so that full length genomic sequencing of viral isolates has become routine. We can now study in complete detail, reconstruct—and using evolutionary algorithms, simulate—the evolutionary patterns of these and other future threats. Analysis of the evolutionary tactics used by small “digital organisms” may permit discovery of yet unknown rules that govern real-world patterns of viral evolution. Another powerful computational approach comes from the field of socio-economics, known as “agent-based” modeling. This involves computational creation of populations of digital entities (“agents”), each of which can be thought of as a person. Sets of rules are written that govern the behaviors of each
agent, either individually or as part of a group, such as movement patterns, likes and dislikes, lifespan, or even resistance and susceptibility to disease. Agents move about in two dimensional space and interact with other agents. Such models can be used to simulate and study the efficacy of public health interventions such as targeted immunization or quarantine measures. Finally, combining agent-based modeling with genetic algorithms may provide a novel strategy to understand—and some day predict—how microbes evolve and emerge as epidemic diseases.
My view is that we who work in the field of epidemiology are far behind other scientific disciplines in exploitation of computational techniques. Some of the current large computational modeling and simulation efforts in other scientific fields include:
The Department of Transportation, in collaboration with the Department of Energy, is using agent-based modeling to model automobile traffic flows in major cities.
The Virgo Consortium has used supercomputers to simulate the structure of the universe. In the simulation, each particle represents a galaxy. One can easily imagine that instead of a galaxy, each particle represents an RNA virus sequence navigating (evolving) through sequence space.
Nuclear weapons are simulated at the Los Alamos National Laboratory’s Strategic Computing Complex. At the complex, there are more than 300 nuclear weapons scientists, modelers, and designers working in a 300,000 square foot facility. Their simulations have a 30 trillion per second calculation capability, which results in 30 million pixel simulation displays. Such computational power could permit simulation of infectious disease epidemics with realistic features.
Computational modeling and simulation should be an important new area for initiatives in our efforts to confront emerging infectious diseases and bioterrorism. Clearly there are some excellent opportunities for developing some very creative new approaches. Arguably a new generation of models and simulations may prove to be the only way that we can predict—and hope to prevent—the emergence and epidemic spread of future infectious disease threats. At the very least, this type of modeling and simulation should have heuristic value—it should be extremely useful in teaching about infectious disease epidemiology, as well as in guiding the design and conduct of field research.
DIAGNOSTICS AND DETECTION METHODS: IMPROVING RAPID RESPONSE CAPABILITIES
David A. Relman, M.D.
Departments of Microbiology & Immunology, and Medicine Stanford University
Ideally, one would like to be able to detect a bioterrorist attack early on, after an agent has been released but before there is any clinical evidence of pathology or overt damage. At this time point, there are opportunities for environmental detection, as well as opportunities for diagnosis of the exposed and infected but pre-symptomatic host. Intervention during this early preclinical incubation phase provides the greatest opportunity for benefit. However, it also poses the greatest challenges: current detection and diagnostic methods are not very useful at this point. It is usually not until a person becomes acutely or severely ill that available methods readily identify and characterize the nature of the attack. Thus, there are several issues related to diagnostics and detection that must be addressed in order to improve rapid response capabilities.
What needs to be done to improve rapid diagnostics and detection?
There are two aspects of diagnostics and detection—environmental and clinical—but their needs are very similar. Environmental detection requires rapid, semi-automated methods that enable detection of a broad spectrum of potential biological agents. These methods need to be standardized, rigorous, reproducible, and based on a thorough understanding of the natural background. Natural background refers to an agent’s genetic variability, antigenic variability, geographic distribution, and how each of these changes over time. It also refers to non-biological aspects of the environment that could complicate and obfuscate detection efforts.
In the clinical setting, diagnostic methods should be rapid enough to reveal causative agents early in the course of disease. They should provide quantitative results, because sometimes the only difference between health and disease with respect to a particular agent is its relative abundance. These methods should be standardized and available at the point of care. As with environmental detection, accurate clinical diagnosis is based on a sound understanding of the natural background of the agent, including related agents, of the normal flora, and of non- or abiotic parts of clinical specimens that could potentially complicate these approaches.
What is the current status of available diagnostic and detection methods?
For environmental detection, there are reasonably good assays for a limited range of specific agents. These assays are based on cultivation, immunologic detection, and nucleic acid-based detection (e.g., PCR). However, current environmental detection systems are generally only capable of intermittent monitoring, plus they require an unacceptably high degree of hands-on maintenance. They are, by and large, designed for idealized conditions and settings that are very simple and free of the extraneous variables that characterize true natural backgrounds. Available information on natural backgrounds is spotty at best. Current protocols are nonstandardized or poorly standardized.
For clinical diagnosis, specific assays are available for most high priority agents. They are based on the same technologies as the environmental detection assays and work reasonably well in fairly idealized conditions. Their use is based on clinical suspicion; they are not generally used for automated implementation based upon broader or less specific information. As with available environmental detection methods, the collection protocols are largely nonstandardized or poorly standardized. And again, they are based on only spotty information about the natural background. In many cases they are too slow to be clinically useful.
Examples of current enabling procedures and technologies
There are some reasonably good collectors available that can sample large volumes of air over short periods of time by concentrating and impacting air content onto either a water filter or solid substrate; the collected material is then introduced into a detection scheme. The swab is another common collection device.
As has been learned from recent anthrax surveillance, swabbing and collecting in a standardized fashion from environmental surfaces and from the internal features of complicated three-dimensional objects is an extremely difficult problem. The most successful currently available collection devices are best suited for sampling air, which one would expect to be one of the simplest entities in the natural world to sample. But air is actually very complex. For example, black rubber particles in the air that come from car tires are a major PCR inhibitor. Still, air sampling methods are better than what is currently available for collecting and concentrating target in water, food, soil, and other materials.
Detection procedures still rely heavily on cultivation. When available, cultivated organisms allow for a more comprehensive evaluation of important phenotypes, than does molecular information generated by newer technologies. Although some phenotypes are not reliable for identification purposes, the availability of a pure culture facilitates acquisition of more reliable molecular information. Thus we should continue to consider ways of optimizing and en-
hancing our ability to recover organisms in culture in addition to other options that we choose to pursue. In theory, this method should allow for the detection of a single organism, but the process is slow. And such a high sensitivity can be problematic if the organism is part of the biological background.
There are many available immunology-based detection platforms, a number of which have received considerable attention in the recent press. Although the so-called “Smart Tickets” offer some true utility, they have major limitations. For example, they rely on a few standard, well-used, and well-tested antibodies. This dependency creates a potential critical vulnerability for detection and diagnostic methods in terms of specificity and reliability. Naturally-occurring epitope variability is only the first problem, as we begin to face the possibility of engineered organisms with deliberately modified epitopes. There are only a few developers venturing beyond a small set of well-characterized antibodies in their explorations of new immunologic platforms. Also, immunology-based assays typically have fairly high lower limits of detection, so sensitivity is another potential problem. The analysis time, however, is reasonably good.
There are a large number of nucleic acid detection technologies that work reasonably well in idealized settings. As with the immunological detection platforms, the fact that these assays generally rely on the sequences of only a few typed strains is a potential vulnerability. The sensitivity of nucleic acid-based assays, however, is much greater than that of immunology-based methods. The level of detection can be as low as a few sequence copies, bacteria, or viral particles.
What are some of the stumbling blocks and problems in the quest for better diagnostics and detection methods?
The diversity of potential bioterrorist agents present a major challenge. This includes not only all of the naturally occurring pathogens scattered throughout the bacterial, viral, and eukaryotic worlds, but also all of the bioengineered, chimeric organisms that may not even yet exist. The immense variability within strains, let alone species, as well as varying degrees of relative abundance or evenness in nature make it very difficult to distinguish causal agents from other agents in the environment or host. With increasing use of the more sensitive PCR-based technologies, it is likely that the natural, variable biological background will show up more clearly. Additionally, the likelihood that even healthy individuals have a quantifiable amount of bacterial and other microbial DNA or RNA circulating through their blood and other sequestered anatomic compartments further complicates the issue.
The causal nature of the association between detected agent and disease in an experimental subject is not easily established; nor are the methods and criteria needed for such an effort well-defined. Likewise, the correlation
between environmental detection of an agent and the risk of exposure and disease in a nearby host is a difficult proposition.
The field is cluttered with unsubstantiated claims and ill-defined validation standards. Tests and methods are often validated for analytical, but not clinical performance characteristics. There are a number of eager vendors on the market who are selling detection and diagnostic kits that have not been well-validated.
There are few standardized collection methods for complex, real-world specimens, such as the insides of computers.
Laboratory surge capacity is inadequate, specimen analysis throughput is low, and turnaround times are slow.
Delivery and implementation of state-of-the-art technologies is poor. For example, although rapid, real-time PCR offers major benefits, it is not available at the point of care and in places where validation needs to occur, such as in the environmental or clinical workplace.
Because of limited sensitivity and inadequate attention to optimized specimen selection and processing, diagnostic methods are limited to the late stages of disease.
What can be done, and where can we be in five years?
Near-term goals include the following:
We need a library of high affinity binding reagents for detection of a wide spectrum of biological agents, their variants and components. These reagents should include not just antibodies but also other high affinity ligands such as aptamers and peptide nucleic acids.
We need an extensive sequence database effort. This is already underway, but we still need to consider its breadth, depth, and what kind of information we need to mine. For example, we still need broad range oligonucleotide sequences as primers or probes for a number of families of pathogens, viral and otherwise, for which we do not have good reagents currently.
We need high throughput laboratories with much greater surge capacity. These laboratories might also be involved in methods and technology development. These could be dual purpose labs that are also used for other routine diagnostic purposes, e.g., influenza virus typing.
We need to focus more effort on standardization and automation of specimen collection and processing in order to analyze more efficiently and accurately large numbers of samples. Devices that are currently being developed need to be validated with real world problems.
We need to consider how to best apply some of the new biotechnologies, such as nanotechnology, microfluidics, and microarrays. Microarrays
in particular will likely have a major impact on strain and host typing; for example, a number of organism-specific whole genome microarrays have illustrated the power of genome-wide hybridization profiles for strain classification and potentially forensic strain typing. Recent developments in genomics suggest that we may now have the capability to perform genomics on a single bacterial cell, such that it will no longer be necessary to cultivate an organism from a clinical specimen or from the environment. DNA microarrays will also facilitate the identification and detection of diagnostic and prognostic host signatures. Expression analysis using DNA microarrays may prove valuable in diagnosing clinical disease and classifying infectious agents by comparing an individual’s gene expression pattern to known pathogen associated patterns. But there are many unresolved issues, such as knowing which cells serve as the best source for signature data, how well and on what basis the host discriminates between different pathogens, and what role host specificity plays. In general, one hopes and expects that variability between healthy individuals is limited in comparison to the differences between healthy and diseased hosts, and that it will not obscure the recognition of a common infectious agent. But, the sources and nature of human host-specific “intrinsic-ness” must be defined.
We need to consider how certain physics applications, such as hyperspectral analysis, may be relevant to infectious disease detection and diagnosis.
NATIONAL SECURITY AND OPENNESS OF SCIENTIFIC RESEARCH
Ronald M. Atlas, Ph.D.
Professor of Biology, University of Louisville
President Elect—American Society for Microbiology
The post September 11 anthrax attack has forced the recognition that bioterrorism is a reality of the 21st century. Infectious diseases pose grave threats to U.S. national security. Our response to these threats requires new and very focused research efforts along with enhancements to the public health infrastructure. Investments aimed at protecting against bioterrorism are best harmonized with the overall efforts to combat infectious diseases. Outbreaks of infectious diseases, whether naturally occurring or intentionally initiated, can represent threats to national and global security. Sustainable efforts that diminish the threat of bioterrorism and lessen the natural occurrences of infectious diseases (a dual function approach) will have the greatest long-term benefits.
Efforts to enhance detection and surveillance systems will be particularly valuable in improving our daily health, and that of our children, while at the same time protecting against acts of bioterrorism. Similarly, the discovery and
development of new vaccines and antimicrobial drugs will be especially valuable in combating infectious diseases, especially with the continued emergence of antibiotic resistance, while helping protect against the potential use of antibiotic-resistant and genetically engineered agents by bioterrorists. Both antibiotic resistance and immune modulation must be considered given that a sophisticated bioterrorist could employ modern molecular approaches to design especially deadly biothreat agents. Broad spectrum antibacterials and antivirals could offer protection against a wide variety of infectious agents and could offer major protection against bioterrorism—this would offer broad protection against unknown biothreat agents while specific narrower spectrum drugs would be most appropriate when the exact nature of an agent had been determined.
Ensuring the vigor of research and development efforts to combat infectious diseases, with an appropriate focus on biothreats, will require an influx of new investments and the strategic redirection of some ongoing efforts through reallocation. If properly directed, the investments in bioterrorism response will be sustainable and will help diminish the overall threat to national and global security posed by infectious diseases.
As we move forward we need to ensure the health of the biomedical research enterprise. While protecting against inappropriate use, we need to ensure that legitimate scientists can access the materials, equipment, and information to move this biomedical agenda forward. We cannot let terrorists undermine our efforts to find new vaccines, drugs, and diagnostics in the battle against infectious disease.
Never before has the biomedical community faced greater challenges about protecting public health from the spread of infectious agents while facing increased scrutiny about the misuse of science by terrorists. Are new mechanisms needed to govern scientific research so as to lessen the probability of the development of advanced biological weapons? If so, what should be done? The National Academy of Sciences and the American Society for Microbiology (ASM) must assume leadership roles in fostering a public discourse that will ensure the advancement of science for the betterment of humankind and enhanced protection against bioterrorism. We must ensure that the biomedical community is assisted and not deterred from finding the diagnostic tools, vaccines, and medicinals needed to combat bioterrorism.
The scientific, biomedical, and public health communities must work with law enforcement to combat bioterrorism. But as Gerald Epstein says in his article entitled Controlling Biological Warfare Threats: Resolving Potential Tensions Among the Research Community, Industry, and the National Security Community, which will appear in the December issue of Critical Reviews in Microbiology, the research and national security communities have different objectives, cultures, and norms, and are likely to weigh the costs and benefits of proposed policy measures differently. It is not surprising, therefore, that these differences would sharpen and that there would be a greater need to seek resolution following
the events of September 11 and the ensuing anthrax attacks. Epstein, who had worked in the Office of Science and Technology during the Clinton Administration raises the following issues, which I will discuss individually:
Tightening restrictions on access to dangerous pathogens;
Restricting access and dissemination of “relevant information,” i.e., classifying research reports and censoring research publications; and,
Imposing restrictions on the conduct of “contentious research,” i.e., limiting fundamental biological or biomedical investigations that produce organisms or knowledge that could have immediate weapons implications.
How can we better control access to potential biothreat agents?
Are there individuals that should not be permitted to conduct certain categories of research, or that should not be given access to dangerous pathogens?
What should be done about physical security at institutions that maintain cultures of potentially dangerous biological agents to help prevent unauthorized individuals from obtaining such agents?
Are locks enough or should armed guards be required to secure laboratories possessing select agents?
ASM has supported imposing reasonable restrictions on access to select agents that pose high risks as potential biological weapons. It has supported legislation and regulation that control the exchange of certain dangerous pathogens including the CDC Laboratory Registration/Select Agent Transfer Program
These regulations, which place shipping and handling requirements on laboratory facilities that transfer or receive select agents capable of causing substantial harm to human health, are designed to ensure that select agents are not shipped to parties who are not equipped to handle them appropriately or who lack proper authorization for their requests. Under the regulations which have been in effect since April 15, 1997, the CDC regulates the shipment of 36 select agents. The list of agents was developed in consultation with the ASM. It includes agents that are especially dangerous, such as the agents that cause smallpox, anthrax, plague, and other deadly diseases. The regulations require adherence to CDC biosafety manual that includes various biosecurity measures. Thus, it begins to limit access. The ASM publicized Select Agent shipping regulations to the scientific community and has repeatedly exhorted microbiologists to adhere to all the regulatory requirements it imposes. But lest there be any excessive sense of security, it must be realized that these regulations apply only to U.S. facilities and that with the exception of smallpox, all the other select agents occur in nature.
ASM also supported the USA Patriot Act which was signed into law on October 26, 2001. This Act imposes restrictions on who may possess select agents, specifically restricting possession of select agents for aliens from coun-
tries designated as supporting terrorism and from individuals who are not permitted to purchase handguns, including some individuals with a history of mental illness or a criminal record. ASM felt that these were reasonable protective measures that would not have a significant adverse impact on biomedical research and that might increase national security by making it more difficult to obtain select agents by individuals who might misuse them. Although ASM sought provisions for exemptions Congress decided otherwise. Thus, the law contains no provision for exemptions under any circumstances
While prohibiting the possession of select agents for purposes that are not for bona fide research and other beneficial purposes, the USA Patriot act does not impose registration requirements for the possession of select agents for legitimate purposes ASM has supported such registration since 1999. On December 4, 2001, the Senate Appropriations Committee approved HR 3338, the DoD Appropriations Bill for FY 2002, which Senator Diane Feinstein (D-CA) and Senator Judd Gregg (R-NH) amended to include Section 8134 Regulation of Biological Agents and Toxins. The amendment they introduced was the same as Section 216 of S l765, the Bioterrorism Preparedness Act of 2001, which was reintroduced by Senator Frist and Senator Kennedy on December 4, 2001. It also is similar to a House resolution introduced by Representative Tauzin. These Bills restate the CDC select agent transfer rules and require safeguards to prevent access to such agents and toxins for use in domestic or international terrorism or for any other criminal purpose. They mandate biennial updating of the select agent list which seems important given the pace of science. They mandate the imposition of regulations and standards for possession of select agents that ensure exclusion of individuals restricted by the USA Patriot Act and require registration for anyone possessing a select agent. Background checks would have to be conducted and steps might also be mandated to ensure that law enforcement could prevent suspected terrorists from gaining access to select agents. Appropriate security requirements for persons possessing, using, or transferring biological agents and toxins would be imposed and information would have to be provided, if available, that would facilitate the traceability of select agents if those agents were ever misused.
Beyond the laws and regulations that limit access to select agents, lies the question of blocking the dissemination of select information that might be useful to bioterrorists, what Epstein terms opacity and what others might call secrecy and censorship. Should more research be declared classified? Should there be criteria that would warrant restrictions on publication or other dissemination of research results? Should we stop revealing genomes? Should some aspects of research be withheld from publication, e.g., methods or selective results? Should there be review boards to consider the national security implications of all publications?
ASM has had to grapple with many of these questions, both before and after September 11. For example, among the information posted at ASM’s website in an effort to provide relevant information were the abstracts of the 4th Interna-
tional Conference that was organized by scientists from the U.S. Army Medical Research Institute, the British Defense Research Agency, NIH, and the Pasteur Institute. One of the abstracts, In Vitro Selection and Characterization of High-Level Fluoroquinolone Resistance in Bacillus anthracis. By L. Price, A. G. Vogler, S. James, and P. Keim of Northern Arizona State University, described a study that showed increasing exposure to ciprofloxacin resulted in the evolution of fluoroquinolone resistance in Bacillus anthracis. This meant that antibiotic resistant Bacillus anthracis could be intentionally produced. Did it represent a roadmap for a bioterrorist? It also meant multiple antibiotic treatment was warranted in cases of inhalational anthrax. Thus did it present information useful to the biomedical community. Should the abstract have been published? Should it have been removed after September 11?
After some discussion we decided to leave this and other information about bioterrorism and anthrax on the ASM website for the education of the scientific community. My view was in favor of the benefits accrued from openness in science—if someone wished to publish legitimate research I did not want ASM to act as the censor. This position in favor of openness of science drew some concern as reported by Eric Lichtblau in his article Response to Terror: Rising Fears That What We Do Know Can Hurt Us that appeared in the Los Angeles Times on November 18, 2001. The article quoted University of Pennsylvania ethicist Arthur Caplan as saying, “We have to get away from the ethos that knowledge is good, knowledge should be publicly available, that information will liberate us...Information will kill us in the techno-terrorist age, and I think it’s nuts to put that stuff on Websites.”
The question of openness of science was considered by ASM years ago when it considered whether the smallpox virus genome should be classified or whether it should be published in the open scientific literature. ASM supported open publication and considered that the sequence data would be especially valuable to understand the virulence of smallpox virus and to provide targets for potential therapeutic drug design. Indeed analysis of the smallpox viral genome has revealed the basis for its virulence including the basis for immunomodulation. The genome analysis revealed targets for vaccine, drug, and detection development, information that seems to be of far more biomedical value than as an aid to bioterrorists.
The same sort of questions have been raised about the genome of Bacillus anthracis. In this case, the nucleotide sequence of plasmid pXO1 was published showing that it contains a “pathogenicity island,” with the three toxin genes (cya, lef, and pagA), regulatory elements controlling the toxin genes, three germination response genes, 19 additional ORFs and 3 sequences that may encode enzymes responsible for the synthesis of a polysaccharide capsule usually associated with serotype-specific virulent streptococci. The conclusion was that major virulence elements of Bacillus anthracis are plasmid encoded. Despite the fact that the critical pathogenicity data was already published the question was
raised about publishing the full genome which recently was completed. In fact some have questioned whether the genomic information already published should somehow be expunged from the open literature? So far the decision has been to continue to release genomic data although there continue to be expressions of concern. Given that the genomic data is viewed as relevant for the identification of targets for therapeutic drugs and vaccines, then it also can be viewed as relevant for identifying targets for increased virulence and the avoidance of current therapies, vaccines, and detection protocols. There is no doubt that there is a duality of potential good and evil encoded within genomes.
The same questions can be raised about other scientific findings, for example the recent demonstration by reverse genetics that a single mutation at position 627 in the PB2 protein of an H5N1 influenza A virus influenced the outcome of infection in mice, i.e., one mutation can greatly increase virulence. Thus, it may be much more simple to create more virulent biothreat agents than previously thought again raising the question of whether such information should have been revealed? Should journals censor such information in such articles? Should meetings remove abstracts and presentations of such information? Should sponsoring agencies require advance review of any presentations and publications so as to restrict release of such information? What would this do to the future of biomedical research in the United States, especially if we were the only country to begin to restrict the communication of scientific results that are of clear biomedical importance.
In an effort to respond to these questions I asked the editors of the 11 journals published by ASM to consider under what conditions they would consider rejecting a paper based upon ethical and national security issues. They responded that they did not want to publish papers that violated the ASM code of ethics nor those that violated other guidelines, including the NIH guidelines for recombinant organisms. They also were very sensitive to the national security implications but were not prepared to restrict the flow of legitimate scientific communications that were clearly aimed at our understanding of microbiology and that held potential for advancing biomedical science. After due deliberation they drafted for me the following statement: “The ASM recognizes that there are valid concerns regarding the publication of information in scientific journals that could be put to inappropriate use. The ASM hopes to participate in the public debate on these issues. Until a national consensus is reached, the rare manuscript that might raise such issues will be reviewed by the ASM Publications Board prior to the Society proceeding to publication.” This statement with an accompanying introduction will be sent to all Editors of all ASM journals in order that they be alerted as to their responsibilities in this matter. Thus, the editors of ASM journals are trying to be responsible stewards of scientific information and communication by carefully balancing national security with the value of advancing science for the benefit of humanity.
Besides questions about communication of scientific information Gerald Epstein also discusses the possibility of constraining research, i.e., restricting researchers from conducting certain types of research. Are there areas of research or types of experimentation that should not be conducted at all? Are there others that should require advance approval? Is molecular biology a threat—will recombinant DNA technology be used to create horrific biothreat agents? Should certain molecular biology experiments and methodologies be prohibited?
Much of this concern emanates from experiments in which IL-4 genes were inserted into mousepox viruses. The result was suppression of the immune response to a much greater extent than anyone had predicted. Virus-encoded IL-4 not only suppresses primary antiviral cell-mediated immune responses but also can inhibit the expression of immune memory responses. A poxvirus can be simply genetically engineered for which immunization will be totally ineffective. The implications for possible genetic engineering of a horrific strain of smallpox virus are enormous. In hindsight some have asked whether this research should have been permitted? Shouldn’t we have known in advance how dangerous the results might have been. Others who clearly were surprised by the results feel that this study alerts us to the need for more research on the immune response and antiviral drugs.
Yet other concerns have been raised about DNA shuffling because of its potential power to create new biothreat agents. Some point to the fact that this methodology potentially allows the rapid production of numerous biothreat agents with enhanced virulence. They raise the fear that DNA shuffling increases threat of being able to create a deadly new pathogen—intentionally or accidentally. They ask whether this methodology is too powerful and hence whether we should prohibit its use. But the potential benefits regarding new drug discoveries seem to outweigh these risks. In my view, the scientific community must move forward as quickly as possible in eliminating the threat of bioterrorism by finding effective preventative measures and cures so that infectious diseases are not a credible threat to humanity.
Beyond the obvious need to further biomedical research and to strengthen the public health infrastructure, one can ask about the appropriate role of the scientific community in identifying misconduct. What obligation do members of the research community have to identify, call attention to, or clarify activities of others that may appear suspicious?
There may be areas of research or types of experiments that pose such sensitivity regarding potential bioweapons application that merit extraordinary obligations for transparency and openness. There are clear aspects of bioethics that require scientists to be whistle blowers when public health is threatened.
Concerning the area of bioethics, the Council Policy Committee of ASM passed a resolution following September 11 affirming the longstanding position of the Society that microbiologists will work for the proper and beneficent application of science and will call to the attention of the public or the appropriate
authorities misuses of microbiology or of information derived from microbiology. ASM members are obligated to discourage any use of microbiology contrary to the welfare of humankind, including the use of microbes as biological weapons. Bioterrorism violates the fundamental principles expressed in the Code of Ethics of the Society and is abhorrent to ASM and its members.
In conclusion I want to share some thoughts from Abigail Salyers, the current President of ASM: “Terrorism feeds on fear, and fear feeds on ignorance. The best defense against anthrax or any other infectious disease is information— information in a form that can be used by scientists and by members of the public to guide rational and effective actions to ensure public safety. Placing major new barriers in the path of the flow of information between scientists and between scientists and the public more likely may ultimately contribute to terrorism by interfering with our ability to prepare and to respond to the threat of the misuse of science by bioterrorists.”
COORDINATING THE INTELLIGENCE, PUBLIC HEALTH, AND RESEARCH COMMUNITIES
Federal Bureau of Investigation
An intentional biological terrorism event requires a law enforcement response. Regardless of whether it was for political, social, or other reasons, the responsible individuals inflicted terror, committed an act of terrorism, and need to be aggressively pursued, investigated and prosecuted. Otherwise, there may be a repeat incident. Thus the role of the FBI in the bioterrorism arena.
The FBI’s capability to apprehend bioterrorists is based on effective laws, federal statutes, and the ability to enforce these laws. Recent legal initiatives include an expansion of the Biological Weapons Anti-Terrorism act, which now applies to the possession of a biological agent that is beyond reasonable means for peaceful prophylactic protective or bona fide research. And, as of November 1, 2001, sentencing guidelines became effective such that anyone who does violate the WMD statute enters into a matrix to determine the sentence received. Prior to that, sentencing was at the discretion of the judge. The new guidelines were established in hope that more structured sentencing would serve as a stronger deterrence factor. During recent events, it has become clear that more consideration must be given to educating prosecuting attorneys and investigators in microbiology. It is also very important that we utilize available resources, including experts within the scientific community.
Equally important is educating the public health community, including public health and state epidemiologists, in how the FBI conducts their investiga-
tions. For example, it is important that the FBI collects environmental swabs and maintains a strict chain of custody in order to ensure that that evidence is in the same or similar condition at the time of trial.
The recent anthrax case has illustrated both a covert and an overt release which the FBI is still investigating. The D.C. incident is an example of an incident in which there existed a known crime scene so the FBI knew where to go and respond. The D.C. incident grew exponentially and also spread to the post offices and the Senate Building. Both types of releases have required the FBI to work very closely with public health and have illustrated the importance of communication, coordination, and the sharing of information—including intelligence information—between the FBI and public health.
Finally, recent events have stressed the need to minimize the overlap between federal, state and local agencies; and the need to set aside personal or agency agendas in order to work together to protect the public and hopefully prevent repeat incidents.
VIRTUALLY ASSURED DETECTION AND RESPONSE: UTILIZING SCIENCE, TECHNOLOGY, AND POLICY AGAINST BIOTERRORISM
Scott P. Layne, M.D., Ph.D.
Associate Professor, School of Public Health
University of California, Los Angeles
Homeland Security and the Biological Weapons Convention
The United States must control bioweapons threats on two major fronts. Domestically, it must seek new ways to boost homeland security and respond to terrorists attacks in several American cities. Internationally, it must seek new ways to overhaul the long stalemated Biological Weapons Convention (BWC) Protocol or propose an alternative way to establish legally binding verification and compliance procedures. The challenges are enormous and demand rapid, reliable, and complete information on which to make decisions.
The development of bioweapons requires three key elements: knowledge, equipment, and infectious agents. These elements have “dual uses” and thereby pose serious challenges to verification, compliance, and security. The general scientific knowledge required to develop bioweapons is conveyed in many microbiologic texts and is not feasible to remove. The United States seeks measures that thwart the migration of technical expertise and first-hand knowledge from past and present bioweapons programs. Likewise, the small-scale laboratory equipment required to create bioweapons is all but impossible to restrict. The United States supports regulations that block the export of industrial-scale
laboratory equipment to potential proliferators. With the exception of variola major (smallpox), the various infectious agents required to create bioweapons are found in nature. The United States seeks regulations that constrains the sale of weaponizable seed stocks to qualified researchers and institutions.
Yet the United States is only one of many countries that supply such knowledge, equipment, and infectious agents. For example, Bacillus anthracis is the subject of research in many countries and conventional forensic methods may not be able to identify the source of B. anthracis used in any particular biological weapon. However, science and technology have opened up an extremely powerful means to address this problem. Infectious disease agents from specific origins exhibit unique molecular fingerprints that are all but impossible to erase (Jackson et al., 1998; Keim et al., 2000). These fingerprints are inherent to many, if not all, bioweapons agents on the A-List, including bacteria and viruses against humans and animals. It is therefore feasible to sequence the genes of such agents, organize that information in large databases, and use this molecular information to strengthen future BWC agreements and homeland security efforts. The elements of the plan are as follows.
The United States, the world’s leader in biotechnology, is in a position to create a new kind of high-throughput molecular forensics laboratory against bioweapons agents. Optimally, there would be two such facilities. The first would be domestically based, used to enhance homeland security, and serve as a model to states that are parties to the BWC. The second would be internationally based and offer improved verification and compliance capabilities to future BWC agreements. These two facilities could generate complementary and corroborative information.
A dedicated high-throughput laboratory against bioweapons agents would offer several important capabilities. First, it would enable exhaustive molecular fingerprinting and taxonomic positioning for a broad spectrum of known threat agents. Second, it would perform such analyses in a consistent and chain-of-custody manner. Third, it would produce high-resolution information within hours to days after sample receipt. In addition, the domestic facility could operate with a “closed” compartment, offering capabilities to the national and homeland security communities, and a separate “open” compartment, offering capabilities to the scientific community. The international facility could operate with capabilities and compartments established by future BWC agreements. Such arrangements would enable the United States to maintain its own molecular forensics and database capability yet share powerful testing methods and technologies with states that are parties to the BWC.
In 1992, the Australia Group identified nearly 100 bacteria, viruses, fungi, and toxins against people, animals, and plants with potentials for weaponization.
To date, however, only about 20 infectious agents have been used to produce biological weapons. A realistic goal would be therefore to fingerprint and catalog this “low hanging fruit.” From a technical, economic, and political standpoint, the result would be to make it more difficult to mount and maintain a secret offensive bioweapons program.
All the necessary technologies are available to build and operate a highthroughput molecular forensics laboratory and database system against bioweapons agents (Layne et al., 2001). More than a hundred companies manufacture the necessary equipment, which generally consist of flexible “plug-and-work” modules, and such technologies are often integrated into one of two kinds of system designs. The first are portable devices offering relatively simple and rapid tests. The second are high-throughput automation and robotic systems offering highly definitive tests. These larger systems must be housed in a semitractor trailer or suitable building, where samples must be brought to them. As outlined below, the optimal system would integrate both designs.
Several portable laboratory devices are available that fit into a suitcase and perform simple (yes/no) detection tests on the spot. The tests are based on polymerase chain reaction (PCR) methods and utilize tailored molecular primers against specific biothreat agents, such as B. anthracis. A larger set of primers is capable of screening for a larger list of biothreat agents. Such portable devices are able to detect very small traces of organisms but cannot actually sequence their genes. They often incorporate a personal computer to control and monitor tests, an Internet link to enable real-time data acquisition, and a global positioning device to automatically track locations. With such technologies, a trained individual can screen about two dozen samples per hour. To increase testing capacity, multiple devices can be deployed.
More definitive molecular forensics tests require more steps. A large assortment of automated and robotic equipment is available for this kind of work. Such industrial-scale technologies (e.g., robotic arms/conveyers, bar code readers, liquid handlers, incubators, genomic sequencers, flow cytometers, and image analyzers) are capable of performing all the procedures required by the proposed high-throughput molecular forensics laboratory. From a design standpoint, the various plug-and-work modules would be integrated into a flexible working system that could be upgraded with the latest commercial technologies. Incoming samples would follow an orderly flow, with different massanalysis lines focusing on different biothreat agents. Because of automation and miniaturization, the entire facility (which permits the growing, extracting, sequencing, and archiving of samples) would fit into a surprisingly compact space that contains biohazardous materials and safeguards workers.
More sequence information is always better for molecular forensics, yet there are tradeoffs between laboratory productivity and definitive identifications. Complete viral genomes range in size from 10,000 to 300,000 DNA or RNA bases, whereas complete bacterial genomes range in size from 1,000,000 to 6,000,000 DNA bases. (In comparison, the human genome is composed of about 4,000,000,000 bases.) To fingerprint and taxonomically position biothreat virus, the molecular forensics laboratory would have to sequence and analyze 50 percent to 100 percent of each isolate’s genome. On the other hand, to do this for biothreat bacteria, the laboratory would have to sequence only 5 percent to 10 percent of each isolate’s genome. Current technologies would enable a highthroughput molecular forensics laboratory to sequencing about 10,000,000 bases per day. This would correspond roughly to fingerprinting and positioning about 500 viruses or 50 bacteria per day. Such procedures could be completed within hours or days after receiving samples.
The high-throughput laboratory would also be able to perform the simpler (yes/no) PCR-based tests described above. A surge capacity of 10,000 samples per day would be feasible with current technologies. At such rates, however, the limiting factors would be sample collection and transportation rather than rapid testing.
The high-throughput molecular forensics laboratory would generate a sizeable database within a few years. In addition to cataloguing molecular fingerprints, the laboratory would also be able to analyze the taxonomic position and natural genetic history of threat agents (genealogies). In reach-back and attribution scenarios, genealogies could prove to be more powerful than fingerprints alone. The most recent generation of teraflop computers, which can achieve speeds of 30 x 1012 calculations per second, would be well suited to analyze the threat agent database. Domestically, the goal would be to support decisionmaking processes and offer surge capacity for public health, emergency medical, agricultural, and law enforcement efforts. Internationally, the goal would be to support United States national security and intelligence operations as well as future BWC agreements. The toolbox for such undertakings includes currently available tracking, mapping, and modeling technologies.
Virtually Assured Detection And Response (VADAR)
The United States has mature policies to deter nuclear attacks, set forth as mutual assured destruction (MAD). It also has established policies to deter conventional attacks, set forth by the ability to fight on one or two major fronts and several minor fronts at once. But the United States has few well-developed policies to deter biological attacks. A high-throughput molecular forensics laboratory and database facility would help to fill this gap by enabling a new policy of virtually assured detection and response (VADAR) regarding biological attacks. The framework is as follows.
The collapse of the system of two opposing superpowers has led to an uncertain world order characterized by one global ultrapower, a majority of responsible governments, several rogue states, multiple religious fringe groups, and some shadowy international syndicates that are forming new networks and posing new challenges to global security. Today, at least 17 countries are known to be developing or producing bioweapons and the list may be expanding.
The scale of global trade also poses a major challenge. For example, more than 14,000 loaded 40-foot marine containers enter the United States each day (Flynn, 2000). Containers routinely travel through the country before reaching a port of entry and the system tracking their intended course and location is rudimentary. Furthermore, few containers undergo any form of inspection and, even when this occurs, specialized inspection technologies are rarely used. The ease of smuggling bioweapons constitutes a significant threat to homeland security, in part because the problems associated with marine containers represents only the “tip of the iceberg” in our leaky border controls.
The foot-and-mouth disease outbreak in Great Britain and Europe, where the economic loss is estimated above £25 billion, reflects another aspect of the problem. Current methods of disease control, which rely on veterinarians inspecting animals for signs of infection, collecting mucosal and blood samples, and analyzing them with manual laboratories, have cycle times of three to five days. Foot-and-mouth disease can spread from one location to another, however, in far less time. Consequently, the current system with manual laboratories cannot support science-based decisions on quarantine zones, animal destruction, and resource allocation. At the heart of the problem is a lack of rapid, accurate, and complete information on which to make dependable decisions. A quantum leap in threat agent surveillance and data analysis is needed.
In a bioattack on the United States, as few as 50 sickened people in one major city could stretch public health, emergency medical, and law enforcement services beyond local capabilities. Larger attacks involving major metropolitan areas would be overwhelming and require the delivery of tons of antibiotics to exposed persons within days, challenging national capabilities. A coherent program that strengthens homeland security thus requires sizeable laboratory and informatic resources that can be organized in terms of four overall phases.
First, in preventing attacks, the United States would rely on the ability to fingerprint and catalogue bioweapons agents with high-throughput technologies. An extensive database of molecular fingerprints and associated origins would offer a new means of rapid attribution and therefore deterrence. It would put rogue states, religious fringe groups, and international syndicates on notice that there is little chance to evade blame for bioattacks.
Second, in the unfortunate event of an attack, public health laboratories would be overwhelmed simply because there would be too many samples to analyze quickly. Manual laboratories would be unable to answer even the simplest questions: Is the agent present? How many different infectious agents were
released? How do they differ? What are the best initial therapies to treat those afflicted and exposed? Information from high-throughput laboratories would reduce confusion and save lives by offering rapid testing in acute situations.
Third, in the aftermath of an attack, public health, agricultural, and law enforcement officials would need accurate answers to another set of questions. What are the geographic boundaries of each infectious agent? What are their stabilities? What are the effects on animals and plants? Information from highthroughput laboratory and mapping systems would speed the recovery process by offering testing for cleanup and investigatory operations.
Fourth, in response to the attack, law enforcement officials must collect evidence in accordance with chain of custody procedures. Intelligence agencies and military services must make accurate attributions and take swift actions to protect national security. Information from high-throughput laboratories and their associated databases could prevent further attacks by rapidly pinpointing suspected sources.
The relatively small anthrax attacks in a few American cities flooded the bioterrorism response network. Thousands of samples were sent to a patchwork of state and federal laboratories which, at best, were equipped to handle about 100 samples per day (Kahn et al., 2000). Even with many laboratories working around-the-clock, they could not keep pace with emergency testing demands.
Strengthening homeland security against bioterrorism needs enhanced public health and emergency medical preparedness at home and expanded human intelligence capabilities abroad. Moving beyond the BWC Protocol stalemate requires reliable disclosure of dual use facilities, timely inspection of suspicious programs, and systematic testing for certain (i.e., a short A-List) weaponizable agents. The common element among such undertakings is rapid, complete, and reliable information on which to make assessments and decisions.
A high-throughput molecular forensics laboratory and database facility would cost several hundred million dollars to build and operate over the first five years. Since the needed technologies already are available, it could be operational within two years.
Such a facility could be operated under the newly created Homeland Security Council. The mission of this new national medical forensics and intelligence support laboratory would be to complement and cooperate with existing government agencies such as health, agriculture, emergency management, justice, defense, intelligence, and the national laboratories. It would support public health, law enforcement, and homeland security programs without usurping their long-established missions. It would provide needed surge capacity in the acute and cleanup phases of terrorist bioattacks. It would also have mechanisms to support certain scientific and technical research.
In building the first molecular forensic laboratory against bioweapons agents, the overall testing methods and high-throughput capabilities would be shared with the scientific community. The design of certain molecular primers against specific biothreat agents and resulting fingerprint and genealogies, however, would be available to the national and homeland security communities only. Such open architectures would facilitate the development a second internationally-based laboratory that parallels the initial design.
In the aftermath of the terrorist attacks on the World Trade Center and Pentagon and organized anthrax attacks in several American cities, there has been renewed debate on the risks of further biological attacks. At present, the risk remains unclear. Yet it is clear that terrorist attacks have become more spectacular and lethal and have now reached our homeland soil. The question is: When will the shift to more devastating forms of bioterrorism take place? The United States now has the opportunity to organize effective prevention, deterrence, and response measures.
The United States must also act on domestic and international fronts. In mitigating bioterrorism, is VADAR a perfect solution? No. Is it an improvement over existing methods and policies? Yes. Is it possible to circumvent? Yes. But with secret offensive bioweapons programs possibly assisting organized terrorism, can we afford to wait?
RESEARCH AND THE PUBLIC HEALTH RESPONSE
Eric Eisenstadt,* Ph.D.
Defense Advanced Research Projects Agency
Technology could help public health enormously; but to help focus the development of technology for public health (as well as for the FBI and other law enforcement agencies who cope with forensic issues that resemble the diagnostic ones faced by public health), the public health community needs to articulate its technology needs. Once these needs are defined, then the science and technology communities, including funding organizations such as (DARPA), can begin to define the science and technology programs required to develop the desired capability. In this way, bridges can be built between public health and the technical community. Indeed, agencies like DARPA are very good at assembling the kind of interdisciplinary scientific and technical efforts—involving academia, industry, and government laboratories—that are required to develop new capabilities. Suppose, for example, the case could be made for a routine molecular
diagnostic capability that would provide a point-of-care physician with the information needed to make a decision about which treatment to prescribe within 30 minutes of taking a blood sample from a patient. A research and development effort might then be mounted to develop this new diagnostic capability by assembling researchers from the appropriate technical and user communities—e.g., molecular biology, materials science, signal processing, and clinical microbiology—to work together to create a new technology. The challenge would be enormous but the magnitude of the development effort will be a strong function of how strongly the case had been made for doing it in the first place.
Genomics-based technologies, for example, have great potential for improving public health. Fulfillment of this potential would be accelerated if the public health community participated in developing a vision of how the application of genomics information could enhance health care. Such a vision might serve to rally the nation to develop technological capabilities that enhance our ability to cope with many of the bioterrorism response and preparedness issues that have been identified in our discussions.
During World War II, for example, it was recognized that radar had tremendous potential for identifying U boats. The proof of principle had been done, but the technology still needed to be developed. A vision of what radar might be capable of doing for the military led to the initiation of the radar program at MIT from which great science and technology emerged including the foundations of the microelectronics industry.
Finally, it is very difficult to bound all of the bioterrorism response capabilities that have been discussed during this workshop. There are simply too many imaginable bioterrorist scenarios (multiple agents and multiple ways to create mischief with them). We do not have sufficient resources to address an unbounded set of problems. So we must try in some rational way to bound bioterrorism and define the set of bioterrorism issues that need to be addressed. We must focus and develop a big vision that the country can respond to. For example, why not identify as a national goal the removal of infectious disease as a public health threat? This does not mean that we need to define how to eliminate infectious disease. When, a few hundred years ago, the British parliament recognized the need “to find longitude” they didn’t know how it was going to be done. But by crisply stating the problem and offering a prize to the one who solved it, some fantastic science and technology emerged. Could we not rally the country behind a campaign to eliminate the infectious disease threat?
Centers for Disease Control and Prevention (CDC). 1999. Update: West Nile-like viral encephalitis – New York, 1999. Mortality and Morbidity Weekly Record 48:890–892.
Collin JF, Melet JJ, Mortot M, Foliquet JM. 1981. Eau d’adduction publique et gastro-enterites en Meuthe-et-Moselle. Journal of Français d’Hydrobiologie 35:155–174.
Jarroff L. 1999. Of mosquitoes, dead birds and epidemics. Time 154(15):74–75.
Miller JR, Mikol Y. 1999. Surveillance for diarrheal disease in New York City. Journal of Urban Health 76:388–390.
Pavlin JA, Mostashari F, Kortepeter MG, Hynes NA, Chotani RA, Mikol YB, Ryan MAK, Neville JS, Gantz DT, Writer JV, Florance JE, Culpepper RC, Henretig FM, Kelley PW. 2001. Innovative Surveillance Methods for Rapid Detection of Emerging Infections—The Results of an Inter-Agency Workshop on Health Indicator Surveillance.
Proctor ME, Blair KA, Davis JP. 1998. Surveillance data for waterborne illness detection: an assessment following a massive waterborne outbreak of Cryptosporidium infection. Epidemiology and Infection 120:43–54.
Rodman JS, Frost F, Davis-Burchat L, Fraser D, Langer J, Jakubowski W. 1997. Pharmaceutical sales a method of disease surveillance. Journal of Environmental Health 60:8–14.
Rodman JS, Frost F, Jakubowski W. 1998. Using nurse hot line calls for disease surveillance. Emerging Infectious Diseases Apr–Jun;4(2):329–32.
Steele KE, Linn MJ, Schoepp RJ, Steele KE, Linn MJ, Schoepp RJ, Komar N, Geisbert TW, Manduca RM, Calle PP, Raphael BL, Clippinger TL, Larsen T, Smith J, Lanciotti RS, Panella NA, McNamara TS. 2000. Pathology of fatal West Nile virus infections in native and exotic birds during the 1999 outbreak in New York City, New York. Vet Pathol 37:208–224.
Thacker SB. 1994. Historical development. In: Teutsch SM and Churchill RE eds. Principles and Practice of Public Health Surveillance. New York, NY: Oxford University Press 8–9.
Thacker SB, Berkelman RL. 1988. Public health surveillance in the United States. Epidemiology Review 10:164–90.
Flynn SE. 2000. Beyond Border Control. Foreign Affairs 79: 57–68.
Jackson PJ, Hugh-Jones, ME, Adair DM, Green G, Hill KK, Kuske CR, Grinberg, LM, Abramova, FA, and Keim P. 1988. PCR analysis of tissue samples from the 1979 Sverdlovsk anthrax victims: The presence of multiple Bacillus anthracis strains in different victims. Proceedings of the National Academy of Sciences, USA 95: 1224–1229.
Kahn AS, Morse S, and Lillibridge S. 2000. Public health preparedness for biological terrorism in the USA. Lancet 356: 1179–1182.
Keim P, Price LB, Klevytska AM, Smith KL, Schupp JM, Okinaka R, Jackson, PJ, and Hugh-Jones ME. 2000. Multiple-locus variable-number tandem repeat analysis reveals genetic relationships within Bacillus anthracis. Journal of Bacteriology 182: 2928–2936.
Layne SP, Beugelsdijk TJ, and Patel CKN (editors). 2001. Firepower in the Lab: Automation in the Fight Against Infectious Diseases and Bioterrorism. Washington, D.C.: Joseph Henry Press.