A Hot Idea

So what are we to make of all the statistical analysis of our wargame? For one, it is clear that, when dealing with a known agent, decisionmakers place considerable value on data that provide positive identification of pathogens. If fact, the Positive ID Group was the only statistically significant group that was viewed as trustworthy for a known agent. In the overall rating, they were the highest ranked individual group and were part of all the other system-of-system approaches that were statistically significant. This conclusion parallels the current focus on sensors as a component of our strategies for homeland security.

As a nation, we are investing a significant amount of money in sensors as part of the BioWatch program. Recent press reports have described a $60 million sensor network deployed in 31 cities with a proposed increase to $118 million for 2005 to cover additional cities. 85 Details of the program are understandably classified, but there has been local acknowledgment of sensors in New York, Washington, Chicago, Houston, San Francisco, San Diego and Boston.86 Government officials will not confirm what agents the system screens for, but they say it is less than a dozen.87 That most likely limits it to the candidate agents listed in CDC’s Category A list. 88

The BioWatch program has not been universally accepted and has plenty of critics. Calvin Chue, a researcher at Johns Hopkins University, points out that BioWatch would likely be effective only in detecting a major atmospheric release. 89 Small-scale attacks, or attacks delivered through food or water, that could result in hundreds—or thousands—of victims would probably go undetected. Moreover, as Chue comments, BioWatch sensors do not monitor any indoor environments.90 In this case, an indoor release will only be detected once it leaves the building and encounters a BioWatch sensor. The Director of the Center for Biological Defense at the University of South Florida—Jacqueline Cattani—summed up her feelings as follows: “If you saw planes going over and releasing major clouds of this stuff, there’s a chance that people would get suspicious a long time before anybody checked the filters.” 91

Other critics further emphasize the fact that the sensors only cover one-half of the U.S. population. 92 They point out problems with positioning of the sensors and question the amount of air sampled. 93 Additionally, they cite the high labor costs involved with the system.94 The filters have to be collected and processed by trained laboratory technicians.

Even the proponents of the program make statements that are somewhat self-damning. One EPA executive was quoted as saying that if an attack were close enough to a sensor, authorities could know about it within 12 hours. 95 He rightly pointed out that 12 hours is quicker than if we waited for victims to develop symptoms.96 However, that if is a major point and raises the issue of spatial distribution and placement, as well as meteorological conditions. It is worth noting that an EPA sensor was in place just blocks from the World Trade Center towers, but following the collapse on September 11th, it did not register the incident—only when the wind direction changed on September 12th did the sensor become “aware” of the incident. 97

The problem of false alarms (positive or negative) must also be considered for the BioWatch sensor network. One senior government official was quoted as saying that of the nearly 500 sensors nationwide, not one had ever raised a false alarm. 98 That is so statistically unlikely as to be considered impossible.

Even 500 sensors are far too few for the coverage sought by BioWatch. The Federal Government is seeking to increase the number of collectors per city. According to Congressional testimony by Department of Homeland Security Under Secretary for Science and Technology Dr. Charles McQueary in late February 2004, the “average” city covered by the current BioWatch program only has ten collectors. 99 Studies indicate that 40-60 would provide the optimal coverage. In addition, according to Dr. McQueary, cities have requested more collectors to cover key facilities, such as transit systems, airports, and stadiums. 100

As noted above, the BioWatch Program is capable of screening for fewer than a dozen known agents (i.e. agents that are predesignated as potential bioterrorist weapons). What about dealing with unknown agents, or known agents that are genetically modified to beat the sensor systems? In our wargame analysis, 101 one of the scenarios tested calls for an unknown agent. In this scenario, sensors only show up as statistically significant when in combination with business reporting—and then only when assessing trustworthiness. (And, as stated in the discussion of the analysis in the previous chapter, this is most likely an anomalous finding, based on how the questions were posed. Re-evaluation of this finding in the next iteration of this study is indicated.) In all other cases, medical reporting is the only statistically significant measure when dealing with an unknown agent. This is an anticipated result, as the Medical Group relied on symptom recognition and required no knowledge of the agent—in other words, unknown agents are identical to known agents.

What is the likelihood an unknown agent will be released on the population? Would it be viral? Bacterial? Contagious? Actually, the human population deals with this issue frequently. An analysis of pathogenic microbes and infectious diseases reveals that, on average, for the 30-year period between 1973 and 2003, one new (i.e., previously unknown) disease emerged annually. 102 These new diseases did not appear only in remote corners of the world. Some of the more memorable ones that impacted the continental United States and Canada include Legionnaires’ disease (1977), HIV/AIDS (1981), West Nile virus (1999), and SARS (2003).

Returning to our wargame analysis, medical reporting also ranks high when dealing with known biological agents. This suggests that rather than spending additional money on a sensor-based system that has applicability primarily for known agents, why not expand the system to include a medical component that has applicability across both known and unknown agents? (While DHS is attempting to develop an integrated, real-time, human-animal-plant surveillance system, as part of its Bio-Surveillance Program Initiative, such a complete system is many years off. 103 The sensor-based BioWatch component is currently a major part of the initiative.) This medical approach seems particularly prudent, given that we are investing in a system designed to detect both the possible (known agents) and the probable (unknown agents). Such a system-of-systems would then be useful both against bioterrorism and for general public health.

The JASON study noted that we already have roughly 300 million biosensors in the U.S.—our population. 104 Not only do humans sample the air when we breathe, but we also concentrate the sample, and our immune systems and innate responses to insult from biological challenges act as a detector. These responses are both highly specific and highly sensitive. If a sample of the population were monitored for the first signs of symptoms of challenge (regardless of route of infection), then the information gleaned could also be timely.

On the basis of these observations and our wargame analysis of the utility of the existing technology solutions, we propose that policemen, firemen, and mail carriers be used as a sentinel population to monitor for possible outbreaks of known or unknown agents. Statistically, they provide a better sample of air than stationary collectors, because they are more uniformly distributed across a metropolitan area. Unlike the current systems, they are not subject to the vagaries of microclimates. Also, during the course of their duties they encounter both outdoor and indoor environments. Their daily routes and movements are fairly well-defined, making it easy to pinpoint affected areas.

Although the performance characteristics of the BioWatch collectors remain classified, it is relatively easy to estimate their capacity to sample the air. Current EPA “high-volume” environmental air samplers are rated at taking in 40-60 cubic feet of air per minute (cfm). 105 Assume a “very-high-volume” capacity for the BioWatch collectors at 100 cfm (equal to 2,832 liters/minute). At that rate, the collectors sample the equivalent of one large room (24.5’x24.5’x10’) per hour, or approximately 170,000 liters of air.

Take the metropolitan Washington, DC workforce as a sentinel population. There are 12,110 police and sheriff’s patrol officers, 4,900 firefighters, 400 EMS workers, 5,940 mail carriers, and 150 parking enforcement workers for a total population of 23,500 people. 106 Assume an average breathing rate of .5 liters per breath, 107 with 16 breaths per minute. 108 (That is a conservative breathing rate, as it is derived from “at rest” figures. Given their level of physical activity, this population is very likely breathing more than that.) So, in one hour each member of this sentinel population samples 480 liters of air. Per hour, the entire workforce samples 11,280,000 liters of air. That is roughly the equivalent of 66 very-high-volume samplers.

Figure 21 – The selected workforce is uniformly distributed across the city and encounters both indoor and outdoor environments.

These samplers are highly specific and highly sensitive, and they have no false positives. What data could be efficiently and economically collected from them, to assess the results of their sampling? Considerable work is being done on identifying various blood components that would indicate the earliest signs of disease. These components would be non-specific and alert us that the body is in the early stages of an immune response. Further blood tests would be needed to identify the agent. Urine, sweat, or breath might also be tested. The perfection of these early identification tests (often referred to as prodromal states) is still well in the future. Plus, there will likely be considerable logistics and economic factors to consider.

At the moment, of all the potential parameters that can be readily measured, body temperature offers the most efficient and economical approach. Moreover, the wargame results indicated medical information is the preferred data when dealing with an unknown agent. Body temperature is a primary piece of medical information.

The technology for measuring body temperature was widely used during the recent SARS epidemic. Various approaches were used, including infrared thermal imagers, oral or ear fever thermometers, and forehead, or temporal artery, infrared thermometers. All focused on measuring specific body temperatures for analysis of any elevation above the expected norm.

The most efficient method for monitoring a workforce is most likely the infrared thermal imagers. While they are not considered the most precise method of measurement, they are good at detecting if someone appears hotter—or colder—than another person.

Figure 22 – In general, the immune system will register the presence of a foreign agent by increasing the body temperature within a matter of hours. These increases could be easily detected by infrared thermal imagers, such as those shown here being used during the recent SARS outbreak.

Every member of a shift could be scanned as he/she came to work. They could all be scanned again at the end of their shift. If members of a shift were exposed near the end of their tour, their temperature might not be elevated by the time they are scanned prior to going home. However, it is likely that the next shift coming on will also be exposed—this time at the beginning of their shift—and will have a measurable temperature increase by the end of their shift. (Baseline data could be stored on all employees and individual temperature variations could be easily accounted for.)

In general, the immune system will register the presence of a foreign agent by increasing the body temperature within a matter of hours. 109 One person showing an elevated temperature might not warrant further investigation, but a cluster would suggest the need for further blood tests.

Given the cost of purchasing and maintaining the systems proposed under BioWatch, it seems worth the time and effort to conduct a test of the feasibility of thermal imaging as a medical monitoring technique. It certainly appears to offer a rapid, simple, and cost effective capability for disease detection that is currently ignored.

By contrast, the military presents a different set of problems, when considering the detection of bioagents. Forces in the field need a standoff detection capability. Not only does it provide them with a detect-to-warn capacity that would not be practical in a civilian setting, it also provides them with a reconnaissance capability. There is no argument that the continued development and deployment of biosensors is critical for combat operations.

The Department of Defense is scheduled to award a $1.1 billion contract for its Installation Protection Plan (IPP). 110 Ultimately, the IPP is to establish a network of chemical, biological, radiological and nuclear-detection sensors at 200 military installations worldwide.111 While the chemical, radiological and nuclear-detection sensors may be important and necessary parts of the Guardian program, this study suggests that the biosensor component might be modified to include thermal imaging. As with the BioWatch program, a short test of this idea seems fiscally responsible, prior to embarking on this five-year effort.

85. Deborah Charles, “US Germ Detection System Active in 31 Cities,” Reuters, Washington, DC, November 14, 2003. See <http://www.stevequayle.com/News.alert/03_Global/031117.bio.detection.html>, accessed December 2003.
86. Sean Whaley, “Outdoor Sensors Take Time to Detect Dangers,” Las Vegas Review-Journal, Carson City, NV, December 27, 2003. See <http://www.reviewjournal.com/lvrj_home/2003/Dec-27-Sat-2003/news/22881907.html>, accessed December 2003.
87. Stephen Prior, PhD, National Security Health Policy Center at the Potomac Institute for Policy Studies, personal communication.
88. The Center for Disease Control and Prevention has standardized a list of the most likely bioterrorism agents. These six agents, identified as “Category A” agents, are anthrax, botulism, plague, smallpox, tularemia, and viral hemorrhagic fevers.
89. “Government Provides Details of Bioterror Sensors in Cities,” The Associated Press, Washington, DC, November 15, 2003. See <http://www.nytimes.com/2003/11/16/national/16TERR.html?ex=1082520000&en=c69bb316d19aa9b7&ei=5070>, accessed December 2003.
90. The Associated Press (Nov. 15, 2003).
91. Ibid.
92. David McGlinchey, “US Health Officials Highlight Surveillance Systems,” Global Security Newswire, October 22, 2003. See <http://www.nti.org/d_newswire/issues/2003/10/22/44363c15-eeed-40ec-8098-a9beae659465.html>, accessed December 2003.
93. David W. Siegrist, Technology-Based Biodefense, Potomac Institute for Policy Studies, September 2003: 5. See <http://www.potomacinstitute.org/pubs/BTIII_Intro.pdf>, accessed March 2004.
94. Charles (Nov. 2003).
95. “Fed’s Sniffing Devices Effective?” The Associated Press, Philadelphia, PA, July 17, 2003. See <http://www.cbsnews.com/stories/2003/07/17/tech/main563765.shtml>, accessed December 2003.
96. Ibid.
97. Stephen Prior, PhD, National Security Health Policy Center at the Potomac Institute for Policy Studies, personal communication.
98. The Associated Press (Nov. 15, 2003).
99. Charles E. McQueary, Statement for the Record, Before the U.S. House of Representatives Subcommittee on Cybersecurity, Science, and Research & Development, February 25, 2004: 6.
100. Ibid., 20.
101. See “Wargame,” this document.
102. “US ignoring dangerous diseases during war on terror, says researcher,” Canadian Press. See <http://story.news.yahoo.com/news?tmpl=story&u=/cpress/20040429/ca_pr_on_he/health_bio_terror_conference&cid=2155&ncid=2155>, accessed May 2004.
103. Tommy Thompson, Bill Frist, Press Conference, January 29, 2004. See <http://www.dhs.gov/dhspublic/display?content=3093>, accessed March 2004.
104. JASON, “Biodetection Architectures,” The MITRE Corporation, McLean, VA, February 2003: 14. See <http://www.fas.org/irp/agency/dod/jason/biodet.pdf>, accessed December 2003.
105. United States Environmental Protection Agency, “Particulate Emission Measurements from Controlled Construction Activities,” Prepared by National Risk Management Research Laboratory, Research Triangle Park, NC, April 2001: 14. See <http://www.epa.gov/appcdwww/apb/R-01-031/EPA-600-R-01-031.pdf>, accessed January 2004.
106. 2002 Metropolitan Area Occupational Employment and Wage Estimates, Washington, DC-MD-VA-WV PMSA. See <http://www.bls.gov/oes/2002/oes_8840.htm#b43-0000>, accessed February 2004.
107. See <http://www.nqinc.com/faq16.html>, accessed March 2004.
108. See <http://transitiontoparenthood.com/ttp/parented/pregnancy/abdobreathe.htm>, accessed March 2004.
109. Christopher Green, MD, PhD, Wayne State University Medical School, personal communication.
110. “Lockheed Martin, CACI, EAI, Fluor and IEM Team For DoD Installation Protection Program,” PRNewswire, CACI International, Inc., Manassas, VA, February 20, 2004. See <http://www.stockhouse.ca/news/news.asp?newsid=2161668&tick=>, accessed March 2004.
111. Ibid.