Protecting public health in the age of bioterrorism surveillance: is the price right?

Introduction

Importance of the Study

Biological surveillance of bioterrorism incidents has gained recent attention in public health and medical research. The bioterrorism incident of October 2001 showed the vulnerability of the U.S. population to biological agents, as well as the high cost of potential attacks. The likelihood of future attacks has not diminished. According to the National Intelligence Council, current trends indicate that the risk of an attack against the United States will increase in the coming years (Frist, 2002). Recent scientific and technological advances that allow weaponization of spores like anthrax also increase the threat. Although a likely scenario is a smaller attack that can sicken thousands and kill hundreds, the U.S. population will not tolerate such "malpractice error of omission" on the part of policy makers and public health authorities (Henretig, 2001). Congress has passed new legislation intended to strengthen the nation's biological surveillance by increasing funding of federal and state biological surveillance, even though the probability and potential size of bioterrorism incidents is unknown (Frist, 2002). Government funding for civilian biodefense increased dramatically from $414 million in 2001 to over $5 billion in 2004 (Schuler, 2004). President Bush's Fiscal Year 2005 budget request includes $407 million for the Bio-Surveillance Program Initiative to strengthen the public health infrastructure (Schuler, 2004). A key component of this initiative will be the expansion and deployment of the next generation of technologies related to the BioWatch Program, a biological-surveillance warning system. Presently, biological surveillance is used exclusively for bioterrorism detection.

[ILLUSTRATION OMITTED]

BioWatch deploys biological sensors intended to detect potential releases of biological pathogens. It provides continuous monitoring of most major metropolitan areas and is designed to protect the nation against bioterrorism and strengthen the public health infrastructure. BioWatch also entails significant operation and maintenance costs, however. It is important that communities not spend more on biological surveillance than they reap in potential benefits from such a system, and public funding should maximize terrorism response capabilities (Fabian, 2002). Society's willingness to pay for prevention of future terrorism incidents should be based on the combined economic and public health damage that such attack can inflict. Therefore, before committing scarce resources to biosafety-enhancing technologies, policy makers should weigh the costs and benefits of biological surveillance.

Biological Surveillance: What Does It Buy?

A bioterrorist incident is considered a low-probability but high-cost event. The odds of such an event look even smaller from a local perspective (Berg, 2004). The costs are high, because biological agents are undetected until the onset of pathology, when treatment is less effective and more expensive, and fatalities are more probable. The human and economic impact of the bioterrorist attack depends on several factors:

* the agent or toxin used,

* the probability of infection,

* the number of people exposed,

* the speed of detection,

* the availability of treatment, and

* the potential for secondary transmission (if a contagious agent like smallpox is involved).

Agents can be detected either directly, with biological detectors like the ones used by BioWatch, or by collecting reports of their effect on a population from health care providers. The deployment of BioWatch allows detection of threatening biological pathogens such as anthrax in a manner timely enough to allow for effective prophylactic treatment. Even though vaccines exist for anthrax, as well as many other biological agents determined to pose a high risk by the Centers for Disease Control and Prevention (CDC), it is unlikely that vaccination of the entire population for a rare disease would be cost-effective. Early detection will allow the exposed population to be treated within a few days of an attack, thus greatly reducing the likelihood that people will develop symptoms, the probability that fatalities will occur, and the probability that a contagious virus will be transmitted. The biological detectors deployed under BioWatch include air-sampling hardware, sample collectors, and lab-testing facilities that can pinpoint within 16 to 36 hours the time and the place of an aerosol release of selected biological agents. The system of detectors cannot prevent the exposure. It can, however, provide early identification of the agent used and of the population exposed, before the onset of symptoms. According to the Department of Homeland Security (DHS) Web site, several dozen BioWatch sites are dispersed around the United States in major metropolitan areas. Understanding and quantifying the benefits of biological surveillance are essential to developing new biological-surveillance technologies and allocating public funds to such programs.

Contribution to the Previous Literature

Previous literature offers ambiguous evidence as to the benefits of biological surveillance. Although some studies find that the huge human and economic costs that could arise from the use of biological weapons justify the investment in domestic preparedness, some researchers find sensor technologies to be of moderate impact. Several studies have attempted to quantify the impact of a potential bioterrorism attack with results ranging widely depending on the agent used.

Kaufmann, Meltzer, and Schmid (1997) model the impact of aerosolized release of three biological agents, Bacillus anthracis, Brucella melitensis, and Francisella tularensis in a suburb of a major city; they estimate the economic impact of a bioterrorist attack at between $477.7 million per 100,000 people exposed for the brucellosis scenario and $26.2 billion per 100,000 people exposed for the anthrax scenario. They also show the importance of early treatment by estimating the high medical costs of a bioterrorist attack relative to the cost of early intervention. The results indicate that delaying prophylaxis increases the risk for loss at a rate resembling a semi-logarithmic scale. StJohn, Finlay, and Blair (2001) find similar results for an aerosolized release of weaponized anthrax and botulinum toxin: Implementing postattack intervention before Day 3 after the event yields the greatest public health benefits. Therefore, biological surveillance systems that decrease time for detection will yield positive economic benefits. Wein, Craft, and Kaplan (2003) examined various potential emergency responses in the event of an anthrax attack. Their study finds that the deployment of biological surveillance produces only minor improvements. This result may be due to their assumption that without surveillance, detection would occur within two days. Although some patients become symptomatic within two days of anthrax exposure, they do not seek treatment, on average, for 3.5 days after the first symptoms appear (Jernigan et al., 2001). The surveillance of symptoms for the early patients often does not allow city officials to identify the exposed population in the way sensor technologies do. Therefore, the statistical model developed by Wein and co-authors (2003) may have underestimated the benefits of biological surveillance.

The research reported here adds to the existing literature discussed above on several fronts. First, the author estimates the lives saved as a result of a decrease in detection time from what current biological surveillance allows. Second, the dollar benefits of biological surveillance were estimated relative to actual government spending for maintenance of biological surveillance sites. Also, the cost-benefit model defines the conditions under which biological surveillance is economically justified--that is, the minimal biothreat probability that equates benefits and costs.

Methods

Benefits of Biological Surveillance

The impact of a bioterrorist attack can be reduced by employing biological detectors. The major benefit of biological surveillance is reduced mortality resulting from lower time to detection. The number of deaths averted can be expressed in monetary units by estimating the value of a statistical life (VSL), a measure of the productivity of an individual. This measure is used as a basic tool in evaluating the societal benefits associated with investments in programs to save lives. The standard expected-future-earnings approach multiplies the number of the deaths averted by the present value of expected future earnings. Since the sum of lifetime earnings overstates the current-year economic value of an individual, the discounted value of the future earnings is the appropriate measure of economic value over long periods of time. The average of future earnings for the U.S. labor force is estimated to be $1,688,595 per worker (Haddix, Teutsch, Shaffer, & Dunet, 1996) or, discounted at 5 percent (Lipscomb, Weinstein, & Torrance, 1966), it is $544,160, or $657,220 in 2004 dollars. The author's estimates of benefits depend on the value of human life. An alternative approach to the expected-future-earnings VSL is to place a dollar value on the life-saving benefits of regulations based on societal willingness to pay for mortality risk reductions. Based on an extensive literature review, the U.S. Environmental Protection Agency (1997) suggests that a reasonable estimate of the value of a statistical life is a mean of $4.8 million in 1990 dollars, or $6.96 million in 2004 dollars.

Other benefits not quantified in this study include medical costs averted because of sensor detection. Without detectors in place, the first indication of the attack would be a number of people seeking treatment for the same symptoms. If infection is anticipated, however, treatment is cheaper and more effective (Kaufmann et al., 1997). Averted medical costs include

* physician visits, treatment costs, and potential hospitalization costs;

* monitoring of naturally occurring diseases;

* agricultural benefits (BioWatch sensors can detect cattle diseases such as anthrax or foot-and-mouth disease, one of the most contagious animal diseases, entailing important economic losses); and

* emotional and behavioral impact of the bioterrorist attack (Stein, Tanielian, Eisenman, Keyser, Burnam, & Pincus, 2004).

The study reported here concentrated on the lives saved as a result of improved speed of detection.

Assumptions

The model presented here calculates the human consequences of a large attack on a civilian population with Bacillus anthracis. In this section, the author considers the impact of a potential bioterrorism attack on a city in which 100,000 people are exposed in the target area. Different studies offer widely varying estimates regarding the infectiousness of an anthrax aerosol release. In 1970, a World Health Organization (WHO) expert committee estimated that casualties following an aircraft release of 50 kg of weaponized anthrax spores over a developed urban area with a population of 5 million would be about 250,000 (WHO, 1970). A report by the Office of Technology Assessment established that from about 130,000 to 3 million deaths would follow the aerosolized release of 100 kg of weaponized anthrax spores upwind of Washington D.C. (1993). Wein and co-authors (2003) used a Gaussian plume model to compute the number of spores inhaled from a release of 1 kg of anthrax in a population of 1.49 million and estimated that 13.1 percent of the population would become infected. Inglesby (1999) modeled an anthrax release at a sporting event in a stadium with 74,000 people and estimated that 16,000 people might become infected.

The study reported here tried to address the wide range of estimates and situations. Thus, sensitivity analysis assesses three scenarios: 1) point release from an aircraft of 50 kg of anthrax over a population in a highly developed urban area; 2) population infected in a long thin region downwind of the point of release of about 1 kg of anthrax spores; and 3) truck anthrax release at a big public event (e.g., the Super Bowl). The first two scenarios represent the range of estimates for an anthrax release in an urban area, while the third scenario is a large-public-event release. See the sidebar on page 10 for assumptions of the model.

It is important to note that information about the sensitivity of the current biological surveillance system is not publicly released. Therefore, the author assumes that if the system is in place, it will be able to detect the kinds of large-scale releases described in the three scenarios above.

Results

Table 1 presents results for the three scenarios under consideration. Benefits are presented in terms of lives saved as a result of biological surveillance and dollar values for the two alternative VSL estimates.

Table 1 shows that benefits of biological surveillance range from 1,688 to 7,290 lives saved per 100,000 people exposed. This wide range is due to the differences in population density at the time of the release, time of release (daytime versus nighttime), weather and wind conditions, and the size of the release, which affects the average number of spores breathed in by an individual. The benefits range from over $1 billion to $50 billion depending on the nature of the release and the assumptions made about VSL.

BioWatch costs per year are approximately $13,672,096 and include labor costs, site upgrades, supplies, travel, training, and other operation and maintenance costs. In addition, lab costs of anthrax testing alone will come to $347,480 per year. These expenditures are justified if

(Probability of Bioterrorist Incident) X (Benefits of BioWatch) [greater than or equal to] (Costs of BioWatch) The higher the estimated benefits of biological surveillance are, the lower the perceived probability at which the expenditures are justified (i.e., if benefits are very high, then even after they are scaled by the probability of a threat, they will still exceed costs). More conservative expected benefits are justified if the probability of the biological attack exceeds 1.26 percent. If the perceived threat is below 1.26 percent, the costs of BioWatch exceed the expected benefits. In other words, the determination of whether biological surveillance passes the cost-benefit analysis rests on policy makers' perceptions of how probable a bioterrorist attack is. It is important to note that for 2005, the budgetary funding for the sensor network will increase to $118 million to cover additional cities, with $47 million to $60 million earmarked for the BioWatch project. Such a sharp increase in spending will further decrease the cost-benefit ratio of the bio-surveillance program but will also improve the population coverage.

The decisions of individual cities to invest in biological surveillance should be based on the benefits such investment will produce in terms of lives saved, as well as the perceived probability of the threat. The degree of threat, however, is unknown and may change over time. Public investment in biological surveillance is long term and costly. Detectors are installed to operate for many years and therefore result in a need for an infrastructure of lab workers and filter collectors. If the threat of bioterrorist attack decreases in the future (although it may never go to zero), dual use of biological detectors for public health monitoring should also be considered.

Discussion and Policy Implications

The results of this study reinforce the conclusions of other researchers (Kaufman et al., 1997; StJohn et al., 2001) by showing that early detection followed by prophylaxis of the exposed population is critical to saving lives. The sooner a biological release is detected, the sooner the population at risk can be identified, and the sooner medical care can be administered to those exposed. Reduction in detection delay from four days to 36 hours (plus 12 hours to locate and deliver the treatment) yields significant life savings in most scenarios.

The future of biological detectors will depend on the perceived bioterrorism threat as well as the public health benefits the detectors provide. Note that this study assumed a modest estimate of public health benefits by taking into account only the lives saved. Additional savings will come from medical costs prevented as a result of early prophylaxis. Kaufmann (1997) estimates the cost of early prophylaxis that is 90 percent effective to be between $51 and $226 per patient, depending on the treatment used. The medical costs following an undetected anthrax attack, however, vary from $422 (per non-hospitalized patient) to $4,541 (per hospitalized patient). Therefore, significant medical cost savings can be achieved because of the large differences between early-prophylaxis treatment and treatment for a symptomatic patient. Medical cost saving will enhance the benefits of early detection and the cost-to-benefit ratio of biological surveillance systems.

The author has estimated the benefits for one biological agent only, because that agent has been used in the United States. The choice of a noncontagious agent understates, however, the potential benefits of biological surveillance. In the case of a contagious disease (e.g., smallpox) early detection is essential to containing an outbreak. People today are incredibly mobile, commuting between urban centers and the suburbs on a daily basis, and traveling across cities and countries regularly. Therefore, the benefits of early detection will be considerably higher for infectious diseases.

Since a perceived bioterrorist threat can decrease in the future, it is important to note other benefits of biological surveillance. The same BioWatch system can be used to monitor communities for other public health threats, such as influenza or SARS, or to study biological agents that are endemic to the area under surveillance. This dual use will enhance the public health benefits of the systems.

Another limitation of this study is the fact that the author did not take into account the costs of a false alarm that biological detectors can yield. False positives may occur as a result of equipment malfunction or in response to naturally occurring strains of bacteria. For example, anthrax may be detected in areas with a large concentration of cattle. False alarms are very distracting and costly. An anthrax contamination false alarm closed congressional office buildings for several months. The author did, however, assume that the first detection of anthrax, which would occur within the first 24 hours of the release, would be rechecked within the next 12 hours by an examination of the background (i.e., naturally occurring anthrax), as well as the anthrax strain, thus minimizing the false-positive rate.

The public health community needs to build infrastructure relevant to both terrorism response and overall public health, including environmental health (Fabian, 2002). Biological surveillance minimizes the impact of a terrorist strike and can be used for the dual purpose of monitoring for outbreaks of infectious diseases. For example, early identification of a flu outbreak can prompt immunization of the vulnerable populations, thus saving lives and medical costs. The probability of a bioterrorist threat may never go away, but biological surveillance will prepare communities to respond to a catastrophe in a manner that minimizes human losses and serves to enhance public health assurance in potentially vulnerable communities.

Corresponding Author: Helen Schneider, Technical Staff Member, Los Alamos National Laboratory, D-3, MS K575, Los Alamos, NM 87545. E-mail: [e-mail].

REFERENCES

Berg, R. (2004). Terrorism response and the environmental health role: The million-dollar (and some) question. Journal of Environmental Health, 67(2), 29-39.

Fabian, N. (2002). Post September 11: Some reflections on the role of environmental health in terrorism response. Journal of Environmental Health, 64(9), 78, 77, 65.

Frist, B. (2002). Public health and national security: The critical role of increased federal support. Health Affairs, 21(6), 117-130.

Haddix, A.C., Teutsch, S.M., Shaffer, P.A., & Dunet, D.O. (1996). Prevention effectiveness: A guide to decision analysis and economic evaluation. New York: Oxford University Press.

Henretig, F. (2002). Biological and chemical terrorism defense: A view from the "front lines" of public health. American Journal of Public Health, 91(5), 718-721.

Inglesby, T.V. (1999). Anthrax: A possible case history. Emerging Infectious Diseases, 5(4), 556-560.

Inglesby, T.V, Henderson, D.A., Bartlett, J.G., Ascher, M.S., Eitzen, E., Friedlander, A.M., Hauer, J., McDade, J., Osterholm, M.T., O'Toole, T., Parker, G., Perl, T.M., Russell, P.K., & Tonat, K. (1999). Anthrax as a biological weapon: Medical and public health management: Working group on civilian biodefense. Journal of the American Medical Association, 281(18), 1735-1745.

Jernigan, J.A., Stephens, D.S., Ashford, D.A., Omenaca, C., Topiel, M.S., Galbraith, M., Tapper, M., Fisk, T.L., Zaki, S., Popovic, T., Meyer, R.F., Quinn, C.P., Harper, S.A., Fridkin, S.K., Sejvar, J.J., Shepard, C.W., McConnell, M., Guarner, J., Shieh, W.J., Malecki, J.M., Gerberding, J.L., Hughes, J.M., & Perkins, B.A., Anthrax Bioterrorism Investigation Team. (2001). Bioterrorism-related inhalational anthrax: The first 10 cases reported in the United States. Emerging Infectious Diseases, 7(6), 933-944.

Kaufmann, A.F., Meltzer, M.I., & Schmid, G.P. (1997). The economic impact of a bioterrorist attack: Are prevention and postattack intervention programs justifiable? Emerging Infectious Diseases, 3(2), 83-94.

Lipscomb, J., Weinstein, M.C., & Torrance, G.W. (1966). Time preference. In M.R. Gold, J.E. Siegel, L.B. Russell, & M.C. Weistein (Eds.), Cost-effectiveness in health and medicine (pp. 214-235). New York: Oxford University Press.

Office of Technology Assessment, U.S. Congress. (1993). Proliferation of weapons of mass destruction (OTA Publication No. ISC-559). Washington, DC: U.S. Government Printing Office.

Schuler, A. (2004). Billions for biodefense: Federal agency biodefense funding, FY2001-FY2005. Biosecurity and Bioterrorism, 2(2), 86-96.

Stein, B.D., Tanielian T.L., Eisenman, D.P., Keyser, D.J., Burnam, M.A., & Pincus, H.A. (2004). Emotional and behavioral consequences of bioterrorism: Planning a public health response. Milbank Quarterly, 82(3), 413-55.

StJohn, R., Finlay, B, & Blair, C. (2001). Bioterrorism in Canada: An economic assessment of prevention and postattack response. Canadian Journal of Infectious Diseases, 12(5), 275-284.

U.S. Environmental Protection Agency. (1997). The benefits and costs of the Clean Air Act: 1970-1990 (EPA 410-R-97-002). http://www.epa.gov/air/sect812/copy.html (25 Aug. 2005).

Wein, L.M., Craft, D.L., & Kaplan, E.H. (2003). Emergency response to an anthrax attack. Proceeding of the National Academy of Sciences, 100(7), 4346-4351.

Wenner, K.A. & Kenner, J.R. (2004). Anthrax. Dermatologic Clinics, 22(3), 247-256.

World Health Organization. (1970). Health aspects of biological weapons: Report of a WHO group of experts. Geneva, Switzerland: Author.

Helen Schneider, Ph.D.

Assumptions of the Model

Population exposed                              100,000
Probability of infection
  Scenario 1                                    0.05 (a)
  Scenario 2                                    0.131 (b)
  Scenario 3                                    0.216 (c)
Time to treatment with surveillance             48 hours
Time to detection                               36 hours
Time to locate and deliver the treatment        12 hours
The epidemic curve for anthrax                  0 cases after less than
                                                24 hours of exposure,
                                                5% cases after 1 day,
                                                20% after 2 days,
                                                35% after 3 days,
                                                20% after 4 days,
                                                10% after 5 days,
                                                5% after 6 days,
                                                5% after 7 or more
                                                days (d)
Probability of fatalities if early prophylaxis  0.1125 (25% develop
  is administered                               symptoms within 2 days,
                                                while fatalities without
                                                prophylaxis are 45%) (d)
Time to detection without surveillance          4 days-6 weeks (e)
Probability of fatalities without surveillance  0.45 (e,f)
Dollar value of statistical life                $657,220 (g)
                                                $6.96 million (h)

(a) World Health Organization, 1970
(b) Wein, Craft, & Kaplan, 2003
(c) Inglesby, 1999
(d) Kaufmann, Meltzer, & Schmid, 1997
(e) Wenner & Kenner, 2004
(f) Inglesby et al., 1999
(g) Haddix, Teutsch, Shaffer, & Dunet, 1996; Lipscomb, Weinstein, &
Torrance, 1996
(h) U.S. Environmental Protection Agency, 1997

TABLE 1 Estimates of Benefits of Biological Detection

                                  Benefits (VSL =
                                  Expected Future  Benefits (VSL =
                                  Earnings in      Societal Willingness
                                  Billions of      to Pay in Billions of
Scenarios            Lives Saved  Dollars)         Dollars)

1. Aircraft release  1,687.5      1.11             11.75
   in urban area
2. Truck release in  4,421.25     2.91             30.77
   urban area
3. Truck release at  7,290        4.79             50.74
   a public event