3

Public Health’s Perspective on the Role
of BioWatch in the Decision-Making Process

Local and state public health agencies sit on the front lines of the BioWatch system. Any autonomous detection system will have to meet the needs of those agencies in order to instill confidence in the results from such a system and to enable public health officials to make the appropriate decisions following declaration of a BioWatch Actionable Result (BAR). To provide the workshop with a public health perspective, Sandra Smole, director of the Division of Molecular Diagnostics and Virology in the Bureau of Laboratory Sciences at the Massachusetts Department of Public Health, summarized the key points she made in the commissioned paper that she had written as background for this session’s discussion (see Appendix F). Five panelists then discussed the role of public health officials and laboratorians in the interpretation of BioWatch results.

Session moderator Suzet McKinney, deputy commissioner of the Bureau of Public Health Preparedness and Emergency Response in the Chicago Department of Public Health, said that the panelists had been asked to consider two questions when they prepared their remarks:

1.   What factors related to BioWatch autonomous detection will affect decision-making response actions after a BAR is reported?

2.   How will jurisdictions determine if action steps are indicated in response to a BAR, and what should those action steps be?

This chapter summarizes the presentations and the ensuing general discussion among the workshop participants. Key points include

•   The technical data that lead to a BAR are just one part of the information that is needed; context is important for decision making.



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



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 23
3 Public Health’s Perspective on the Role of BioWatch in the Decision-Making Process Local and state public health agencies sit on the front lines of the BioWatch system. Any autonomous detection system will have to meet the needs of those agencies in order to instill confidence in the results from such a system and to enable public health officials to make the appropriate deci- sions following declaration of a BioWatch Actionable Result (BAR). To provide the workshop with a public health perspective, Sandra Smole, direc- tor of the Division of Molecular Diagnostics and Virology in the Bureau of Laboratory Sciences at the Massachusetts Department of Public Health, summarized the key points she made in the commissioned paper that she had written as background for this session’s discussion (see Appendix F). Five panelists then discussed the role of public health officials and laboratorians in the interpretation of BioWatch results. Session moderator Suzet McKinney, deputy commissioner of the Bureau of Public Health Preparedness and Emergency Response in the Chicago Department of Public Health, said that the panelists had been asked to consider two questions when they prepared their remarks: 1. What factors related to BioWatch autonomous detection will af- fect decision-making response actions after a BAR is reported? 2. How will jurisdictions determine if action steps are indicated in response to a BAR, and what should those action steps be? This chapter summarizes the presentations and the ensuing general discussion among the workshop participants. Key points include  The technical data that lead to a BAR are just one part of the in- formation that is needed; context is important for decision making. 23

OCR for page 23
24 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH  Public health officials are in the position of making high-regret decisions in the face of uncertainty—the information provided must be accurate and reliable.  Public health stakeholders—as the end users—want to be en- gaged at all stages (design, development, and testing).  Desired features and considerations for a detection system include o faster detection time; o remote access; o instrument-specific performance data; o information on agent identity, viability (alive or dead), and susceptibility to medical countermeasures; o spatial resolution (information about release location and amount); and o archived sample data for later testing. INFORMATION NEEDED FOR DECISION MAKING In her overview of BioWatch as it relates to public health, Sandra Smole stressed that BioWatch is viewed in the public health community as a surveillance tool that is used in conjunction with other tools, such as syndromic surveillance, for detecting the presence of an infectious agent or toxin of public health significance. In that respect, information from BioWatch contributes to and is interpreted within the context of the larg- er surveillance picture. As was mentioned in an earlier presentation, no two BioWatch jurisdictional operations within the 30-plus locations where BioWatch is permanently deployed are exactly alike. As a laboratorian, Smole said, she sees the definition of a BAR as hav- ing two components (see Figure 3-1). In the laboratory setting, a BAR is the production of a reliable, reportable laboratory result, backed by a laboratory- quality system. In the operational setting, a BAR is a determination that a result is actionable within the context of other key pieces of information. She said that autonomous detection technology has the potential to improve the depth of information provided by the BAR by, for example, improving pathogen identification and strain discrimination capabilities. It also has the potential to provide a quantitative BAR result that offers useful information for those considering the response to multiple collector “hits” and develop- ing an appropriate response, particularly with respect to indoor and transpor- tation venues.

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 25 FIGURE 3-1 Two components of a BAR: laboratory and operational. SOURCE: Smole presentation, June 25, 2013. Smole said that from a public health perspective, the biggest issue stemming from the use of autonomous detection systems will be the need to manage a new timeline for response actions that will result from an increase in testing frequency. Any new technology will have to come with a robust quality-assurance program and a secure data-management infrastructure that will be capable of handling input from each instrument within a more extensive network, Smole noted. She continued by saying that an autonomous detection system will need to provide flexibility in the panel of agents that it can detect; to allow for varied collection dura- tion or shorter time between samplings, particularly after a BAR; and to be able to store a portion of a test sample for additional verification by other test methods. It will be important for local jurisdictions to be able to perform basic maintenance of the detectors. Concerning the information needed to determine a BAR from a new autonomous detector, Smole said that the public health laboratory direc- tor or designee will need prior knowledge of parameters such as sensitiv- ity, specificity, reproducibility, robustness, and inherent limitations that will be critical for interpreting test results. Once an instrument provides a positive signal, the public health laboratory director or designee will need access to instrument performance indicators, positive and negative controls, threshold settings, and historical data from the detectors. She noted that while BAR data may indicate a detection event, they will still require interpretation within the context of other information.

OCR for page 23
26 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH Smole also said that any new technology should be benchmarked against the current technology. It is important to remember, she said, that data from the current system have served to establish a baseline for sensi- tivity, specificity, reproducibility, and an understanding of the inherent limitations of the current technology as far as interpreting and acting up- on results. “There is a specific DHS [Department of Homeland Security] protocol in place on our BioWatch portal that describes the mechanism for introducing a new platform and/or method to the BioWatch Program. It should be assumed that the new technology must perform equal to or better than the current technology before it is adopted.” As a final com- ment, Smole added that, in her experience, public health is supportive and poised for adoption and implementation of new technologies to pro- tect the nation’s health. FOUR REGIONAL PERSPECTIVES Houston, Texas, and Surrounding Harris County Two of the strengths that public health agencies bring to the table are their versatility and their ability to make decisions even when sufficient information is not available, said David Persse, emergency medical sys- tem physician director in the City of Houston’s Public Health Authority, who serves as the BioWatch Advisory Committee (BAC) chair from Houston. It is important, he said, that the technology community under- stand the nature of the job that he and his counterparts in public health take on regarding BioWatch, particularly the time pressure they will be under to make critical decisions from the moment they receive a notifica- tion from the laboratory that there is a BAR. From the time Persse is no- tified of a BAR, he has an hour to gather further intelligence; put the BAR in context, given that intelligence; and arrange a phone call with the appropriate local officials, including the local office of the Federal Bureau of Investigation (FBI) and the Joint Terrorism Task Force, his colleague Umair Shah from Harris County, and the jurisdictional coordi- nator from the state of Texas, who will also need to have gathered infor- mation by the time of the call. During that call Persse will solicit input from his colleagues, a key benefit of the BioWatch program, which en- courages collaboration between different agencies, but the final decision as to whether the BAR signal indicates a threat to public health rests with him. “This is not a group decision or a democracy,” he said, “there is a

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 27 health authority that is responsible for the lives of the people within that jurisdiction.” One hour after the local call, Persse and the other appropriate offi- cials in the jurisdiction prepare for the national call. “At the end of that phone call, we’re either going to do something or not do something [e.g., obtain additional samples, temporarily close facilities, alert local hospi- tals, begin prophylaxis, etc.], but at that point, from our perspective, BioWatch ends, and the response begins.” Today, Persse and his colleagues in public health have a degree of comfort with the current Generation 1 and 2 system because of their fa- miliarity with the way that the cycle threshold (Ct) values from the PCR assays are reported and with the peculiarities of each of the portable sampling units (PSUs) located in their community. A Ct value indicates that a certain DNA sequence (representing a target organism) is present at a given location—and a higher Ct value corresponds to lower concen- trations of the agent detected—but how it got there and what it means must be interpreted in the context of other information. In Houston, for example, some of the portable sensor units are more active than others because of an organism that is endemic to the area, and some are affected by wind and weather patterns or even the season or time of day. All of this information helps Persse put a BAR in context (see Box 3-1). In a perfect system, Persse said, BioWatch would identify the organ- ism, determine whether it was alive or weaponized, and provide forensic information that would pinpoint who released it—something that he acknowledged was not realistic. “The BioWatch system is much like a clinical laboratory test,” Persse said. A clinical lab test provides a set of numbers that have to be interpreted in the context of the individual patient, regardless of how sophisticated the test is or which technology was used to run the test, and the same is inevitably going to be true for BioWatch, no matter how sophisticated it becomes, he noted. “As much as I would like for an autonomous detection system to be able to tell me that we have a terrorist attack with this organism, this much of it distrib- uted at this time, in this location, that’s not going to happen,” Persse said. “I have to take the information that’s given to me, put it into context, and move forward.”

OCR for page 23
28 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH BOX 3-1 Variables Analyzed to Put a BAR in Context  BAR—yes or no?  Agent detected  Number of collectors positive  Day, month, year  High, low, and average temperature  Heating degrees (65 − Avg = # of heating degrees). Use when the av- erage temperature (Avg) is less than 65.  Cooling degrees (Avg − 65 = # of cooling degrees). Use when Avg is greater than 65.  Departure from normal temperature, degrees Fahrenheit  Precipitation for the day, inches  Precipitation total up to and including the day for 2, 3, 4, 5, 6, and 7 days prior  Snowfall depth (?)  Fraction of daytime minutes with sunshine (1 to 10 scale) o 0 = overcast conditions all day o 5 = partly cloudy o 10 = sunny all day  Average and maximum wind speeds  Average/resultant wind direction  Gust wind direction  Degrees of change in wind direction  Weather codes (thunderstorm, mist, haze, fog, etc.) NOTE: BAR = BioWatch Actionable Result. SOURCE: Persse presentation, June 25, 2013. Art of Decision Making Umair A. Shah, executive director of the Local Health Authority, Harris County (Texas) Public Health and Environmental Services, stated that although the workshop was focused on the science of biodetection, he wanted to address the art of decision making. “This is not just a sci- ence, but it’s what we as clinicians and public health practitioners do all the time, which is really taking the contexts, the nuances, and making that part of our decision-making process,” he explained. “Public health decision making is still an art based on the experiences of the individuals and the agencies that are part of the process and done in the contextual framework of the available information.” In Harris County, which is the third most populous county in the na- tion and one that encompasses a diverse set of communities in an area the

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 29 size of Rhode Island, making a decision after being notified by Houston public health of a BAR requires considering those detection data, along with information from disease surveillance and contextual intelligence. Disease surveillance includes examining zoonotic patterns reported by local veterinary clinics and the state zoonosis surveillance system as well as data on human disease patterns that may have been reported by area hospitals or other health departments. Contextual information includes details about environmental patterns and unusual security threats or secu- rity patterns. While the decision-making process is occurring, the county begins mobilizing its crisis risk communication resources and makes sure that its operational support functions are ready. County officials also take fiscal constraints into account when making a decision on how to re- spond to a BAR. Fiscal constraints, Shah said, are having a real impact on the value proposition today. For example, a decreased investment in the technologies that enable syndromic surveillance and automated dis- ease reporting, combined with decreased staffing for surveillance, is di- minishing response capabilities, which in turn makes the decision on how to respond to a BAR even more art than science. Shah concluded his remarks by offering some considerations for those in the audience who are involved in developing and deciding on Generation 3 or other future BioWatch technologies. The first thing to think about, he said, is that a laboratory positive is not the same as a pub- lic health positive, and the issue of false-positives is likely to be a bigger issue with a new autonomous detection system with more cycles, more tests, and more results on a daily basis. Second, a biodetector that has the capability to automatically signal a BAR without human input or addi- tional context (red light/green light) may be appealing from a technology perspective, but from the public health perspective such a feature would take away the ability to engage in nuanced decision making. It is also important, he said, to remember that the integrity of public health is critically important. “How does the public view us if we do launch or don’t launch a response based on incorrect or incomplete in- formation?” Shah said. He added that he shares Persse’s view that the BioWatch BAR system is like a clinical laboratory test. As a clinician, if I had a woman who walked through the door and I said, “You have a spot on your mammogram, and I’m going to imme- diately send you for a bilateral modified radical mastectomy, based on that spot,” immediately, my days as a physician are numbered. That is the challenge here. What we are really trying to do is take that spot on a

OCR for page 23
30 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH mammogram, figure out what other diagnostic and contextual infor- mation we need to put to the puzzle, and then figure out what to do with that information. Persse said he strongly agreed with that view, adding that the only currency public health has is credibility. The bottom line, Shah said, is that any new technology must make public health more effective. The San Francisco Bay Area The San Francisco Bay area covers more than 8,000 square miles and is inhabited by a highly mobile population of 7.2 million people liv- ing in both urban and rural communities, said Erica Pan, deputy health officer and director in the Division of Communicable Disease Control & Prevention, Alameda County (California) Public Health Department, adding that San Francisco itself is one of the smallest counties in the re- gion with a residential population of less than 800,000 that doubles dur- ing the day. In total, the health jurisdictions of six counties, including San Francisco, and the city of Berkeley are included in the San Francisco Bay Area BAC. One unusual feature of the Bay Area is that the regional BioWatch laboratory is the California state laboratory, which is not in the Bay Area region. Furthermore, while San Francisco does not have any indoor detectors, San Francisco International Airport does, and a repre- sentative of the airport is also a member of the advisory committee. The BAC has developed a decision matrix to discuss during the local conference call and guide its response to a BAR (see Table 3-1). The matrix, developed from local experience, ranks the likelihood of a poten- tial problem organism in the environment after a BAR. For example, a BAR resulting from the detection of Francisella tularensis with a Ct val- ue greater than 35 by a single detector is rated as an unlikely release event because F. tularensis is an indigenous organism that has been de- tected previously in this BioWatch region. By contrast, a BAR resulting from a single detector identifying an organism not previously seen by BioWatch in the region would be rated as a probable release event, as would a BAR resulting from the detection of F. tularensis with a Ct val- ue less than or equal to 35. A BAR would also be judged to be a probable release event if additional collectors provided positive signals or if high concentrations of the organisms were detected. At that point, Pan ex- plained, the advisory committee looks for other information, such as the

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 31 TABLE 3-1 San Francisco BioWatch Advisory Committee Decision Matrix Likelihood of Example of BioWatch Actionable Result (BAR) Organisms in the Environment Unlikely Single collector positive of organism not previously detected by BioWatch in the region, or F. Tularensis (FT) with a Ct value >35 Possible Single collector positive of organism not previously detected by BioWatch in the region, or FT with Ct value ≤35 Probable Single collector positive of organism not previously detected by BioWatch in the region and the following information may in- crease the risk of release: a. Additional collectors e. Overt unusual/adversarial activ- with confirmed positive ity consistent with threat signals f. Evidence of weaponized or b. High concentrations of genetically modified agent agent detected g. Recent bioterriorism incident or c. Positive environmental credible threat in the United surface swipe sample(s) States d. Recent outbreak of hu- h. FBI-designated special event man or animal cases in near positive collector the region SOURCE: Pan presentation, June 25, 2013. NOTE: Ct = cycle threshold; FBI = Federal Bureau of Investigation. presence of positive environmental surface swipe samples, recent outbreaks of human or animal disease in the region, overt unusual or adversarial activi- ty consistent with a threat, evidence of a weaponized or genetically modified agent, a recent bioterrorism incident or credible threat in the United States, or an FBI-designated special event near the positive collector. Decision making is aided by a detailed list of organism-specific cri- teria that account for factors such as incubation period, transmissibility between humans, past incidence of human disease, the likelihood of nat- ural environmental occurrence in the region, and the potential risk to public health. At that point, the BAC orders immediate environmental testing, using a plan developed by its Environmental Protection Agency (EPA) and FBI partners, and it may declare a potential local emergency while waiting for input from other state and federal agencies. Pan said that the BAC has discussed increasing the frequency of retrieving filters from detectors, but it decided that such an action would not be useful

OCR for page 23
32 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH with current technology and given the limited laboratory capacity for testing. Other decisions that will be made on a case-by-case basis include how to conduct public health surveillance, either with enhanced passive techniques that involve sending out a health alert to emergency medical staff and clinicians to be on the alert for symptoms of a certain disease or via active surveillance with teams visiting hospitals to search for cases. Although veterinary surveillance would also be useful, Pan said that the region has few resources in this area; this is one area in which the BAC would like to expand the available resources. The advisory committee will also need to consider post-exposure prophylaxis for the filter retrieval staff and local emergency responders, Pan said, and it will have to alert the personnel that will be needed to coordinate a potentially large medical surge. Another important decision will be whether to issue an alert to the Strategic National Stockpile and to regional staff that would be needed to distribute medical countermeas- ures. Pan said that the Bay Area BAC has had internal discussions as well as talks with other advisory committees around the country about what kind of exposure reduction recommendations it would make. She noted that her committee has decided that the distribution of post- exposure prophylaxis will take priority, and it has worked closely with the Bay Area Mass Prophylaxis Working Group to create a unified screening algorithm for all of the region’s points of distribution and also an associated website that would go live immediately after a release event was confirmed. Thoughts on Future Technology Regarding the impact of newer technology, Pan reiterated earlier re- marks that it will be essential to build confidence in the results that is at least equal to the comfort level that public health has developed with the current technology. “We would like to see equal sensitivity and specifici- ty and some ability to have some oversight or understanding of the quali- ty control of any autonomous detection,” she said. More frequent testing periods would provide more timely notification and potentially reduce the window of likely exposure and the number of people exposed, which in turn would reduce the demand for prophylaxis. This would be particu- larly true in the case of indoor detectors, where faster detection would decrease how far someone who was exposed would have traveled. One

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 33 potential downside of more frequent testing may be an increased possi- bility of false-positive test results. Pan concluded her remarks by listing other information that new technology might be able to provide that would help public health’s abil- ity to respond to a BAR. Her wish list included  quantitative results;  the ability to distinguish between pathogenic species, closely re- lated non-pathogenic species, and environmental species;  rapid confirmation of viability;  antibiotic susceptibility;  indications of weaponization or intentional genetic modification;  feasible decontamination strategies; and  the flexibility to include or remove organisms from the screening panel. Pan noted that public health has been calling for studies to identify endemic organisms that make up the environmental background. “We have found some of them incidentally through this program,” she said, “but it would be much more useful if we knew for sure the background data for certain endemic organisms in different areas.” She also reiterated Shah’s earlier remarks about the impact of budget constraints on the public health infrastructure and said that because of the dramatic cuts to public health funding, it is difficult to find qualified people to fill positions that are available. Developing an anthrax vaccine does no good, she said, if there is no one available to distribute it. New York City In his presentation, Colin Stimmler, director of the BioWatch Pro- gram, New York Department of Health and Mental Hygiene, said that it is important to remember that the goal of BioWatch is not just to detect an airborne biological agent for the sake of detection. “It is detection to respond, to hopefully save lives,” he said. In New York City, only 40 percent of the BioWatch portable detection units are outdoors, with the remaining 60 percent in indoor locations. Another unusual feature of the New York City BioWatch network is that it is administered by the New York City Police Department, with the Department of Health and Mental Hygiene serving as the lead scientific agency responsible for testing, en- vironmental sampling strategies, disease surveillance, and implementing

OCR for page 23
34 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH the public health response plan. All told, there are 16 agencies in the New York City Stakeholder Group, that area’s version of a BAC. Indoor Detection New York City has had indoor detectors since 2003, and their loca- tion is based on police department threat assessment, not on the percent- age of the population covered. A sizable percentage of the indoor detectors are in transit hubs that are managed by seven stakeholders, Stimmler said, with some of those stakeholders reporting to regional au- thorities and not New York City. Regardless, he said, all of those stake- holders must be integrated into planning and also into any response. Stimmler also noted that filters are collected multiple times per day from the indoor transit hub detectors. To put the importance of these indoor collectors in context, he noted that the busiest airport in the world—the Atlanta international airport—processes about 250,000 people per day, while three times that number pass through New York City’s Grand Cen- tral Station daily, and more than 5 million people use the New York City subway system every workday. With that information as background, Stimmler presented a scenario as a means of illustrating the challenges that New York City public health faces because of its extensive indoor program. In this scenario, which he characterized as the one that is most likely to present him and his colleagues with the necessity of making an immediate, potentially high-regret decision, the public health laboratory reports a BAR for a biological threat agent from an indoor detector in a major transit hub. At that point, a number of things are known: the number of detectors report- ing a BAR, the location of the detector or detectors reporting a BAR, the Ct values, the sample period range for each detector, the performance of the detectors in terms of sensitivity and specificity, and syndromic sur- veillance data from the previous day. What will not be known is whether there has actually been a release, if the organism is viable or not viable, and, if it is not viable, whether it is dead because it was dead upon re- lease or because it died from desiccation on the filter. Other unknowns will include  whether the release is an isolated incident or part of a series of incidents;  when or where the release occurred;  how and how much of the material was released;

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 35  where the material has migrated since release;  whether the material was altered or weaponized and o whether it is more or less virulent, o has a longer or shorter incubation period, o is resistant or susceptible to medical countermeasures, or o if the size of the particles makes them more conducive to transport; and  whether there has been an error in the analysis of the sample despite all of the safeguards to prevent such a mistake from happening. With those gaps in mind, there are many possible explanations for a single detector reporting a BAR. One possibility is that there was a large, point-source release that then dispersed and became a series of line- source releases. This situation has the potential for wide dissemination, and a possible explanation for only a single detector having reported a BAR is that this one detector is identifying the edge of the large-scale release. Another, far more benign possibility is that a vendor selling products made from animal hides set up his or her display right in front of the portable detector. These two possibilities have two vastly different consequences, yet there will be no way of knowing if either is true when public health has to report a BAR. In fact, Stimmler and his colleagues assume that the only information available at the time of reporting a BAR will be the laboratory results. The stakeholder group’s initial response in this scenario will be to seek answers to four questions: Has there been a release of a biological threat agent? If there was a release, where, when, and how did the release occur? Who might be at risk of infection? How can the stakeholder group target the available public health interventions to protect the greatest number of potentially at-risk people? Answering the first question, Stimmler said, leads to the default hypothesis upon which all other ac- tions will be based, but given the limited information available at that moment, he said, “all of our actions are potentially high-regret and are instantly public.” The unique challenge of having an indoor program is that there will have to be a decision about closing the facility in which the signaling detector is located, Stimmler said. To help with that decision, environ- mental sampling teams in protective gear will make an additional collec- tion of samples with surface wipes and a retrieval of the detector filters— a very public display that is unavoidable. Stimmler then made what he characterized as the understatement of the year: “If we get this wrong

OCR for page 23
36 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH and we overreact, the economic and political impacts are huge. If we get it wrong and underreact, the public health impacts of doing nothing are even bigger.” With that scenario and its consequences in mind, Stimmler discussed how BioWatch Generation 3 or autonomous detection would change the situation. Currently, the time between potential release and detection by New York City BioWatch ranges from 9.6 to 37 hours, which means the system is basically one that detects in order to treat. Autonomous detec- tion could reduce that time to 2 to 4 hours, which would turn the system into one that detects to protect. An autonomous system could make it pos- sible to respond quickly enough to limit additional exposures and make it easier to estimate who was affected by the release. However, any new technology needs to have as close to zero percent false-positive and false- negative rate as is possible. Because of the scale of the New York City BioWatch network, even a 0.01 percent failure rate—which Stimmler said would translate to two false-positive BARs per year—would be unac- ceptable, given the consequences. Moving into the realm of what he acknowledged might be science fiction, Stimmler provided a public health wish list for Generation 3 or autonomous detection technology that was very similar to the list Pan presented. Additional items on his list included the ability to resolve the time (in minutes) and the location of the release as well as estimates of the aerosol mass that was released. Chicago The Chicago Transit Authority (CTA), the second largest transit sys- tem in the nation, moves 1.6 million people per day among 145 rail sta- tions and 11,500 bus stops. The CTA has its own BioWatch response plan because, as John Plante, senior manager of emergency preparedness for the CTA, described the situation, the system’s 1,200 railcars are not only a target but could also serve as a vector for transmitting a released bioweapon. The system connects the city; contamination could spread not only within the subway, but to surrounding streets, businesses, and the airport (see Figure 3-2). This indoor response plan is based on a number of assumptions, including that every one of those railcars will have to be inspected, that detection in the subway system might result in cessation of operations for a period of time, and that detection in the sub-

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 37 FIGURE 3-2 The CTA rail map. SOURCE: Chicago Transit Authority. way system may necessitate decontaminating not only the railcars, but also the platforms, ancillary rooms, and other structures. Once a decision is made that affects the subway system, the response will include the management of patrons, management of employees, management of the transportation system, management of the facilities involved, and risk communication. Plante said that the city has practiced an evacuation scenario and has consulted with both the Department of Defense and the EPA on how to test and decontaminate a railcar. He said

OCR for page 23
38 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH that the mayor holds the final authority to declare an event but that the CTA, with advice from the city’s department of public health, maintains control of the subways throughout the closure and evacuation process. The CTA, he added, will follow the advice of the Chicago Department of Public Health regarding whether or not to keep the rail system running. Plante reiterated what the previous panelists had said regarding the need for an autonomous detection to be as accurate and reliable as the current system. The key for any new technology, he said in closing, will be to provide fast, accurate information that will enable the end users— managers such as himself—to make a decision about what is happening, how bad it is, what to tell customers and employees, if a facility will need evacuation, and how and when the affected facilities can be reopened. DISCUSSION To open the discussion, Eric Gard, a scientist with the Defense Biol- ogy Division at Lawrence Livermore National Laboratory (LLNL), asked the panelists to talk about when they would use atmospheric mod- els of potential exposure regions and backward models to establish the point of release. He also asked them if current information limits the util- ity of those models in either the outdoor or the indoor environment. Persse replied that these models are used early on, almost as soon as the initial notification of a BAR is received. He said that his team has a mechanism in place to feed data to the team’s federal partners to start the modeling process and added that modeling might be more of a benefit to law enforcement than to public health because in the event of a release, everyone who wants prophylaxis is going to receive it. Pan agreed with those comments but thought that modeling may help if it comes to priori- tizing who gets vaccine in the case of supply limitations. Raymond Mariella, Jr., from LLNL noted that all of the panelists mentioned the desire to determine if an organism was alive or dead when detected and asked if there is a particular time frame that would be useful for making that determination. Smole said that it is possible now to pro- duce that information in 24 to 48 hours, but her impression was that the information would be more useful if it was available in a matter of a few hours. Stimmler said it would be most helpful if the time frame was minutes rather than hours because that determination affects the decision to release medical countermeasures and trigger procedures to prevent the spread of an infectious agent.

OCR for page 23
PUBLIC HEALTH’S PERSPECTIVE 39 In response to a question by Thomas Slezak from LLNL about net- working an autonomous system, Smole said that it would be important for the laboratory to be part of any such data communications network, and Persse said that any network would have to be transparent. “If there was a delay between when the feds found out and when the locals found out, you are going to create an air of mistrust,” he said. Plante added that it was essential that any network be robust and secure. Robert Kadlec asked about the importance of including the ability to detect antibiotic resistance in future BioWatch developments, and Pan replied that antibiotic resistance is a basic and important piece of infor- mation that decision makers need when they make a request to the Stra- tegic National Stockpile. Shah agreed with that comment and added that such information allows public health to make a more nuanced decision in a timely manner. Session rapporteur Beth Maldin Morgenthau, assistant commissioner for the Bureau of Policy, Community Resilience and Response within the Office of Emergency Preparedness and Response at the New York City Department of Health and Mental Hygiene, voiced her concern that a system that produces results every few hours could strain what are al- ready limited and diminishing public health resources and said that the BioWatch program needs to keep that in mind as it contemplates rolling out the next version of this technology. She also raised the point that alt- hough everyone agrees with the need to have flexible panels whose bio- logical agents can be added or removed as needed, it will be necessary to have a response plan in place before adding a new agent to the panel. Again, with diminishing resources, it is important to consider how public health will respond to new agents. Maldin noted, too, that as BioWatch is deployed in additional indoor settings, it will be important to remember that the public will be more aware of the system when public health moves into a setting to gather additional information needed to declare a BAR. She then reiterated Plante’s points about the difficulty in cleaning up after a release and the difficulty of determining when to declare a con- taminated area clean. Maldin also said that she agreed with Smole’s and Persse’s com- ments about the need for transparency with any networked system. “Our labs should look at the results first, then public health should get to talk about the results before anyone else sees them,” she said. Shah noted that ultimately the decision of how to respond to the release of a biological weapon must be a local one. “Our citizenry, our residents, expect local governance and local decision making,” he said, which implies both a

OCR for page 23
40 TECHNOLOGIES TO ENABLE AUTONOMOUS DETECTION FOR BIOWATCH need for transparency and a need for local public health offices to man- age the data from a networked system. Donald Prosnitz, an independent consultant, asked the panelists if they would take BioWatch into their jurisdictions today, given what they know now and the experiences they have had with BioWatch. Plante said that he would because, for all its limitations, BioWatch is still better than syn- dromic surveillance in terms of shortening the time for detection and sav- ing lives. Stimmler said yes as well but added that he would want to have more input in terms of the agents being tested for and the logic behind those choices. Pan agreed, but said she would insist on more local in- volvement at the beginning of the rollout process. She added that resource constraints today would make BioWatch seem like an unfunded mandate, a sentiment that McKinney seconded. Persse agreed that BioWatch was valuable and said that he would have it in his jurisdiction again because, in addition to the information it supplies, it has also created a network among state, local, and federal officials that is valuable, a comment that Smole seconded. Shah was the most reluctant, saying that at the time BioWatch was originally deployed, there was no good understanding of what a BAR meant. However, assuming that state and local officials have more input and information as future programs are deployed, he said he is looking forward to the autonomous detection system. He noted that people need to view it in the context of all the tools available in the surveillance toolkit and also recognize that resources are limited.