Sandra Smole, Ph.D.
Director, Division of Molecular Diagnostics and Virology
Massachusetts Department of Public Health
Bureau of Laboratory Sciences
A white paper prepared for the June 25–26, 2013, workshop on Strategies for Cost-Effective and Flexible Biodetection Systems That Ensure Timely and Accurate Information for Public Health Officials, hosted by the Institute of Medicine’s Board on Health Sciences Policy and the National Research Council’s Board on Life Sciences. The author is responsible for the content of this article, which does not necessarily represent the views of the Institute of Medicine or the National Research Council.
OVERVIEW OF BIOWATCH AS IT RELATES TO
PUBLIC HEALTH OFFICIALS AND
The BioWatch Program was put in place in 2003 to monitor the threat of an aerial release of a high-consequence infectious agent within a populous area. It is a first-of-its-kind, bioaerosol monitoring network that spans federal agencies from national security, public health, and environmental protection. The program, overseen by the Department of Homeland Security (DHS), works closely with the Centers for Disease Control and Prevention (CDC), Federal Bureau of Investigation (FBI), Environmental Protection Agency (EPA), and, more recently, with the Department of Defense. While it is managed by the DHS’s Office of
Health Affairs, BioWatch itself is maintained and operated at the local level as a unique entity within each jurisdiction. The decision makers within these jurisdictions represent a composite from municipalities, county, state, and regional authorities and may include both public and private entities. No two BioWatch jurisdictional operations, within 30-plus locations, are exactly alike. In general, laboratory testing and much of the program’s operational response lie within public health as the lead entity in determining response to an infectious disease event with potential for high-consequence impacts. In response to an intentional release, a strong, collaborative partnership with law enforcement is required to coordinate parallel criminal and public health investigations. Partnerships with FBI and other local law enforcement have been strengthened by the BioWatch Program.
Within the public health community, the BioWatch Program is viewed as one surveillance tool among many; another such tool is, for example, syndromic surveillance, used for detecting the presence of an infectious agent of public health significance. The uniqueness of the BioWatch Program is in its ability to act as an environmental early warning system detecting the presence of a specific agent prior to the appearance of significant human disease. It contributes to the larger surveillance picture. There is a broad array of opinions regarding the need for such an early warning system. A system identifying rare events measured against resources necessary for national public health surveillance systems that identify many public health threats is bound to generate controversy among public health professionals. It is difficult to measure the value of such a system until it has been triggered by the purposeful release of a biothreat agent.
WHAT IS A BAR? WHAT ANALYSIS OR ACTIONS MAY BE TAKEN AS A RESULT OF A BAR?
In oversimplified terms, a BioWatch Actionable Result (BAR) is a laboratory test result indicating that multiple nucleic acid signatures specific to an individual biological agent have been detected in the environment. Prior to determining a BAR, a two-tiered testing process occurs: (1) screening samples for a single signature per agent and (2) confirming any reactives by testing for additional signatures for that agent. None of the results indicate that a viable, or live, organism has been detected. The BAR leads to the public health laboratory director, or a designee, review-
ing laboratory test data and determining whether all internal controls and testing processes are consistent with a “true positive”—that is, the detection of the biological agent. This review includes ensuring that the result is not a laboratory-generated error due to a cross-contamination issue or reagent failure. The laboratory director, or designee, has the option to contact the CDC’s Bioterrorism Rapid Response and Advanced Technology Laboratory and DHS’s BioWatch Special Program Office (BWSPO) for an outside technical consultation prior to determining a laboratory BAR. This might be done, for example, if there is a known reagent or instrument issue that may impact the result interpretation. Ultimately, the verification or “sign off” of the result by the laboratory director, or designee, assigns professional responsibility in ensuring reliability of the test result.
Once a laboratory BAR has been determined, a number of rapid local and national notifications occur within the jurisdiction, including notification to convene a conference call within 2 hours to discuss the laboratory BAR with the local BioWatch Advisory Committee (BAC), or a BAC subset with key representation. Early in these notifications, the jurisdiction’s local FBI weapons of mass destruction coordinator is given a heads-up. Resources may be sought to initiate a request to the BWSPO to seek plume modeling at the national level; some locations use local resources, such as the National Guard’s civil support team or other agencies, as a supplementary resource for modeling. In many jurisdictions a local BAC call occurs just prior to the BioWatch national conference call. Some jurisdictions have indicated that the notification time frame is complicated to manage if the BAR is received after work hours or during evening commutes. The BAC chair, typically a public health official, coordinates the jurisdictional BAC conference call. The agenda is quickly covered using a customizable conference call script provided by the BWSPO. The information reviewed includes a concise presentation of the laboratory BAR data, including (1) assurance that the laboratory has reviewed its internal quality testing checklist and (2) a quick primer on how to interpret the data including its limitations, followed by (3) the actual laboratory details, such as number and location of the BioWatch collectors that were positive, the organism detected, the cycle threshold (Ct) value for each collector, and the weather conditions for the preceding 12 to 48 hours.
The BAC chair guides the conference call by requesting supplementary information, such as available intelligence information, presence in the jurisdiction’s environment, and any human or veterinary surveillance
data related to the organism detected. At this time a DHS BWSPO request for source reconstruction and plume modeling may already have been initiated, and it is understood that these national laboratory resources require a 6- to 8-hour lead time. Modeling data are used primarily to assist with environmental sampling decisions. A rapid assessment of the perceived threat and the public health risk is performed based on the information in hand. The BAC then decides on a course of action for the following: (1) a decision to initiate environmental sampling; (2) a decision to alter the sampling intervals for the collections (i.e., collecting every 6 to 12 hours versus every 24 hours) or add additional collectors; (3) a decision to initiate a finer focus on syndromic surveillance system results based on the agent detected; (4) identification of the need for federal resources (e.g., support for environmental sampling or testing, initiation of strategic national stockpile resources, if known); (5) implementation of public information and messaging plans; and (6) estimation of the time for the next BAC conference call. The BioWatch national conference call occurs immediately following the local jurisdictional BAC call and begins with a summary by the BAC chair of the current situation, follow-on actions, requests for federal assistance from the various agencies (DHS, CDC, FBI, EPA, or the strategic national stockpile) and a decision regarding the next conference call time.
BENEFITS, LIMITATIONS, AND ISSUES RELATED TO THE GEN-2 DATA GENERATED BY THE CURRENT BIOWATCH PROGRAM
Within the current Biowatch Program, data are generated on a 24-hour cycle by laboratory staff. These data consist of either positive (Ct values) or negative results (not detected, i.e., below the signal threshold) for nucleic acid signatures representing a panel of select biological agents. Millions of results have been generated across the country from a variety of geographic locations using the current Gen-2 biodetection technology: real-time detection polymerase chain reaction (RTD-PCR). A positive aspect of the current RTD-PCR technology is that positive results (Ct values) are semiquantitative.1 The Ct value is a numerical value indicative of the relative concentration of nucleic acid detected in the
1BioWatch Portal: BioWatch laboratory assay.
sample. This information has been useful when considering response actions following a single- versus multiple-collector “hit.”
Data generated by the BioWatch Program represent samples collected from a variety of environments (primarily outdoor, but with some limited indoor locations) and spanning the full seasonal variation across the country. The depth and experience with these data have served to establish an important baseline for sensitivity, specificity, reproducibility, and an understanding of the inherent limitations of the current technology in interpreting and acting upon results. One very important aspect of the current Gen-2 system and reagents is that they have not generated laboratory test false-positives but they have detected agents from the panel in the environment, which has operationally complicated the response to a BAR (IOM and NRC, 2011). Additionally, implementation of a robust quality assurance program by the BWSPO2 in 2010 has resulted in data supporting the reliability of the results and engendering additional confidence. Specifically, qualified laboratory testing staff must demonstrate ongoing individual competency as well as participate in overall laboratory proficiency assessments. Each laboratory is responsible for maintaining all of the components of a comprehensive quality assurance plan (standardized methods, staff training, proficiency tests, stringent record keeping, equipment maintenance, corrective action protocols, etc.) to ensure that high-quality results are reproducibly obtained. The BioWatch Program should be recognized for its significant effort in partnering with each public health laboratory to ensure high-quality results from that BioWatch laboratory.
As indicated earlier, the periodic detection of some of the biothreat agents on the test panel has occurred due to their presence in the natural environment. Several agents in significant enough concentrations in the environment have been detected by the BioWatch network within several geographic locations across the country. While these detections were still “true positives” from the laboratory perspective, they represent laboratory test data that needed to be interpreted in the context of other information. In the first example, one agent (Brucella sp.) was detected on multiple occasions in the environment, but the organism was later deemed not to be of sufficient public health concern, and monitoring for this agent was suspended early in 2008.3 The second example involved the detection of a closely related subspecies of the agent of tularemia
2BioWatch Portal: DHS BioWatch Program quality assurance program plan.
3BioWatch Portal: DHS memorandum: Recommendation to discontinue monitoring for Brucella species from biomonitoring effort.
(Francisella tularensis subspecies) not considered to be a human pathogen and therefore not of public health concern. At the time, the laboratory test reagents within the test panel were not specific enough to differentiate between virulent and avirulent subspecies of this organism. Since this event BioWatch has introduced an additional set of laboratory reagents to rapidly differentiate between virulent and avirulent subspecies.4 These two examples highlight the need for flexibility in the panel of agents detected—to improve upon the performance of a reagent, remove the organism from the panel, or adjust a local threshold—if repeated detections in the natural environment act as a “red herring.”
One limitation of the current system has been the length of the turnaround time—the time between when the air sample collection begins and a laboratory test result is reported. Essentially, the existence of the turnaround time translates into the possibility that an aerial release could have occurred any time within a 10- to 36-hour window. For the purposes of accurately pinpointing when an exposure might have occurred, shortening the turnaround time to 4 to 6 hours (or even less if the technology is capable) would provide a significant advantage for public health response. This will be particularly true if the jurisdiction in question expands detection systems to include the transportation sector (e.g., subways, commuter rail, airports). A 4-hour (or less) collection-to-result turnaround time is much more useful in the transportation sector where the exposed cohort rapidly becomes difficult to identify for the purposes of prophylactic treatment or vaccination.
One other limitation to the current technology should be noted. The BioWatch result signifies detection of an agent’s nucleic acid signature, but this does not allow the determination of “viability,” that is, that the agent is alive and infectious and therefore a risk to human health. The ability to rapidly test agent viability within the sample would be valuable. It is recognized that this may be outside the limits of current technology. Part of the response to each BAR is the collection and testing of environmental samples to determine the presence of viable organism. The culturing of supplementary environmental samples can add an additional 24 to 72 hours in determining the viability of an agent and its full public health significance.
4BioWatch Portal: BioWatch Program Francisella tularensis guidelines for responding to an outdoor BioWatch Actionable Result.
CONSIDERATIONS RELEVANT TO NEW AUTONOMOUS DETECTOR TECHNOLOGY
The BioWatch Program’s initiatives to modernize its bioaerosol detection capability, among other strategic objectives, not only are relevant to the program mission but also are in alignment with a national movement to strengthen the public health infrastructure. The capacity to rapidly identify disease threats or emergent infectious diseases has been increasing, and with it there has been a concomitant need for effective communication and response (Gargano and Hughes, 2013). Recent public health threats and active responses include those to novel Middle East respiratory syndrome coronavirus and avian influenza A (H7N9). There is an urgent need to rapidly improve and modernize laboratory-based technologies that are used to detect and characterize pathogens of public health concern within the national public health laboratory network. To ensure the public health mission of staying abreast of modern disease diagnostics, CDC has initiated a fiscal year 2014 budget request to fund the Advanced Molecular Detection and Response to Infectious Disease Outbreaks initiative.5 While this effort focuses on next-generation sequencing methods and the application of bioinformatics to public health science and practice, other methodologies, such as mass spectrometry, are rapidly emerging on the clinical diagnostics front (van Belkum et al., 2012) and being adopted within the public health laboratory network. Public health is supportive of the opportunity to adopt new technologies to protect the nation’s health.
As a replacement for the Gen-2 system, the following families of biodetection technologies are being considered for an autonomous detection system: nucleic signatures (PCR, microarrays, and other probe-based systems); genomic sequencing; immunoassays and protein signatures; and mass spectroscopy. The goal is to identify a more cost-effective, field-based, autonomous detector with acceptable performance specifications that would result in the ability to improve upon population coverage. Some of the technologies being considered perform best on samples for which the agent has been purified from a complex environment or for which the agent is of sufficient concentration to perform detection and identification. It is likely that a combination of more than one technology
5Advanced Molecular Detection and Response to Infectious Disease Outbreaks initiative, http://www.cdc.gov/fmo/topic/Budget%20Information/factsheets/AMD_Factsheet.pdf (accessed August 4, 2013).
will be necessary to achieve collection, concentration, purification, and amplication from a complex matrix followed by detection to meet the specifications required of the new system (Sapsford et al., 2008). Some of the new technologies may afford the ability to “fingerprint” the agent detected by providing strain-level discrimination. While this information is important to confirm a source, it may not add immediate value to public health response decisions unless the information improves upon the ability to discriminate between closely related species. Examples include differentiation between virulent and avirulent Francisella tularensis subspecies or the resolution of closely related species such as Burkholderia mallei and Burkholderia pseudomallei. This level of specificity would be useful in weighing disease risk and the need for prophylaxis.
While public health is receptive to adopting a new technology, many of same considerations and responsibilities for interpreting a Gen-2 BAR will still apply for the new autonomous detector technology. A simple “red light/green light” result does not provide sufficient data to a public health laboratory official to make a professional determination regarding the quality of the result. This is especially true of adopting a new technology for which the laboratory test would be performed at a remote location. There is no precedent for this in the public health laboratory realm with any technology. The closest is the U.S. Postal Service Biological Detection System (BDS) for which a “positive” result is not confirmed until the BDS cartridge is brought to the public health laboratory and confirmed using CDC’s Laboratory Response Network (LRN) reagents. To call a BAR on data generated by an instrument in the field, there must be access to data to review the following: instrument quality performance indicators (these will vary depending on the technology), positive and negative controls, threshold settings, and the actual test value (qualitative or quantitative). Other parameters are critical to test interpretation: preknowledge of the autonomous detector’s sensitivity, specificity, reproducibility (machine-to-machine, day-to-day); robustness (environment-to-environment, season-to-season); and limitations (adverse impacts of environmental contaminants, such as pollen or brake dust). It will be important for any new technology to be benchmarked against the current technology. A specific protocol is in place describing the mechanism for introducing a new platform or method to the BioWatch Program.6 It should be assumed that the new technology must be at least equal to the current technology (Gen-2 RTD-PCR) in sensitiv-
6BioWatch Portal: New method, assay, and platform acceptance testing guidance.
ity, specificity, reproducibility, and instrument reliability before it is adopted. As with Gen-2, data generated by the new autonomous detection system will still need to be interpreted within the context of other information.
Additional operational and technical considerations include the following:
• Reducing the detection turnaround time to 4 to 6 hours means that the testing frequency will be approximately 5 to 7 times more than the current Gen-2 technology. The new technology will have to perform proportionally better to not suffer the consequences of an increase in “actionable” results that turn out to be false-positives or instrument errors.
• Being able to have some flexibility in the agent panel by allowing addition of new threat agents or removal of others, should the need arise, would be useful.
• Being able to adjust the duration of the collection and testing window in response to a BAR should be considered.
• Ensuring secure and reliable electronic results from each instrument is of concern. Information security and internet stability will require redundancy in the system. Preventive measures to prevent spoofing a sensor or modification of transmitted data must be ensured.
• Replicating key aspects of the current laboratory-based quality assurance program will be necessary to ensure high-quality results from each instrument and the system as a whole.
• Other factors that will affect the placement of an autonomous detector include instrument footprint, weight, and operating noise levels.
• Colocation of environmental monitors for such measures as temperature and humidity would be useful in evaluating the impact on instrument performance from—as well as providing data on—the environmental conditions at the time of an instrument BAR.
• Saving a portion of the test sample to verify by other test methods would be highly desirable.
• Provision for the option for some jurisdictions to perform basic autonomous detection reagent loading and quality checks is desirable and would facilitate a familiarity with the technology and its inherent limitations. This “boots on the ground” feature would
provide valuable insight into technical problems and their solutions and would ensure a greater confidence in interpreting data from the autonomous detector by the responsible laboratory director, or designee.
In conclusion, the evaluation and adoption of a new technology to improve and automate the BioWatch network will be supported if the technology’s performance characteristics are equal to or better than those of the existing Gen-2 laboratory-based operation. The public health laboratory must be provided with sufficient data to be familiar with the instrument’s performance, including its limitations, in order to instill trust and confidence in the quality of the data. This is critical to ensure technical competence and expertise in interpreting the results of an autonomous detector BAR for public health officials and other key stakeholders involved in response actions.
Gargano, L., and J. Hughes. 2013. Emerging microbial threats: Communication challenges and opportunities. Microbe 8(5):205–211.
IOM (Institute of Medicine) and NRC (National Research Council). 2011. BioWatch and public health surveillance: Evaluating systems for the early detection of biological threats. Washington, DC: The National Academies Press.
Sapsford, K. E., C. Bradburne, J. B. Delehanty, and I. L. Medintz. 2008. Sensors for detecting biological agents. Materials Today 11(3):38–49.
van Belkum, A., M. Welker, M. Erhard, and S. Chatellier. 2012. Biomedical mass spectrometry in today’s and tomorrow’s clinical microbiology laboratories. Journal of Clinical Microbiology 50(5):1513–1517.