Information Sharing and Collaboration: Applications to Integrated Biosurveillance: Workshop Summary (2012)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

5 Discussion of a Scenario To highlight the complexity and importance of collaboration and information sharing, participants from a number of federal agencies en- gaged in a scenario-based discussed of a fictional disease outbreak mod- erated by William Raub. The scenario consisted of five moves detailing successive stages in the investigation of the outbreak. Participants in the discussion were Joseph Annelli, Animal and Plant Health Inspection Ser- vice (APHIS), Department of Agriculture (USDA); Christopher Braden, Centers for Disease Control and Prevention (CDC); Cory Bryant, Food and Drug Administration (FDA); Selwyn Jamison, Federal Bureau of Investigation (FBI); Donald Kautter, FDA; Teresa Quitugua, National Biosurveillance Integration Center (NBIC), Department of Homeland Security (DHS); Kevin Russell; Armed Forces Health Surveillance Cen- ter (AFHSC), Department of Defense (DOD); and Regina Tan, Food Safety Inspection Services (FSIS), USDA. The scenario posited that in 10 widely separated U.S. cities over a period of several days, patients (adults and children) arrive at hospital emergency departments exhibiting bloody diarrhea. Other children pre- sent with symptoms of anemia, abnormal bleeding, and acute kidney failure. In each city, a few of the patients die; most of the remainder require hospitalization—in some cases, intensive care. In a number of cases, particularly among those patients requiring hospitalization, stool cultures are performed. Across the affected cities, these cases of illness come to the attention of local or state health departments in a variety of ways and in varying time frames. In some instances, physicians who recognize the illness as characteristic of infection with Shiga toxin-producing Escherichia coli 41 

42 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION (STEC) contact their local health department to report their concerns about possible clusters of this disease. As stool cultures become positive for STEC over 2 to 3 days, laboratories report the positive tests to health departments, in some instances using new electronic reporting capacities. Over the following week, some clinical laboratories forward speci- mens to state public health laboratories, which perform DNA fingerprint testing (PFGE testing) and share this information with the national PulseNet system. In other instances, health department syndromic sur- veillance systems that include monitoring for patients with symptoms of “bloody diarrhea” detect an increase in such illness. As state health de- partments begin to recognize and assess this situation, because of its se- verity, they are likely to notify CDC and post their concerns on the Epi-X network (a confidential information-sharing network used by local, state, and federal public health officials). As CDC begins to assess the situa- tion, it enlists FDA and USDA counterparts in a joint investigation. Simultaneously, local media in some communities are alerted to the situation by parents of ill children, and reporters who contact local hospi- tals learn of additional cases of similar illness. A wire service editor notes the occurrence of similar reports in multiple local or regional newspapers and publishes a national report, speculating about the possi- bility of a common cause. Within a few hours of the national wire service feed, Al Qaeda in the Arabian Peninsula claims responsibility, asserting specifically that sleeper cells in the United States were the perpetrators but being ambiguous regarding how the infections were introduced. This announcement triggers widespread coverage by television, radio, and print media. NBIC, through its daily monitoring of media reports and other open- source information outlets, takes note of the wire service report and the Al Qaeda claim and decides to seek, obtain, and analyze pertinent sur- veillance information. That same day, CDC, FDA, and USDA apprise NBIC that they are investigating whether the cases in the various cities are related and, in particular, whether they are the result of a foodborne pathogen from a common source. FIRST MOVE The first move of the scenario involves tracking the relevant mortality and morbidity, which continues for the duration of the incident. This move also includes investigation of the terrorists’ claims of responsibility. 

DISCUSSION OF A SCENARIO 43 Raub asked specifically about CDC’s BioSense, which was designed in part to collect real-time information about prescriptions and over-the- counter pharmacy sales. Could this information provide near-real-time, semiautomated reporting that something might be amiss, as opposed to the information coming from individuals who are motivated to examine and report unusual occurrences? Braden replied that it depends whether the data BioSense provides are sufficiently specific, because “there’s a lot of diarrhea and vomiting out there, and picking a signal out of such a high background is actually not easy to do with syndromic surveillance.” The detection of bloody diarrhea, in this particular circumstance, is more helpful, and laboratory information is most helpful. Having that information has been a revolu- tion for CDC, Braden said, in investigating food- and waterborne out- breaks. Public health laboratories are critical in the initial stages of an outbreak. The genotyping systems they use are very good at detecting cases that are related and, just as important, cases that are not related. Laboratory information can detect just a few common cases spread through a collection of states, whereas just a few cases would not trigger a syndromic surveillance system. Collecting exposure information quickly is important, said Braden, in addition to detecting aberrations above a background level of disease. Having this information is essential in forming realistic hypotheses to pursue in tracking the origins of the disease. If public health officials see something that suggests intentional contamination, Braden continued, they will reach out to law enforcement. Also, at both the federal and state levels, partner agencies will be in- volved early. Braden observed that an emergency operation center (EOC) would be stood up early in the scenario, as would an operation center for the Department of Health and Human Services Secretary. CDC and other agencies also would participate in the Multiple Agency Coordination Foodborne Illness Outbreak (MACFIO) process. CDC would ask an FBI agent to participate in the EOC, and it would put someone in FBI head- quarters in Washington, DC. “We did that during [the anthrax crisis]; I think that was a lesson learned that you need to do that.” Similarly, if municipal drinking water were involved, CDC would have a quick ex- change of personnel with the Environmental Protection Agency to ensure that information flows between the two agencies. Other information flows would be operational as well. PulseNet al- lows data to be shared widely—for example, these data would be fed into the public health information system at USDA. In addition, a memoran- 

44 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION dum of understanding with Canada allows data to be exchanged between the two countries, which could be a precursor for a much larger interna- tional system. Braden said that determining the source of the illness and getting people away from that source is a multiagency job. CDC’s particular job is to implicate the source using multiple lines of evidence and whatever information is available. When a convergence occurs with other agencies that are doing a traceback, the agencies have something to pursue. “It is multiple prongs of approach—and in a hurry.” Kautter said that FDA relies strongly on CDC and on state and local officials to triage the information that comes to the agency. FDA also has a liaison at CDC, and CDC has one at FDA. In the case of the scenario, FDA would be asking whether the substances would be related to a food or to something else and if a product was regulated by FDA or by USDA. Tan said that FSIS has a senior epidemiologist embedded in CDC who can help generate hypotheses and relay information to USDA. USDA also has access to data streams that can help track suspected vehi- cles, such as VetNet and a consumer complaint monitoring system. But USDA needs to be specific in knowing where an organism came from and how it was distributed to consumers. Annelli mentioned that APHIS also has a liaison at CDC who would be immediately involved in the conversations. APHIS would be particularly interested in information tying an outbreak to a farm of origin, which is where it would be directly engaged. Tan emphasized the importance of the local and state health depart- ments. Using this information, USDA’s compliance investigation divi- sion would work closely with the applied epidemiology division and with state and local officials to find, as quickly as possible, such information as lot codes, sell-by dates, production dates, and establishments that are involved. The compliance investigations division also works with any cases that might have criminal intent, and it is closely linked with the applied epidemiology division. Russell said that the AFHSC within DOD also monitors outpatient visits. In this scenario, when a cluster was detected, the center would ex- amine its syndromic surveillance system to look for cases among military personnel. Also, DOD has reportable medical events, and STEC is re- portable, which would also be an avenue for information. In addition, the center, like other federal agencies, has a liaison in CDC, and that person would inform DOD about the disease outbreak in the scenario. 

DISCUSSION OF A SCENARIO 45 Raub asked about the claim of responsibility by Al Qaeda, and Jamison responded that such a claim would generate a threat credibility evaluation process at the FBI, which would rely in part on information from the FBI’s liaison at CDC. NSS and DHS also would be engaged in processes to determine whether the incident was intention. E. coli would not be an agent of choice for most terrorist organizations, said Jamison, but the agencies still would try to make that determination. Jamison mentioned that the traceback is the most important infor- mation to determine if the event was intentional. To make this determina- tion, a joint investigation with other federal agencies is preferable whenever possible. Interviews done by other agencies could help the FBI. Another consideration, he pointed out, is that once the media have gotten ahold of the story, the people doing the work at the agencies will be receiving calls from their superiors asking for information. “That makes everyone’s job that much more difficult, because now you’re not only going to be talking to our partners here at the table, but you’re going to have to talk to your boss . . . to let them know what’s going on, so they can go talk to whoever it is they need to talk to.” At this point in the scenario, NBIC would be getting information from several sources, said Quitugua. Information would be coming in from CDC, from other agencies, and from media reporting. NBIC would be looking for contextualizing information, such as the characteristics of the outbreak. It would seek information from CDC about ongoing actions and from the FBI about its investigations of responsibility. Based on this information, NBIC would determine, in consultation with the involved federal agencies, whether to activate the National Biosurveillance Inte- gration System protocol, which is focused on information sharing, coor- dination, and collaboration. SECOND MOVE In the second move investigations are under way to determine if the cases are related. Though no credible evidence of terrorists’ involvement exists, the investigation continues. Kautter described some of the challenges involved in the food sector. All that is known initially is that there are common symptoms that may or may not be related. The purpose of the investigation is to go from many unknowns to pinpointing specific food, the specific agent, the spe- cific lot, or whatever information is needed to get that food off the mar- 

46 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION ket as fast as possible, but even that is a difficult process. “If I asked you, right now, what you had as a side dish Saturday night, and what were all the different ingredients in that side dish, who could answer that ques- tion? . . . These are the kinds of questions that we would need to address to pinpoint: one, is it even food, and then, two, is it one common food?” Braden drew a distinction between data and information. Some data remain at the local level and never go to the state or federal level, and that is appropriate, since there is no need to bring together all of the data in one place. However, the information derived from those data does need to come together. The curation of this information extends from the media relations staff to the scientific staff, all of whom are responsible for part of the task. Braden described three data streams: laboratory, epidemiologic, and environmental. Each of those data streams can be very broad. A ques- tionnaire searching for common elements may come up empty, requiring that the questionnaire be changed and that new cases and controls be inter- viewed. Doing this work can take time, Braden said, and a database that can be used to sort out this kind of information has not been established. In response to a question about whether computerized systems could combine information from different sources to detect linked cases that rise above a background level, Braden pointed to some of the difficulties with food. A food product can end up 100 miles from where it is pro- duced before it ends up 2 miles from where it is produced. Furthermore, symptoms are the result of self-reporting, and “especially for GI [gastro- intestinal] illness, people are notoriously inaccurate about where they think they got an illness.” In response to a question about communication with and among the states, Braden observed that PulseNet is constantly detecting clusters that need to be assessed. In many of these cases, CDC asks states for infor- mation, and the states then go to local health departments to find what information has been gathered. Based on that information, a cluster may be judged an outbreak, in which case further investigations begin. For some clusters, such as botulism or E. coli O157 that is causing deaths among children, the EOC would be stood up very quickly, whereas other outbreaks do not require that level of response. CDC also has conference calls involving lots of people, said Braden. All of the investigators and partners are on the calls, so “information sharing occurs early and often and on the phone.” Kautter added that 50- state conference calls may occur two, three, or more times per week. “Good old-fashioned conference calls—you can get a lot of really good 

DISCUSSION OF A SCENARIO 47 information that way, and it allows the states to report some of the in- formation [they have], and for all the states to hear that latest information as well.” Early in the process, said Braden, CDC, through its EOC, would be taking many of the calls. Later, in the scenario described and according to the current structure, the MACFIO would be in charge. Regarding the provision of information, Braden said that CDC has learned that the web can be extremely helpful. “We will put as much in- formation as we can out on the web about the number of cases, how se- vere they are, where they are, all those types of things, pretty early in this investigation. And we would update that probably daily.” In response to a question about how government officials should re- spond to endless requests for information generated by media accounts of possible terrorist involvement, Bryant emphasized partnerships among agen- cies as a way of handling requests for information. Even before this outbreak was linked to a food, the FBI would be in contact with FDA through its Office of Criminal Investigation, as well as with other agencies. The FBI and CDC also would be working together under such circumstances. Annelli pointed out that the biological assessment threat response (BATR) process would likely be involved in this scenario. Any federal agency can call for the BATR process, and all agencies should be in the loop as it proceeds. White House communicators also would be involved, and, as Annelli said, “quite frankly, from what I’ve seen in other events, the communicators end up driving some of that external messaging and communications more than the operations folks do. We’d have a com- pletely parallel kind of organization working on the messaging piece, while the operations folks are trying to figure out what’s really going on.” In response to a question about the MACFIO, which emerged from the Food Safety Working Group of the Domestic Policy Council to cre- ate a higher level of coordination, Tan said that relationships and com- munications structures exist now outside the MACFIO. The MACFIO would be stood up in extreme circumstances, but the day-to-day work of liaisons and hosts makes things run smoothly. “What is critical for the investigation is what’s happening on the ground at the local and state health departments, what information they’re getting, and what hypothe- ses they’re generating. . . . That’s where it all comes together on a na- tional scale.” Tan said that USDA has an Office of Public Affairs and Congressional Relations that would handle requests for information from the public while also engaging in cross-agency coordination. However, Russell pointed out that there still is no one interagency location to go to 

48 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION for information in the situation described. DOD and NSS are both inter- ested in such a capability, he said. DOD’s involvement, said Russell, would not yet be determined at this point in the scenario, but it would be providing information to CDC and ensuring that communication occurs within DOD. “We feel strongly a responsibility to our own service members, and that would not be ne- glected at the expense of handing it over to the CDC.” A questioner asked specifically about the curation of information that extends beyond public health to the national security arena, and several panelists pointed to a tension that exists in this area. Russell, for exam- ple, said that DOD is divided between the public health sector and the security sector. “Efforts are under way to see if those communities can communicate better, but it’s an uphill battle.” Trust will be essential for these different communities to be at the table together without being threatened or jeopardizing trusted relationships. In response to a question about the investigation of terrorist in- volvement, Jamison said that if the situation is domestic and involves terrorism, the FBI will have the lead for the investigation. But until ter- rorism is established, it would be conducting a joint investigation. Raub observed that many government officials have experienced what he called “communication with our partners by press release, which is not exactly conducive to collaboration.” He then asked Quitugua how NBIC would be seeking to establish situational awareness. Quitugua re- plied that NBIC would be engaged primarily in collecting and assem- bling information, “because we have a very small group internally, and we can’t possibly replace the expertise of all the authorities that are al- ready working it.” NBIC would be trying to establish the size of the event, how bad is it, and whether it is going to get worse. It would reach out to other agencies for modeling results, for example, or for other in- formation. It would be particularly interested in turnaround times—what information might be available, and when will it be available. “We’re not doing an independent analysis of a completely separate data set,” said Quitugua. “We’re trying to link people who may have data sets that need to be brought together. A lot of them already are linked. But sometimes there are new things that become available.” In response to a question about who would be curating information at each agency, Tan said that the curator at USDA would be the operational team. Braden said that the situational awareness team at CDC would be putting information together from different sources. Kautter said that the FDA has established a new Coordinated Outbreak Response and Evalua- 

DISCUSSION OF A SCENARIO 49 tion program, which is now the repository for information in food safety events. In DOD, said Russell, each of the services has the responsibility of reporting to their chain of command through to their service Surgeons General, but also the Assistant Secretary of Defense for Health Affairs and the AFHSC. Audience members offered some insight regarding their expertise and local context. Joel Ackelsberg of the New York City Department of Health said that in a scenario such as the one described, state and local public health departments would be “ramping up their resources.” These kinds of cases would catch people’s attention. If necessary, the New York City Department of Health would activate its incident command and be able to bring in resources from across the agency. The foot sol- diers who are collecting information and helping to generate hypotheses would be at the local level. Stephen Redd of CDC observed that external communication needs a structure to ensure that the information being released is consistent. Also, he said, it is important for the spokesman to be a scientist to enhance credibility. “That’s something that we learned with H1N1 is really im- portant for trust.” THIRD MOVE In the third move, the source of the exposures turns out to be bean sprouts contaminated with E. coli. There is no credible evidence of ter- rorism, but the investigation continues. Braden emphasized the difficulty of identifying such a source. Implicat- ing bean sprouts could mean finding a person who ate bean sprouts, knows where the bean sprouts were bought, and had a shopping card linked to a shopping card database to show when the bean sprouts were bought. “Get- ting that kind of information is quite hard. People don’t remember that they ate bean sprouts—it’s often kind of a stealth vehicle.” Yet an accurate trace back is essential to keep partner agencies from being led astray. Braden also observed that identifying a specific pathogen increases the ability to predict what is going to happen. For example, E. coli can produce different types of Shiga toxins, which will influence the charac- teristics of the outbreak. This information in turn would need to be con- veyed to public health authorities at all levels. He also pointed out that this system is currently threatened, because more and more clinical la- boratories are adopting non-culture-based diagnostics, which means that 

50 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION cultures are not available to characterize infectious organisms. “We’re not going to have that isolate and be able to characterize it to be able to know what to do.” Braden also described the difficulty of acting as a communicator at CDC while participating in an investigation. “Sometimes the incident commander is spending half the time in the studios in front of the camer- as.” Joint press conferences with all of the involved agencies on the phone, along with as many as 150 reporters, also take up time. Tan reiterated that the epidemiological information starts with state and local health departments. USDA can rely on internal databases to investigate hypotheses, but without enough information the investigation cannot proceed. “The boots-on-the-ground element cannot be underesti- mated, because there’s often not enough information for us really to un- derstand exactly what establishment [was involved], what were the production dates, what are the lot codes. Often, people unwrap their food and then they toss the package, and all of our information goes into the garbage. And our folks on the ground are not adverse to going into the garbage and getting the information for us. They can, they have, they remind me often.” Kautter observed that implicating bean sprouts is a significant ad- vance because it provides a path to pursue. Bean sprouts are not particu- larly seasonal, but if the contaminated food were seasonal, the first question to ask would be where the food is grown that time of year. If the food was imported, were import samples taken, or are there domestic samples that can be tested? Can the genetic fingerprint of an infectious organism from the food be linked to a clinical sample? Also, once a source is known, industry calls become much more numerous, because industry or trade associations could have information about the origins and treatment of the food. On this point, Quitugua said that NBIC would not necessarily have all of the information that other agencies have. For example, NBIC does not have access to PulseNet, though it does have information conduits to the agencies. NBIC was originally envisioned as a place with liaisons from all of these agencies, but agencies are spread very thin and exper- tise is difficult to retain. Instead, the relationships are virtual through emails, phone calls, and collaboration that does not require people to see each other every day. “Face time is preferable, but I spent a long time run- ning a reference laboratory, where I only saw three people all day. . . . People are completely capable of doing that.” 

DISCUSSION OF A SCENARIO 51 Standing Committee member Merrie Spaeth of Spaeth Communica- tions pointed out that in the foreseeable future, people will not go to the cameras, but the cameras will come to them, in the form of small webcams that can post images directly to the web. She also applauded the involvement of communications experts being involved in the infor- mation dissemination process. In this scenario, Raub emphasized, the President, the Cabinet secre- taries, and a host of other people are leaning on the National Operations Center, on NBIC, and on the pertinent agencies. “Who’s got what, and what do you have, and what can you get and what do you need, all be- come part of the issue.” FOURTH MOVE The fourth move focuses on the effort to determine the site or sites at which contamination of the bean sprouts occurred. No evidence of con- tamination during processing or distribution appears. Until the identification of a source, said Braden, CDC would lead the investigation, with the incident manager who stood up the EOC coordi- nating the overall investigation. If an investigation were to become a criminal investigation, the FBI could take over, or if a facility were iden- tified as the origin of the contamination, USDA could take the lead. But until then, CDC would be the lead agency. Braden said that CDC would be doing targeted testing of sprout samples. Random testing early in out- breaks is not very productive, but once a target has been identified, tests can look for specific organisms. Kautter said that if a positive sample were taken of the sprouts either in a restaurant or at someone’s home, FDA would work with its federal, state, and local officials to try to determine whether it was one brand or numerous brands, one distributor or numerous distributors, or perhaps a single lot. Trace backs would investigate the sprouting facilities, the source of the seeds, and the companies involved. FDA also would be bringing in industry members that are sprouters and their trade associations to help get information faster and sooner. At this point, CDC and FDA would proba- bly be conducting joint press communications, with regular updating to keep the public informed. The situation would be somewhat different if the seeds were imported from another country, Kautter added. If so, the appropriate countries would 

52 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION be involved in the investigation. FDA also would communicate with DOD to see if military personnel were affected. Mugno asked when the information about bean sprouts would be communicated to the public. Kautter said that the issue with food safety events is always how much information to convey when information is uncertain. Telling the public not to eat a food product can be a major economic burden, but the FDA leans on the side of caution for public health. With 10 cities and a number of deaths, “we are going to get more information out sooner than later,” he said. However, no food safety out- break is the same as another. “It is really a discussion point between usu- ally CDC and FDA regarding the public messaging,” said Kautter. “It would be nice to have a one-stop shop template for all outbreaks, . . . but every single outbreak is different and the information that we get from every outbreak is different. We have to triage that and determine where our public messaging is.” In response to a question about how quickly information would be delivered to the public, Quitugua said that in significant events, DHS personnel may have just 60 minutes to prepare the Secretary to talk with the press, and that “all efforts” are made to ensure interagency coordina- tion. Spaeth emphasized the importance, in the current media environ- ment, of making immediate public contact to establish that the situation is being managed. She also emphasized that “there is an enormously rich amount of [communications] resources today at your disposal which ought to be factored into the planning at some point.” Quitugua noted that the information being collected is also what NBIC wants to know. In addition, it wants to know what kinds of ques- tions agencies are being asked, because if one agency is being asked those questions, other agencies are likely to get them, too, and some agencies but not all may have answers. In this way, NBIC could act as a clearinghouse to provide people with a common operating picture. In- formation “doesn’t always get spread evenly,” said Quitugua. A cross- check by NBIC could ensure that everyone has the same information. Meanwhile, said Jamison, the information being gathered by other agencies would be the same information the FBI would need to assess the possibility of intentional contamination. The bureau would be look- ing at individuals who work with the bean sprouts in the processing plant or in the farm to try to determine if they have a terrorist connection. For instance, once a strain was identified, an investigation could determine whether individuals had access to those strains. 

DISCUSSION OF A SCENARIO 53 Nancy Carter-Foster, representing the State Department, said that the State Department would be involved in this episode and would expect to work closely with the FBI, DOD, and others. “We are part of the security agencies as well as the diplomatic agencies,” she said. Also, the terrorist claim would make this outbreak a national security incident, and the State Department would be investigating it in that context as well. Final- ly, the messages conveyed to the public need to be as verified as possi- ble. Messages should “let people know what the problem is as well as what it is not.” For example, a recent salmonella outbreak led to the de- struction of an entire crop of Mexican tomatoes and peppers, causing losses of hundreds of millions of dollars, but tomatoes and peppers turned out not to be involved. “The food industry is a global industry. Things that you think are just messaging to protect our health can have much broader implications.” A workshop participant asked if agencies have ways of integrating information that is received before an outbreak occurs. Jamison replied that the FBI has procedures and liaisons throughout the government to respond to intelligence pointing toward an attack. The exact response depends on the information received and the source of the information. But the FBI would let the public know about a possible incident, as it has with credible evidence for terrorist events in recent years. Russell observed that DOD also has organizations that are ready to receive such information, such as the regional commands throughout the world. But, he added, procedures to do so at the AFHSC remain subop- timal, even though this is an organization that needs to work well. Quitugua also expressed a concern that the conduits for intelligence information are not clear. “I don’t know for a fact exactly how I would get that information from the FBI. I have a feeling it would go a very secure route that would take a long time and maybe completely bypass the center and a number of NBIC partners.” People may say that they are going to share information, but they do not exercise those communica- tions routes in each event. Jamison agreed, though he also described the Interagency Policy Committee through which intelligence information would be communi- cated to agency representatives. “At that point it is incumbent upon each department and agency to let their respective officers know what they need to know, and that is often where the breakdown occurs.” Also, he observed, public health departments are often left outside the loop. “We are working to alleviate that problem, but we are definitely not there yet.” 

54 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION Tan observed that USDA has a specialized part of the organization that is equipped to receive classified information, but getting clearances elsewhere in the organization so that information can be disseminated can be a problem. Annelli reiterated that USDA has systems to acquire classified information. The challenge is converting classified information into actionable information that can be more widely distributed for use. “How do we take that information and get it to the people who need to know regardless of their level of security clearance . . . and then down to the state and local level?” Raub added that information can occur in dif- ferent versions. One version might be promulgated broadly, while anoth- er version goes only to people with the appropriate clearance. Annelli pointed out that the person who has information is often the one deciding who else needs to know it, yet that person often does not know what other people need. For example, USDA’s Food and Nutrition Service does not buy bean sprouts for the school lunch programs. This piece of information would be useful to add to the messaging, but it may not necessarily be conveyed without special attention to interagency communications. Kautter reminded the workshop participants that with a claim by ter- rorists of involvement, as in this scenario, FDA or CDC are probably not going to be the main source of news. Rather, department secretaries or even the President will be on the news. A good protocol is available for more routine episodes, but this episode is dramatically different “if CNN is running this 24/7 that terrorists have contaminated the American food supply.” On this point, Jamison noted, the terrorist claim would certainly create the public perception that the FBI is leading the investigation. Jamison said that the agencies have discussed having a single spokesman, whether someone from NSS, from DHS, or from CDC. In this way, all agencies could be represented and could all provide the same message. Kautter said that information exchanges with the FBI are not a prob- lem at FDA, which has numerous people with the necessary clearances. Ackelsberg noted that New York City has people with the necessary se- curity clearances. Furthermore, based on the relationships built over the years, city officials would be confident that they would be told of a threat regardless of their security clearances. Gibson, however, was less confi- dent that his agency would get the information it needs as they prepare for the 2012 Super Bowl in Indianapolis, partly because the law enforcement and health care communities do not necessarily understand the needs and role of public health. 

DISCUSSION OF A SCENARIO 55 FIFTH MOVE In the fifth and last move, no signs of contamination are found in the processing, packaging, and distribution systems. Instead, the source of contamination turns out to be a particular farm. Meanwhile, investigators determine that terrorism is very unlikely. USDA has sources of information involving plant and animal pro- duction systems, said Annelli, but it does not necessarily have the author- ity to investigate these systems unless the cattle are shedding E. coli into water supplies, said Annelli. USDA may have the expertise, but it also may need to exercise that expertise under someone else’s authority, such as FDA or CDC. USDA also would be working with state agricultural counterparts who have local expertise. For example, it has a veterinary assessment team that could look at farm sites and perhaps do some sam- pling to locate the organism and the route of contamination. Gibson ob- served that county public health directors have remarkable amounts of authority in cases where there is a threat to public health. If experts indi- cate that something needs to be done, “there is huge power at the county health director level,” he said. Braden pointed out some of the complications with bean sprouts, which are sprouted in facilities from seeds that often come from overseas and are not considered a food. The first part of the investigation would go to the sprouting facility, and FDA would probably be leading that in- vestigation with its local counterparts. But experience would suggest that the seeds were actually contaminated before coming into the facility. In response to a question about how information is communicated up the chain of command at an agency, Braden said that information is commu- nicated in many different ways. Multiple reports would come out of the emergency operations center for different audiences. During conference calls with other investigators the information will be more detailed, but information is summarized for reporting up the chain of command. Russell described the need to make information “pertinent and actiona- ble.” He added that compiling exact numbers for an outbreak can become counterproductive as time goes on, and agency personnel may need to resist the demand for numbers from authority figures. Kautter said that FDA and its state and local counterparts would be tak- ing samples of the water, the seeds, and the final sprouts. If the trace-back pointed to a specific facility, investigators would determine what other products that facility was producing that could be contaminated. Once a 

56 BIOSURVEILLANCE INFORMATION SHARING AND COLLABORATION particular sprouting firm was identified, recall recommendations would be issued for products, including other products made at that facility. Quitugua noted the difficulty in some episodes of tracking the many kinds of information that are relevant. Agencies may craft exactly what they want to say, but as that information gets aggregated and summa- rized, the underlying messages can change. Ackelsberg noted that a very similar process would be occurring at the state and local level as at the federal level. Jurisdictions would estab- lish incident command centers so that they would be able to gather in- formation and respond to questions as best as they can. Formal communications would be going up and down command structures, with informal communications across agencies. Public messaging would be a major focus with a disease like this that is contagious and can cause sec- ondary infections. State and local public health departments would be working with their partners to find the source of the outbreak and stem its spread as quickly as possible. Ackelsberg reiterated that a biosurveillance system already exists— at the local, state, and national levels—and it is being used every day. “This isn’t the time to start reducing the capacities of these systems that have been built up over the last 10 years with great investment. We are doing good work. That often goes unseen because it is just what we do every day.”