Risk Communication Planning for the Aftermath of a Plague Bioattack
Total Page:16
File Type:pdf, Size:1020Kb
Risk Analysis, Vol. 28, No. 5, 2008 DOI: 10.1111/j.1539-6924.2008.01080.x Risk Communication Planning for the Aftermath of a Plague Bioattack ∗ , Elizabeth A. Casman1 and Baruch Fischhoff1 2 We create an influence diagram of how a plague bioattack could unfold and then use it to identify factors shaping infection risks in many possible scenarios. The influence diagram and associated explanations provide a compact reference that allows risk communicators to identify key messages for pre-event preparation and testing. It can also be used to answer specific questions in whatever unique situations arise, considering both the conditions of the attack and the properties of the attacked populations. The influence diagram allows a quick, visual check of the factors that must be covered when evaluating audience information needs. The documentation provides content for explaining the resultant advice. We show how these tools can help in preparing for crises and responding to them. KEY WORDS: Bioterrorism; influence diagram; plague ecology; risk communication; rodents; Yersinia pestis 1. INTRODUCTION a method for developing communications for such situations. It uses an influence diagram to organize In a public health emergency, there is little time the facts relevant to the decisions that individuals to develop health communications. Unless that work might face. In advance of an emergency, prototype has been done in advance, public health officials must messages wouldbe developed and evaluated for sce- improvise—at the risk of saying wrong things (be- narios spanning the range of possible emergencies. cause the situation has not been analyzed properly) When an actual emergency arose, the prototype mes- or of saying right things wrongly (because messages sages would be adapted to the specific circumstances, have not been tested for effectiveness). If officials fail drawing on the information organized with the influ- the public, then they can cede the stage to less quali- ence diagram. We demonstrate the approach with a fied voices, offering confident, incompetent, and con- plague bioattack, one threat with multiple possible tradictory messages. scenarios. There are guidelines for systematically develop- ing and evaluating communications for well-specified risks.(1,2) But what happens when a threat’s de- 1.1. Current Risk Communication Planning tails cannot be predicted in advance? We propose for Plague Attack In focus groups convened by the Centers for Dis- ease Control and Prevention (CDC), participants re- 1 Department of Engineering and Public Policy, Carnegie Mellon ported wanting information that would help them University. to prevent and detect exposures, identify symptoms, 2 Department of Social and Decision Sciences, Carnegie Mellon and treat infections, along with background informa- University. ∗ tion providing them with basic understanding of the Address correspondence to Elizabeth A. Casman, Carnegie Mel- (3) lon University, Department of Engineering and Public Policy, hazard. Such information is available on the CDC (4) Pittsburgh, PA 15213 USA; [email protected]. bioterrorism website. 1327 0272-4332/08/0100-1327$22.00/1 C 2008 Society for Risk Analysis 1328 Casman and Fischhoff As seen below, plague risk reflects the in- ical in bioterror attacks. That requires behavioral teraction of multiple, complex processes. Without measures—and communications supporting them. decision-focused analysis, communications can miss critical facts or bury them in irrelevant details. Mem- 2.1. Naturally Occurring Plague in the United States bers of the public cannot be expected to set informa- tion priorities about topics where they lack expertise, Plague arrived by ship from China more than a even with more systematic data-collection methods century ago, causing rat-borne human epidemics in than focus groups, whose only proper use is in the port cities on the Pacific and Gulf coasts. Aggres- most preliminary, formative stages of research.(5) sive rat control and plague surveillance stopped its spread.(11) Similar measures prevented further urban outbreaks, but not before plague had moved into na- 1.2. Influence-Diagram-Based Rapid tive rural rodent populations in grassland, forest, and Communication Method shrubland habitats, where it is now endemic in the (9,12) There are large peer-reviewed and gray liter- western United States. atures about plague, its natural ecology, and con- CDC receives about a dozen reports of human trol with contributions from many disciplines. Influ- plague cases annually, with 78% traced to flea bites, ence diagrams can organize such disparate facts,(6,7) 20% to direct contact with infected animals, and 2% representing critical factors as nodes and their depen- to inhalation (the latter almost always involving do- (13–15) dencies as connecting arrows. mestic cats). Although epizootics (epidemics in Section 2 gives an overview of naturally occur- animals) can be geographically widespread, few hu- ring and postattack plague risk. Section 3 presents a man cases have resulted. basic influence diagram model of a plague bioattack, focused on factors relevant to decision making. Sec- 2.2. Zoonotic Potential of a Plague Bioattack tion 4 focuses on measures that disrupt model links. Section 5 describes how the model can be used to Most analyses of plague bioattacks have ignored identify and organize facts needed for effective com- the zoonotic dimension. For example, a major World munication. We focus on the United States, although Health Organization assessment assumed no ani- many conclusions apply elsewhere. mal uptake in a scenario involving 50 kg of plague bacteria dropped from a plane.(16) One of the De- partment of Homeland Security’s (DHS) 15 disas- 2. PLAGUE ATTACK SCENARIO AND ter planning scenarios has aerosol releases causing BACKGROUND INFORMATION thousands of human cases of pneumonic plague, (17) Plague is a rapidly progressing, often fatal dis- but no zoonotic involvement; DHS’s first three ease caused by the bacterium Yersinia pestis.Itisa TOPOFF planning exercises also had no zoonotic CDC Category A select agent, i.e., an organism suit- dimension. able for bioterrorism. Plague can infect many warm- In nature, plague is a zoonotic disease of ro- blooded animals, often lethally.(8,9) dents (rats, mice, chipmunks, squirrels moles, voles) and lagomorphs (hares, rabbits, pikas), presumably In order to infect humans, plague bacteria must (18) be inhaled, swallowed, or enter broken skin. Natu- susceptible to aerosol infection. Flea bites and rally occurring cases are mostly bubonic, transmit- contact with dead animals can infect humans and ted by flea bite and characterized by painful swollen companion animals like cats and dogs. The risk to hu- mans lasts until an epizootic burns through suscepti- lymph nodes (buboes). Some flea-borne infections (13) become septicemic, infecting the bloodstream. Pneu- ble animal populations. That could be prolonged, monic plague is an infection of the airways, and if illness (or fear) undermined the municipal services is usually contracted by inhaling infectious fluid that control plague risk: pest extermination, garbage collection, lawn mowing, sewer maintenance, animal droplets. (19–21) If begun within 18 hours of the first symp- shelters, etc. toms, antibiotics can treat most naturally circulating 3. INFLUENCE DIAGRAM plague strains,(4) though some drug-resistant strains DOCUMENTATION FOR A have been observed in Africa.(10) The Soviet Union PLAGUE BIOATTACK is thought to have developed multidrug-resistant strains. Because creating antibiotic-resistant bacte- Fig. 1 shows the basic relationships between hu- ria is straightforward, preventing transmission is crit- man and animal plague infections. An aerosol release Risk Communication Planning for the Aftermath of a Plague Bioattack 1329 3.2. Plague in Animal Populations 3.2.1. Enzoonotic (Maintenance) Cycle In nature, plague can persist at low levels of infection in reservoir hosts and their fleas, in the enzootic cycle. In the western United States, reser- voir hosts include wood rats (Neotoma species), deer mice (Peromyscus species), voles (Microtus species), kangaroo rats (Dipodomys species), and grasshopper mice (Onychomys leucogaster).(29,30) Plague can persist in the environment without living rodents in burrows,(31) carcasses,(32) soils,(33) grains, dry sputum, flea feces, and buried hu- man bodies.(27,34) At near-freezing temperatures, it can live for years. Animals digging through contaminated soil can become infected, initiating new enzootic cycles.(35–37) Reported survival times Fig. 1. Basic structure of plague influence diagram. in animal and human remains indicate that cold temperatures increase bacterial persistence. Plague (top) causes pneumonic plague in humans and was recovered from exhumed human bodies in SE animals, initiating a zoonotic cycle that can then Russia and Manchuria after 180 days in winter and cause bubonic plague in humans with the health 30 days in summer. More relevant to bioattacks, outcomes that follow. Fig. 2 elaborates these rela- plague persisted in guinea pig carcasses after 109 tionships, including bacteria transfer between hosts; days at –3 to –5◦C and mouse carcasses after 22 days factors that affect exposure, detection, countermea- at 1–10◦C, after 9 days at 10–22◦C, and after 7 days at sures; and information flows. 22–30◦C.(27) We now summarize the science underlying Wild animals can transmit plague to urban ro- zoonotic aspects of a bioattack. Others have summa- dents when their habitats overlap,