DIMACS Working Group on Modeling Social Responses to Bio-terrorism Involving Infectious Agents

May 29 - 30, 2003
DIMACS Center, CoRE Building, Rutgers University

John Glasser, CDC, jwg3@cdc.gov
Ellis McKenzie, NIH, McKenzel@mail.nih.gov
Fred Roberts. Rutgers University, froberts@dimacs.rutgers.edu
Presented under the auspices of the Special Focus on Computational and Mathematical Epidemiology.


Carlos Castillo-Chavez, Cornell/Los Alamos/Arizona

Title: A Simple Model for Toronto's SARS Outbreak

In this talk, I will present a simple SARS' model of Toronto's SARS' outbreak and the reasons behind its successful control. If time permits, I will discuss the potential impact of transient populations on disease spread in the context of New York City.

Claudio Cioffi-Revilla, GMU

Title: Dominance principles in social network dynamics and terrorism.

Dominance principles from the theory of political uncertainty (Cioffi-Revilla 1998, chs. 6-7) provide models for understanding the effect of different event and process structures on overall behavioral probabilities of terrorist attack, detection, response, failure/success, an= d other critical (compound) events. While most of these principles can be expressed in closed form, it seems difficult to avoid misconceptions about their true nature. Simulations (e.g., agent-based models) and other computational strategies may offer helpful visualizations, by implementing improved functionality in existing or future models.

Lee Clarke, Rutgers Sociology

Title: Leadership and Bioterrorism: Problems of Elite Panic and Public Communication

We know that public panic in disasters is rare. This fact is, however, routinely ignored by officials who must make policies regarding disaster response and risk communication. When public over-reaction does happen, it is sometimes because elites have panicked, thus creating the very condition they were trying to avoid. I will use previous research on bio-events to illustrate this point and to draw out the implications for present policies.

Paul Dreyer, Rand Corp.

Title: Modeling Lessons Learned from the SARS Outbreak

Although it is highly unlikely that SARS is an example of bioterrorism, the reactions to the disease worldwide provide some interesting insights that should be considered by anyone modeling responses to bioterrorism. For example, the counter-intuitive responses by individuals (breaking quarantine, hoaxes, failing to seek treatment) suggest the need to include methods for incorporating such people into models (particularly agent-based models). This talk will discuss some of the recent events that have occurred during the SARS outbreak, reactions by individuals and governments to the outbreak, and what we, as modelers, should learn from them.

Stephen Eubank, Los Alamos National Lab

Title: Bioterrorism and EpiSims

The EpiSims epidemiological simulation system relies on estimates of contact patterns among all the individuals in a large urban region. These estimates are derived from activity patterns associated with each household. Currently, only a few simple deviations from the pre-planned activities are allowed, such as staying at home or reporting to an isolation location. The simulation itself has been designed to be far more flexible, but we need a better understanding of how activity patterns will change in the event of disease outbreak, perhaps coupled with other unusual events, and how people will react to planned interventions.

John Glasser, M. Reynolds, B. Schwartz, and M.S. Cetron, CDC

Title: Control of Severe Acute Respiratory Syndrome via Isolation of Infectious People or Quarantine of Those Whom They May Have Exposed: Evaluation via Mathematical Modeling.

Background: Severe Acute Respiratory Syndrome (SARS) originated in China, where conditions facilitate transmission of avian and porcine pathogens to human hosts, but was quickly disseminated worldwide via air travelers. Methods: We have modeled SARS via systems of differential equations 1) to elucidate social phenomena affecting transmission, particularly the probabilities of infected people seeking medical care for prodromal symptoms, being correctly diagnosed and effectively isolated, and authorities quarantining their probable or possible contacts, and 2) to assist in deploying available public resources most advantageously. I. If people were infectious before developing respiratory symptoms, the incubation period would overestimate the requisite latent period, which may determine if SARS is controllable via isolation alone. We are estimating the interval between exposure and infectiousness, conditional on durations of prodromal and respiratory symptoms, by minimizing disparities between predictions and observations in defined populations. Similarly, prodromal and respiratory infection rates, products of rates of interpersonal contact and probabilities of transmission on contact (i.e., infectiousness), may differ in our models. II. As awareness of SARS increases, people with compatible symptoms should seek medical care earlier and attending physicians become increasingly likely to diagnose their afflictions. These social phenomena could ensure the timely isolation of infectious people or increase the proportion of face-to-face contacts quarantined. Accordingly, contacts of people seeking medical care during the prodrome and being correctly diagnosed may be quarantined soon after exposure in our models, whereas contacts of those who are misdiagnosed or present with respiratory symptoms may be identified too late for quarantine. Given estimates of the biological parameters, we are assessing a) sensitivity of control to these probabilities, particularly if quarantine is required or justified, and b) temporal variation associated with increased awareness, which official health communiqués could facilitate or impede. Results: We will present these sensitivity and other preliminary results.

John F. Heinbokel and P. Jeffrey Potash, CIESD

Title: Freeze, Fight, or Flee: Behavioral Choices in a Natural Plague Outbreak as a Surrogate for a Bioterrorist Event

We have used system dynamics to model the natural outbreak of pneumonic plague in Surat, India, in September 1994. These models currently include three behavioral responses (seek treatment or prophylaxis from the medical community, flee the city, and isolate oneself from contact with others = fight, flee, or freeze, respectively) as exogenous drivers. Use of exogenous drivers was a consciously preliminary step to explore the relative or potential importance of those behaviors in affecting the course of an outbreak. Our current challenge is to utilize recent social science research on judgment and decision making to make these behavioral components endogenous elements of the model. With those behaviors becoming fundamental elements of the controlling feedbacks, these models will have greater predictive value and credibility and better serve as guides to policy decisions.

Jim Kvach, AFMIC

Title: Infectious Disease Modeling: Why is Defense Intelligence interested?

In any given year, the United States deploys military units to over 150 countries on a wide range of missions. On any given day, the United States has over a hundred thousand troops stationed worldwide. Historically, infectious diseases account for the vast majority of military hospitalizations during war. Military deployments present unique risks, behaviors and environments. Infectious disease modeling holds the promise of reducing health risks to US military forces due to naturally-occurring or intentional disease introduction.

Ian Lustick, Penn

Title: Prophylactic vs. Response Strategies for Bioterrorist Threats: Modeling the Problem in PS-I

We will display a point and click capacity for combining operationalizations of numerous key variables including geography, incubation, quarantine, contagiousness, host mortality, recovery, immunization, etc. Our intent is to put agent-based modeling simulation capabilities directly into the hands of non-programmer health professionals.

Martin Meltzer, CDC

Title: perceptions of probabilities

I would like to talk about perceptions of probabilities and how those changes perceptions may change valuations. The key point is that, when considering policy options there is a need to also consider different perspectives. Many models consider the societal perspective. But, for the same problem, the individual perspective can give an entirely different answer. This does not mean that one model/ perspective is "correct" and the other model/ perspective is "wrong." Rather, it means that different perspectives can produce different conclusions. This realization can be very important piece of information for policy makers. I believe that mathematical modelers can aid policy makers by building models and discussing their results with direct reference to different perspectives.. I plan to illustrate this concept by showing examples from the economics of routinely vaccinating prisoners against hepatitis B, the economics of vaccinating restaurant workers against hepatitis A, a risk-benefit analysis of pre-exposure smallpox vaccination (from an individual's point of view) and considering the implications of positive predictive value (PPV) and false positives for sentential systems.

Enrico L Quarantelli, University of Delaware

Title: Human Behavior In Crises: Similarities and Differences in a Bio-Terrorism Occasion.

My intent would be to indicate what is known from social science research on how people (and groups) react in the face of crises (such as natural and technological disasters as well as conflict situations such as 9/11 and other kinds of terroristic happenings). Along many lines the research findings are very well grounded, along some lines, there is much yet to be learned. I do not visualize my talk as a lecture but as a presentation of some talking points which might be relevant to the focus of the proposed workshop.

Sarah Roberts, NYC Dept. of Health

Title: Communicating with the Public about Public Health Preparedness

The talk will describe some of the communication systems and strategies that the NYC Department of Health and Mental Hygiene has been developing to assist it in responding to a public health emergency (either naturally occurring or caused by terrorists). As a case in point, I will describe how the Department has been communicating with the public about SARS.

Peter Sandman, PSandman.com

Title: Risk Communication Aspects of Bioterrorism Preparedness and Response

How people respond to the threat of bioterrorism depends partly on how those in authority talk to them about what might happen; how people respond to the actuality of bioterrorism depends partly on how authorities talk to them about what has happened. I will sketch in a model that tries to position communication about bioterrorism in the overall risk communication terrain. Then I will identify a few of the recommendations that can be derived from the model, focusing on those that seem likeliest to be mishandled.

Monica Schoch-Spana , Johns Hopkins

Title: How to Vaccinate the Public against Panic

Anticipating public panic, social disorder, and even violence, decision-makers and disaster planners have rarely approached the general public as a critical partner in containing the effects of a bioterrorist attack. This session counters the prevailing negative preconceptions regarding mass response to crisis with a series of principles for creating and sustaining an active, positive role for the public in bioterrorism response. Key strategies for supporting a "people's role" in biodefense include volunteer mobilization, awareness raising (i.e., pre-event education, crisis communications), and participatory decision-making. Incorporated in the session will be a review of relevant historic and contemporary case studies (e.g., 9/11, anthrax letter attacks, 1918 pandemic flu) and their implications for encouraging the public's constructive collaboration in confronting a future bioterrorist attack

Martin Shubik, Yale University

Title: How many is too many or too few in a terrorist group?

These brief comments explore why this is an important problem the different aspects of how to approach the study of small groups. Technology, communications, weapons types, game theory, operations research, lawyers and sociology-anthropology. The argument is that the selection of mixed groups is both difficult and absolute necessity.

John Sorenson, Oak Ridge

Title: Modeling Human Response to a Chemical Weapons Accident

This talk will discuss research conducted at Oak Ridge National Laboratory on integrating behavioral considerations into dose assessment models for chemical weapons accident. The work focuses on the warning diffusion process and the response of people to warnings. I will also note some of the important issues to be considered in developing models for bioterrorism.

Ping Yan, Health Canada

Title: Roles of Statistical Science and Mathematical Modeling in the Effort to Control SARS

The knowledge of the incubation period, case fatality, and the spread of the Severe Acute Respiratory Syndrom (SARS) are crucial the current fight against this emerging disease. They must be obtained early when the window opportunity of eradication and control still exist. Current knowledge in these areas are based on empirical observations. They are often biased. Statistical science and mathematical modelling play important roles.

We question the current prevailing consensus in several very important areas. We call for collaboration of different disciplines: biostatistics, epidemiology and operations research, and call for international data sharing.

Previous: Participation
Working Group Index
DIMACS Homepage
Contacting the Center
Document last modified on May 27, 2003.