DIMACS Workshop on Game Theoretic Approaches to Epidemiology and Ecology

October 15 - 17, 2007
DIMACS Center, CoRE Building, Rutgers University

Alison Galvani, Yale University, alison.galvani@yale.edu
Tim Reluga, Los Alamos National Laboratory, treluga@lanl.gov
Presented under the auspices of the Special Focus on Computational and Mathematical Epidemiology.


Sanjay Basu, Yale University

Title: Do Men Matter? A Game Theoretic Model to Assess the Effect of Herd Immunity on the Impact of HPV Vaccination

HPV vaccines have provided an opportunity to significantly reduce the incidence of cervical cancer. To optimize future cancer prevention programs, public health planners must anticipate the degree of public adherence to voluntary vaccination recommendations. This will assist in determining the degree to which preventative screening will be required in the post-vaccine environment, and whether incentives may be necessary to reach vaccination coverage targets for the community. A currently-available quadrivalent vaccine covers some of the HPV types related to genital warts, providing incentives for men to vaccinate in addition to women. It is unclear how many women or men would accept vaccination, and the degree to which herd immunity effects could provide community-level benefits. We parameterized an epidemiological game-theoretic model of HPV vaccination with questionnaire data on actual perceptions about cervical cancer, genital warts, and HPV vaccination, to estimate the vaccination strategies that may be adopted by women and men in the United States. Our results revealed that the cost of vaccination to the individual may have to be substantially reduced to reach an optimal balance of costs and benefits in the population. This may require incentives to vaccinate. Male vaccination may provide significant benefits in some circumstances. However, a majority of incentives should be directed towards women.

Chris Bauch, University of Guelph

Title: Game theory and game dynamics of vaccinating behavior under a voluntary policy: a tragedy of the commons?

Voluntary vaccination policies for childhood diseases such as measles present parents with a subtle challenge: because of herd immunity, if a sufficient proportion of the population is already immune, either naturally or by vaccination, then even the slightest risk associated with vaccination will outweigh the risk from infection. As a result, individual self-interest might preclude complete eradication of a vaccine-preventable disease. In this talk we review previous game theoretical analyses of this "Tragedy of the Commons". We also present more recent work on the game dynamics of vaccination behaviour, which describe the time evolution of the frequency of vaccinators in the population as well as the disease dynamics. The models predict that a quick increase in perceived vaccine risk will tend to induce a rapid decline in vaccine uptake, whereas an equally quick (or slower) restoration of perceived vaccine risk to pre-scare values will induce only a very gradual restoration of vaccine uptake. This pattern of a rapid drop in vaccine uptake followed by a gradual recovery was observed in a pertussis vaccine scare in England and Wales in the 1970s.

Cláudia Torres Codeço, Fiocruz

Title: Vaccination in a disease-free scenario: how to decide?

Concern regarding natural or induced emergence of infectious diseases have raised a debate on the pros and cons of preemptive vaccination of populations under uncertain risk. In the absence of immediate risk, ethical issues arise because even small risks associated with the vaccine are greater than the immediate disease risk (which is zero). The model proposed here seeks to formalize the vaccination decision process looking from the perspective of the susceptible individual, and results are shown in the context of the emergence of urban yellow fever in Brazil.

Flavio Codeco Coelho

Title: Modeling the impact of vaccine safety belief dynamics on vaccination coverage

Individual perception of vaccine safety is an important factor in determining a person's adherence to a vaccination program. This perception, or belief, about the safety of a given vaccine, is not a static parameter but a variable subject to environmental influence. To complicate matters, percetion of risk (or safety) does not correspond to actual risk. In this paper we propose a way to model the dynamics of such beliefs in the context of a realistic epidemiological scenario. The methodology proposed is based on Bayesian inference, and can be extended to model more complex belief systems associated with decision models.

Nina H. Fefferman, DIMACS

Title: Playing Games at School: Parents, Public Schools, and Children's Health

Both parents and schools want to maximize both the education and health of children in their care, but the group- vs. individual-level concerns with regard to the behavior of any one child can lead to conflict in the desired policies for how schools should handle any one sick child. These conflicts only get worse once the economic factors of school funding based on attendance and home finances incorporating missing work or paying for childcare. We'll propose a framework for quantifying these seemingly qualitative decisions and present an overview of some of the interesting games that result.

Eili Klein, Resources for the Future

Title: Impact of Incentives for Surveillance and Reporting on the Spread of an Influenza Epidemic

The emergence of a new flu strain or a new disease similar in infectiousness and virulence to SARS is a significant preoccupation of the scientific community as well as policy-makers. In this paper we examine how incentives for surveillance and reporting impact the spread of a novel influenza virus from a source country. Specifically we are concerned with changes in both the sensitivity and specificity of testing regimes and the impact this has on a countries incentives to do surveillance and report. Using a game theoretic approach, we examine the number of deaths and the benefits and costs of reporting across a range of surveillance regimes in an SIR model with two countries. Our findings suggest that increased surveillance is not always the optimal solution for a potential source country.

Beth Kochin, Emory

Title: Selfish Vaccination can be Higher than Utilitarian Vaccination for Chickenpox

Vaccination is the primary public health measure for reducing the transmission of chickenpox. It has been repeatedly found that selfish vaccination is lower than utilitarian vaccination. Chickenpox severity increases with age, which we find reverses the typical relationship between utilitarian and Nash vaccination levels. An epidemiological model is used to determine effects of different cost structures and vaccine efficacies on the utilitarian and Nash vaccination levels. We find that the Nash vaccination level can be higher than the utilitarian vaccination level, a situation that has not been previously appreciated.

Ramanan Laxminarayan, Resources for the Future

Title: Infection Control Games Hospitals Play

Hospitals play a central role in healthcare delivery but they can also become important foci for pathogen transmission. Many pathogens, especially some antibiotic resistant bacteria have become adapted to modern hospitals. Hospital efforts to control these infections are influenced by the behavior of other facilities since those actions determine the likelihood that a patient entering the focal facility is colonized with a resistant pathogen. In this talk, I cover the role of game theory both in describing interactions between hospitals but also in determining optimal subsidies for encouraging better infection control in hospitals.

Anup Malani, University of Chicago

Title: Incentives for surveillance and reporting of infectious diseases

Containment of potential pandemic threats such as SARS and avian flu depends on prompt reporting of outbreaks. This paper examines the incentives of source countries to identify and disclose outbreaks. The analysis yields four non-obvious findings. First, there may be less loss of life than in the first best case with full information and coordinated action. Because trading partners do not internalize the economic cost of sanctions on source countries, they sanction more than in the first best and sanctions limit the spread of disease. Second, whereas sanctions by trading partners following the formal report (so-called reporting sanctions) of an outbreak discourages reporting, sanctions based on fears of an undetected or unreported outbreak (so-called non-reporting sanctions) encourage disclosure by reducing the relative cost of reporting. Third, improving the quality of diagnostic technologies or the veracity of rumors of outbreaks may discourage disclosure by encouraging trading partners to raise reporting sanctions. Fourth, more accurate rumors are less likely to discourage reporting than more accurate testing. Whereas more accurate rumors encourage higher non-reporting sanctions, more accurate technology reduces non-reporting sanctions and non-reporting sanctions encourage reporting.

Jan Medlock, Yale University

Title: Ethical Considerations in Vaccination Policy

It has been argued that the traditional outcome measures of deaths averted and infections averted are too simplistic because they do not take into account ethical considerations of the way society values of people of different ages. One alternate outcome measure is years of life saved, which prioritizes children's lives. Another alternative adjusts for parental and societal investment in children, prioritizing young adults. A few studies have been done using these alternate outcome measures, but they miss many of important epidemiological details, such as age-dependent transmission. In the context of influenza vaccination, I will discuss the evaluation of various vaccination policies using different outcome measures. Similarities and contrasts between the different outcome measures will be highlighted.

Timothy Reluga, Penn State

Title: Population Games for Vaccination and Epidemiology

In recent years, game theory has gained attention as a method to study behavior in situations where the value of an individuals decisions depends on the decisions of others. After a brief review of the some of the recent history, I'll describe how the combination of classic epidemiology models with Markov decision processes is used to formulate population games and study problems in public health settings. Example applications to influenza vaccination and polio infection will be discussed. I'll close by discussing some directions for future development of these techniques and applications.

Eunha Shim, Yale University

Title: Antiviral Intervention During Pandemic Influenza: the Prophylaxis and Treatment Coverage Levels Driven by Individual and Societal Interest

Antiviral agents are expected to play a critical role in mitigating the next influenza pandemic through prophylaxis and therapy. The administration of drugs has epidemiological and evolutionary effects with repercussions for both the individual patient and the community in general. The individual benefits from increased survival, but potentially suffers from adverse drug effects. The positive externality of drug treatment is reduced transmission. The negative externality is the selection for drug resistance. We evaluate how the balance among these factors determines the discrepancy between coverage levels driven by self-interest and the community optimum. We assess the impact of the transmission fitness of the resistant-type virus relative to the wild type on this discrepancy. We find that this discrepancy increases as the transmission fitness of the resistant strain increases. Furthermore such discrepancy is larger for the use of antiviral as prophylaxis than for antiviral therapy.

Christopher M. Snyder, Dartmouth College

Title: Which Vaccines Deserve the Largest Subsidies? An Integrated Economic and Epidemiological Model

Vaccines benefit the unvaccinated by reducing the prevalence of the disease. This positive externality drives a wedge between the social welfare and profit from producing a vaccine. We analyze vaccine externalities in an integrated economic and epidemiological model in which producers have market power. A result of our model is that externalities per vaccine dose are largest for rare diseases, implying that - holding constant total disease burden - vaccines for common but less serious diseases such as the flu are more profitable than vaccines for rarer but more serious diseases. We use this model to examine optimal government subsidies and also consider extensions to various oligopolistic market structures.

Claudio Struchiner, Fiocruz

Title: On the Persistence of Mosquito Susceptibility to Dengue and Malaria

Insects' resistance to infectious agents is essential for their own survival and also for the health of the plant, animal and human populations with which they closely interact. The recently sequenced Drosophila, Anopheles, and Aedes genomes provide a detailed and comparative view on their immune gene repertoires that in combination with post-genomic analyses can help dissecting a variety of effector mechanisms that can control infection. Mosquito defense mechanisms comprise a variety of immune responses, both cellular and humoral, and the refractory traits are under the control of quantitative trait loci (QTL). We review the natural distribution pattern of susceptibility to infection of the dengue and malaria mosquito vectors and discuss possible mechanisms explaining how selection shapes the genetic architecture of those quantitative traits and the evolutionary forces that maintain genetic variability.

Previous: Program
Workshop Index
DIMACS Homepage
Contacting the Center
Document last modified on October 11, 2007.