Game Theory, Braess’ Paradox, and the Spread of Epidemic Disease
The authors of the linked article have created a game-theoretic model of human behavior in the face of epidemic, with the goal of informing policy makers on how to most effectively quell the spread of epidemic diseases like H1N1 and SARS. It is commonly known that widespread vaccination is by far the most effective means of presenting the spread of such diseases. If a large proportion (typically between 85 and 90 percent, depending on the disease) of an at-risk population is vaccinated, then so-called herd immunity can be achieved — making further spreading of the disease mathematically improbable. However, individuals often choose self-protection (by means of various behavioral cautions e.g. washing hands frequently, staying indoors) as opposed to vaccination for a number of reasons, including but not limited to: the cost of vaccination, the perceived risk of vaccination, the perceived risk of contraction, the perceived severity of the disease, and the perceived curability of the disease.
Interestingly enough, analysis of this model uncovers a phenomenon in the study of the spreading of disease correspondent with Braess’ paradox in the study of traffic patterns. In other words, the article’s authors find, somewhat counter-intuitively, that increasing the success rate of self-protection also increases how effectively diseases are able to spread. They conclude that this phenomenon occurs regardless of network structure, but that it is much more pronounced in more heavily interconnected (or, in their language, “localized”) networks. The authors explain that as individuals realize the effectiveness of self-protection, they become less and less likely to vaccinate, driving the population further from herd immunity. In these more interconnected networks, information is transferred quickly and frequently, and thus the knowledge of the effectiveness of self-protection is able to become more widespread.
This article demonstrates that topics covered in the course (game theory, network structure, Braess’ paradox) are used to inform decisions of great consequence. While these principles can be used, as discussed in class, for matters as trivial as the prediction of traffic patterns or the interpretation of interpersonal relationships, here we are able see that they also underlie extremely significant behavioral patterns.
https://www.nature.com/articles/srep03292