Synthesizing and Predicting Infectious Diseases while accounting for Endogenous Risk (SPIDER)
E.P. Fenichel, C. Castillo-Chaves, P. Dazsak, R.D. Horan, C. Perrings
Funded by the National Center for Mathematical and Biological Synthesis (NIMBioS)
There is increasing interest in modeling risks associated with emerging infectious diseases (EIDs). Most EIDs are zoonotic in nature, and many infect valuable livestock and wildlife resources. Disease risks, like the risks associated with invasive species, are endogenous – a function of human decisions. However, most current attempts to model EID risks treat risk as exogenous and intrinsic. In order to successfully manage, predict, and develop surveillance strategies for new emerging diseases it is imperative that the human decision making processes, which influence disease risks are formally included in the decision process. Recent mathematical advances offer new opportunities to do so.
SPIDER is a working group funded by NIMBioS, and will meet four times during 2009-2010. SPIDER brings together a group of ecologists, epidemiologists, mathematicians, and economists to develop mathematical models of disease risk. SPIDER models are different than the standard epidemiological models of disease risk in two ways. First, SPIDER focuses on emerging infectious diseases (EIDs). These are infectious diseases that are caused by novel pathogens. Second, SPIDER is will focus on understanding the importance of endogenous risk. Endogenous risk is when the probability and adverse consequences of an event such as disease emergence is a function of human decisions. The concept of endogenous risk, though intuitively obvious, is often neglected in epidemiological modeling.
A traditional epidemiological model of EID risk follows the solid arrows in the figure below. Risk is treated as a conditional probability and each solid arrow is appropriately multiplied into the product. The ecological and epidemiological literature often treats the attributes of the links as properties of the invader or the invaded system. The nature of these links determines the probability of invasion. Yet, the attributes of the links are determined by human behaviors and human responses to the expected costs of invasions. The attributes of the links – and thereby EID risks - are endogenous. The feedbacks that shape the filters as functions of human behaviors are shown as dotted lines in the figure. Connecting the dotted and solid arrows cause an interaction between the probability of the event and the human response. A predictive model must account for these feedbacks.
