Incorporating Human Behaviour in Epidemic Dynamics: A Modelling Perspective

Abstract

The past few years have seen the development of a suite of extended epidemic models that take into account the “active” nature of individuals and/or population. Many models start from the natural premise that individuals are not “passive” but, on the contrary, receive and process information about potential or ongoing epidemics. Therefore, risk perception and behaviour change play a major role in shaping and changing the outcome of an epidemic. Incorporating such aspects into classical epidemic models poses many challenges. First of all, there are many open questions about how information is generated, its availability locally and globally, its routes of dissemination and diminishing returns of “old” information. All these factors lead to a significantly extended state space with many more variables and parameters compared to standard epidemic models. Thus, apart from issues around measuring and quantifying risk perception and/or behaviour change driven by information, a major modelling challenge revolves around model complexity. More precisely, how to achieve an optimal balance between model accuracy and tractability. In this chapter, starting from a pairwise model that accounts for the concurrent spread of an epidemic and information, modelling complexity and results are discussed by (1) evaluating the effectiveness of various information generating and transmitting mechanisms followed by (2) the deconstruction of the pairwise model to a simpler variant and by (3) discussing concrete modelling alternatives (i.e., pairwise and effective degree models for dynamic networks) and potential future modelling trends in the area of coupled models of human behaviour and disease transmission.