EPIDEMIC PREDICTIONS AND PREDICTABILITY IN COMPLEX ENVIRONMENTS

The spread of epidemics is inevitably entangled with human behavior, social contacts, and population flows among different geographical regions. The collection and analysis of datasets which trace the activities and interactions of individuals, social patterns, transportation infrastructures and travel fluxes, have unveiled the presence of connectivity patterns characterized by complex features encoded in large-scale heterogeneities and unbounded statistical fluctuations. These features dramatically affect the behavior of dynamical processes occurring on networks, and are responsible for the observed statistical properties of the processes' dynamics and evolution patterns. In the context of large-scale propagation of emerging infectious diseases, the air transportation network is known to play a major role in shrinking distances around the globe, by connecting far apart regions and allowing infectious travelers to potentially spread the disease to different geographic areas in a relatively short time. Here we will present a large-scale stochastic computational approach for the study of the global spread of emergent infectious diseases which explicitly incorporates real world transportation networks and census data. The simulated spatio-temporal pattern of epidemic propagation is analyzed in relation to the heterogeneous properties of the underlying complex architecture. Specific quantitative indicators are introduced to evaluate the predictive capability of the computational approach with respect to the intrinsic stochasticity of the disease transmission and of human interactions and movements. The interplay of the complex properties of the transportation infrastructure with the disease dynamics leads to the emergence of epidemic pathways as the most probable routes of propagation of the disease, selected out of the huge number of possible paths the disease could take by following airline connections. A case study for risk assessment analysis and comparison with historical epidemics is analyzed.