How contact patterns destabilize and modulate epidemic outbreaks

Abstract

The spread of a contagious disease clearly depends on when infected individuals come into contact with susceptible ones. Such effects, however, have remained largely unexplored in the study of epidemic outbreaks. In particular, it remains unclear how the timing of contacts interacts with the latent and infectious stages of the disease. Here, we use real-world physical proximity data to study this interaction and find that the temporal statistics of actual human contact patterns i) destabilize epidemic outbreaks and ii) modulate the basic reproduction number. We explain both observations by distinct aspects of the observed contact patterns. On the one hand, we find the destabilization of outbreaks to be caused by temporal clustering of contacts leading to over-dispersed offspring distributions and increased probabilities of otherwise rare events (zero- and super-spreading). Notably, our analysis enables us to disentangle previously elusive sources of over-dispersion in empirical offspring distributions. On the other hand, we find the modulation of R0 to be caused by a periodically varying contact rate. Both mechanisms are a direct consequence of the memory in contact behavior, and we showcase a generative process that reproduces these non-Markovian statistics. Our results point to the importance of including non-Markovian contact timings into studies of epidemic outbreaks.

arXiv comment10 pages, 4 figures plus supplementary information

References in corpus (18)

• Epidemic processes in complex networks
• Identification of influential spreaders in complex networks
• The origin of bursts and heavy tails in human dynamics
• Link communities reveal multiscale complexity in networks
• Why social networks are different from other types of networks
• Modern temporal network theory: A colloquium
• COVID-19 lockdown induces disease-mitigating structural changes in mobility networks
• Localization and Spreading of Diseases in Complex Networks
• Fundamental structures of dynamic social networks
• Causality-Driven Slow-Down and Speed-Up of Diffusion in Non-Markovian Temporal Networks
• Can co-location be used as a proxy for face-to-face contacts?
• Calling patterns in human communication dynamics
• The Strength of Friendship Ties in Proximity Sensor Data
• Equivalence between non-Markovian and Markovian dynamics in epidemic spreading processes
• Description of spreading dynamics by microscopic network models and macroscopic branching processes can differ due to coalescence
• Small inter-event times govern epidemic spreading on temporal networks
• Building surrogate temporal network data from observed backbones
• Periodicity in Movement Patterns Shapes Epidemic Risk in Urban Environments