Structural and Epidemic Properties of Contact Networks from Epigames

Infectious diseases spread through highly structured contact networks between individuals. Such contact networks can capture the population-level shape and nature of these contacts- but often explicitly recorded networks are unavailable, intractably large, or retrospective. Therefore, in many cases, outbreak modelling that includes information about contacts effectively makes simplifying assumptions about the structure and dynamics of these networks. Understanding how patterns of inter-personal contacts vary across contexts is key to predicting epidemic potential. Experimental epidemic games (“epigames”) are controlled situations in which participants join a simulated epidemic via a Bluetooth-enabled gamified smartphone app. Using multiple analyses including structural, spectral, temporal, and epidemic metrics, we characterised differences between networks generate by five epigames from two international conferences and three university campus deployments in the period 2020-2025. Our analyses demonstrated that epigames can capture realistic, high-resolution contact networks across diverse social contexts. The conference deployments reproduced the dense, cohesive structure observed in empirical studies of short, intensive events, while the university campus deployments exhibited the sparser, more modular networks characteristic of larger and more heterogeneous populations.