Disease-driven network adaptivity: implications for epidemic dynamics.

Our study introduces a dynamic approach to modelling adaptive networks, where the maximum number of links per individual adjusts based on the total number of infected individuals. This departure from fixed-link assumptions aligns with real-world observations, supported by the evidence from disease outbreaks like COVID-19, where mean contact numbers decrease significantly. Using a logistic function, we model the evolution of mean contacts during COVID-19 outbreaks, corroborated by CoMix survey data from various countries. Through simulations and analysis, leveraging the effective-degree ordinary differential equation formalism and stochastic network simulations, we demonstrate how adaptive networks alter epidemic outcomes, affecting critical thresholds and final epidemic sizes. We find that while adaptive behaviours can lead to substantial reductions in epidemic outcomes, there are cases where insufficient adaptivity can be less effective compared with static networks due to delayed responses. Additionally, networks with higher adaptivity strength show greater resilience in managing highly transmissible diseases, whereas lower adaptivity strength tends to be more beneficial in scenarios with lower transmission rates. Our findings underscore the importance of incorporating dynamic network responses for accurate disease modelling and intervention strategies.