Cyclical infection prevalence in small groups

To investigate variability of endemic infection prevalence within groups, we contrasted three SIS models. A deterministic compartmental (DC) model was used to estimate means without stochastic variation; a stochastic compartmental (SC) model calculated variation around an equilibrium prevalence. An agent-based model (ABM) enabled serial observations of individuals. Under scenarios in which most infection arose from within the group and if the group was small, then Fast Fourier Transform analysis of the ABM showed significant low frequency oscillations of infection prevalence. This suggests that estimation of endemic prevalence among small groups should include some serial measurements to detect oscillatory behavior.