The role of social interactions among etiologically distinct
subpopulations in the seasonality of infectious disease incidence

N. Fefferman, E.N.Naumova

Background

Periodic trends in infectious disease are among the best known and worst understood phenomena in the study of disease dynamics.
Seasonal patterns in infectious diseases have traditionally been incorporated into epidemiological modeling via a variety of mechanisms of external forcing.
More recent investigations have examined the possible roles of stochasticity in generating these observed oscillations.

To demonstrate that seasonal patterns can arise solely from the differences in social interactions among etiologically distinct subpopulations.

We created a series of 14 stochastic population models to explore the roles of etiological heterogeneity and the effects of its influence on disease spread patterns in the context of a social network setting.
Examined a simplified social structure that consists of six interacting subpopulations based on age.
Exposure to pathogen was modeled via two distinct mechanisms.
" Once exposed, individuals became infected based on their subpopulation specific probabilities according to a state-dependent Markov process.
We plotted the mean rate time series of symptomatic and immune cases for four age groups with all adult subpopulations combined or for the entire population.

Different patterns of interaction among the subpopulations yield very different disease incidence and immunity curves. When the interaction rate with children is inflated, low magnitude oscillations are seen in the older children and in adults, but these decay over time and the other subpopulations do not exhibit such oscillations (Figure 2, Panel A). However, when interaction rates are inflated in only those adults who have children (either older or younger), this produces pronounced oscillations in all age groups, although those outside of adults have a lesser magnitude of fluctuation (Figure 2, Panel B). These results suggest that the emergence of oscillation is highly sensitive to the interaction rates of the contributing subpopulations.

Seasonal trends can arise solely as a function of etiological diversity and societal behavior.
Different patterns and trends in these oscillations can result from different pathways of introduction for primary exposure into the population through different subpopulations and how, once introduced, small oscillations within these subpopulations can be inflated by the continued interaction with the greater community.
Further research will be needed into a great many aspects of these social-network mixing-based transmission models in order to determine appropriate boundary conditions for particular sets of subpopulations.
These results will provide crucial insights into the underlying mechanisms of annual and seasonal outbreaks of infectious disease.