Host-level influenza virus-respiratory pathogen interactions are often reported. Although the exact biological mechanisms involved remain unelucidated, secondary bacterial infections are known to account for a large part of the influenza-associated burden, during seasonal and pandemic outbreaks. Those interactions probably impact the microorganisms’ transmission dynamics and the influenza public health toll. Mathematical models have been widely used to examine influenza epidemics and the public health impact of control measures. However, most influenza models overlooked interaction phenomena and ignored other co-circulating pathogens. Herein, we describe a novel agent-based model (ABM) of influenza transmission during interaction with another respiratory pathogen. The interacting microorganism can persist in the population year round (endemic type, e.g. respiratory bacteria) or cause short-term annual outbreaks (epidemic type, e.g. winter respiratory viruses). The agent-based framework enables precise formalization of the pathogens’ natural histories and complex within-host phenomena. As a case study, this ABM is applied to the well-known influenza virus-pneumococcus interaction, for which several biological mechanisms have been proposed. Different mechanistic hypotheses of interaction are simulated and the resulting virus-induced pneumococcal infection (PI) burden is assessed. This ABM generates realistic data for both pathogens in terms of weekly incidences of PI cases, carriage rates, epidemic size and epidemic timing. Notably, distinct interaction hypotheses resulted in different transmission patterns and led to wide variations of the associated PI burden. Interaction strength was also of paramount importance: when influenza increased pneumococcus acquisition, 4-27% of the PI burden during the influenza season was attributable to influenza depending on the interaction strength. This open-source ABM provides new opportunities to investigate influenza interactions from a theoretical point of view and could easily be extended to other pathogens. It provides a unique framework to generate in silico data for different scenarios and thereby test mechanistic hypotheses.

© 2024 Aaron A. King
3038 Biological Sciences Building
1105 North University Avenue
Ann Arbor MI 48109-1085 USA