Modelling the Propagation of Infectious Disease Through a Nexus of Multi-Regional Contact Networks When There Arespatial Disparities in Vaccine Coverage

We develop an agent-based computational model for the spatial spread of an epi- demic to explore how the inter-regional variability in vaccine coverage a§ects the global epidemic size as well as the individual risk of infection. Our agent-based model imple- ments a standard SIR model of epidemics, but has the capacity to complement and enrich standard analysis of the model. The benchmark version of the model is speciÖed using county-level áu vaccination claims rates from California. Then, holding overall vaccination levels constant, we change the variance in the distribution of regional vac- cine coverage. The mean size of the global epidemics increases as the variance in regional vaccination rates increases. This relationship ampliÖes when the model has a greater share of inter-regional contacts. As expected, increasing the global rate of vaccinations decreases the risk of infection for unvaccinated individuals. In addition, the infection risk of an unvaccinated agent decreases when the number of vaccinated contacts in her contact network increases. However, we observe no signiÖcant corre- lation between the vaccination rate within a speciÖc region and the infection risk of unvaccinated individuals in that region. In summary, regions with low vaccination rates experience larger outbreaks but the probability that an unvaccinated agent is infected does not depend on the regional vaccination rate. Thus, we Önd a non-monotonic re- lationship in the geographic resolution of vaccination rate measurement. These results have implications for the assessment of an unvaccinated individualís risk of infection as measured at di§erent levels of information aggregation such as global, regional, or individual network-speciÖc infection rates