Large-scale infectious disease outbreaks spread at a rate proportional to the prevalence of infectious individuals, as mass action attenuates the effect of variation in individual transmission rates, even in highly organized host populations. By contrast, when the number of cases is relatively small, systematic variation in transmission rates among hosts has a bigger impact: for instance, allowing diseases to persist and spread despite high average rates immunity or low average rates of infectious contact. In these scenarios, infectious diseases propagate by ’stuttering-chains’ of transmission, leading to more frequent extinctions as well as sporadic bursts of new infections. These dynamics are especially important in emergence/elimination scenarios, as exemplified in the current resurgence of measles in the US. To achieve efficient control in these scenarios, we need epidemic models that link stuttering chain dynamics to underlying variations in the host population structure.
Results/Conclusions
We describe a demographically-informed framework for ’stuttering-chain’ dynamics that models systematic variation in host transmission potential as autocorrelation in transmission rates along a branching process of disease spread. As a demonstration, we fit the model dataset on the spread of canine influenza virus (CIV) across the US. Accounting for systematic variation in spread rates in CIV reveals a lower value for the basic reproductive number (R0 = the expected number of secondary infections caused by an index case in a susceptible population; R0 ≈ 0.7 for CIV). This contrasts with canonical epidemic theory based on mass action which predicts that diseases with R0<1 are doomed to rapid extinction. Rather, we demonstrate how contagion processes can persist because of ‘heritable’ variation in the propensity for ’super- spreading.’ Rather than superspreading events being random, they arise from incremental movement away from the mean transmission rate, as the pathogen encounters localized areas of high host population density. Understanding how localized variation in host population structure drives disease dynamics during emergence and elimination can be essential for forecasting and control.