Many acute immunizing human infections, like measles, exhibits complex spatial dynamics with frequent local extinctions and recolonization among cities, towns and villages. Regional persistence is generally thought to depend on either local persistence in a small number of large communities above the critical community size and core-satellite spread, or metapopulation persistence and extinction-colonization dynamics among asynchronous locally non-persistent epidemics. Mass vaccination is arguably one of the most intense disturbance to the spatial dynamics of any consumer-resource system. We use a variety of statistical methods to investigate transitions in spatial dynamics en-route to elimination of measles in the United Kingdom. We further ask what transportation models best accounts for the hierarchical spatial spread of infection among local communities. We employ the uniquely-detailed urban spatio-temporal incidence data set for measles during the prevaccine era (1944-1967) to the infection’s vaccine-induced nadir in the 1990s.
Results/Conclusions:
We show that core-satellite dynamics with disproportional importance of a handful of large cities determined synchrony and dominated the spatial dynamics in the prevaccination era. However, as vaccination levels increased, this was replaced by true metapopulation dynamics. The resulting decorrelated and erratic spatial dynamic is a significant impediment to goals for regional elimination because of the frequent rescue-effects. We propose a parameterized model with a transition from gravity-like to spatially-random coupling to capture the effect of vaccination on spatial dynamics. We further explore alternatives to the gravity-model — such as Fotheringham’s competing destinations model and Stouffer’s model of intervening opportunities — for capturing spatio-temporal disease dynamics.