Persistence of measles at the regional level can be seen as a classic metapopulation problem (Levins, 1969), the outcome of which depends on the balance between the frequency of local extinctions due to fadeout, and the rate at which infection is reintroduced from elsewhere to locales that have undergone fadeout. The potential for reintroduction to counteract local fadeout depends on the degree to which the incidence of infection is synchronized between different locales. If outbreaks—and hence fadeouts—are highly synchronized then the infection will tend to undergo simultaneous (or nearly simultaneous) extinction in many locales. A high level of synchrony between outbreaks, therefore, reduces the probability that reintroduction will occur, reducing the persistence of the infection (Bolker and Grenfell, 1995; Ferguson et al., 1997; Grenfell and Harwood, 1997; Keeling, 2000).
The importance of synchrony in understanding persistence of infection has led to numerous studies examining the spatial synchrony of epidemics. These have provided both a description of the patterns seen in spatially resolved incidence records and a fair understanding of the mechanisms that give rise to the observed synchrony (Bartlett, 1956; Murray and Cliff, 1977; Cliff et al., 1992a,b; Bolker and Grenfell, 1996; Lloyd and May, 1996; Rohani et al., 1999; Keeling, 2000; Grenfell et al., 2001; Keeling and Rohani, 2002).
Two distinct patterns of synchrony are seen, depending on the sizes of communities being observed. For a collection of cities (or other locales) that are each above the critical community size, a high degree of synchrony is often seen between their recurrent epidemics. On the other hand, for a collection of smaller-sized towns or locales that surround a large population center, wave-like behavior is often seen as infection spreads outwards from the large city into the surrounding region. This second pattern exhibits lagged synchrony, with phase differences between outbreaks in different locales. We shall discuss these two patterns separately.
Was this article helpful?