Evolution and emergence of novel human infections

Arinaminpathy, N.; McLean, A. R.

The work presented here underlines the unpredictability of infectious diseases undergoing adaptations for human transmission. Nonetheless, it also illustrates that general patterns of epidemiological behaviour can be associated with different evolutionary pathways. The scenarios presented here have widely different implications for public health, and each presents its unique problem in terms of containment. A pathogen following a punctuated route to adaptation is liable to emergence without any warning. Moreover, under this scenario, large outbreaks tend to be caused only by adapted pathogens, and so such outbreaks will be comparatively difficult to contain. By contrast, the gradual route offers some warning of adaptation in the form of large but self-limiting outbreaks. Both these clusters and the early stages of emergence are composed mainly of cases of a partially adapted pathogen, rendering containment comparatively easier. However, a danger is that repeated false alarms would elicit repeated containment efforts, potentially draining valuable resources. There are many possible refinements to the simple model of transmission presented here. One shortfall in the model is where large numbers of infection are predicted: in reality, large outbreaks will tend to be limited by local depletion of susceptibles, as well as being likely to trigger spontaneous social distancing and public health interventions. Host heterogeneity may also play a role. For example, Lloyd-Smith et al. (2005) point out in the context of a non-adapting pathogen that individual variation in transmissibility, hidden by a population-level value for R 0 , can have a strong effect on the outcomes of introductions, making extinction more likely than in a homogeneous population. Yates et al. (2006) explore different types of heterogeneity, including susceptibility to infection, and make the elegant distinction that, for nonadapted stages, the rate of adaptation has a stronger effect on probability of extinction than heterogeneity, and conversely for adapted stages. Extending this discussion to cluster sizes before emergence, these and other aspects of host population structure would be areas for refinement in models more detailed than those we have presented here. Nonetheless, the general insights offered by our approach are likely to remain valid. While mathematical models can be no replacement for detailed epidemiological investigations in the field, such as contact tracing and laboratory analysis, we hope we have shown here that they can offer valuable, objective insights into potential pre-emergence scenarios. Together with established frameworks for rapid case identification and management, mathematical models can play an important role in our toolkit for preparedness in public health. N.A. gratefully acknowledges funding for this work from the James Martin 21st Century School and A.M. is a Senior Research fellow of All Souls College whose support is gratefully acknowledged.