Spatial Heterogeneity, Host Movement and Mosquito-Borne Disease Transmission

Acevedo, Miguel A., et al. “Spatial Heterogeneity, Host Movement and Mosquito-Borne Disease Transmission.” PloS one 10.6 (2015): e0127552.

Acevedo et al. note that most modeling efforts of vector transmitted disease assume spatially homogeneous transmission. This assumption may be problematic because of the many potential environmental drivers of heterogeneity in vector demographics and transmission rates. To address this gap they formulate a multi-patch Ross-McDonald model. Within each patch the infection dynamics of humans and vectors proceed according to a typical Ross-McDonald model but humans are allowed to move between patches at some fixed rate (k). All parameters in the Ross-McDonald model are held constant across patches except for m – the ratio of mosquitoes to humans in the population. Varying m across patches creates a spatial heterogeneity in transmission as well. The degree of variation in m is represented by the coefficient of variation (CV) of m across the network.

The effect of heterogeneity in transmission on R0 and human disease prevalence is then tested analytically, with a two patch model, and in simulation, with a ten patch model.  In the two patch model they prove analytically that R0 (for the two-patch network) and total prevalence increase with increasing CV (which in this case represents a larger difference between the m value in each patch. This dependence of R0 on CV decreases as the movement between the two patches increases. Similar trends were observed in the 10 patch simulation. Both R0 and prevalence increase with increasing CV, which in keeping with previous findings, but degree of human movement plays a large role in determining the value of these terms. Increased movement necessarily decreases R0 to a horizontal asymptote (where CV based differences are preserved) while low, non-zero, levels of movement maximize prevalence. When movement is very low most humans remain in their original patch, isolating disease in the high transmission hotspots and forcing the infectious mosquitoes in these areas to bite many already infected humans. Once movement is slightly increased then the number of humans exposed to these hotspot regions significantly increases thereby increasing prevalence until movement increases enough to begin to limit the amount of time any individual spends in a hotspot patch, thereby increasing their infection risk. As a result R0 and prevalence do not have a direct positive correlation in this system.

This paper provides a good example of the incorporation of spatial heterogeneity into typical infectious disease models. It also undermines the common assumption that increased R0 equates to increased prevalence, arguing for an incorporation of host movement into the modeling of such relationships.