Disease transmission at the wildlife-livestock interface: multi-species modeling of bovine tuberculosis in Michigan

Background/Question/Methods

Disease causing organisms are rarely restricted to a single-host. Multiple host species may be involved as a part of a parasite’s complex life-cycle, as dead-end hosts that are infected but are unsuitable for continuing transmission, or as part of a system of host species with continuous intra- and inter-species transmission. We present a model of coupled transmission of bovine tuberculosis (bTB, Mycobacterium bovis) among a free-ranging white-tailed deer (Odocoileus virginianus) population and a spatially-structured cattle population applied to the case of bTB in Michigan, USA. We ask, can observed bTB dynamics in both species be maintained as if both species are part of a single interacting host population? Or is one species the primary reservoir with directional spillover to the other?  We developed a spatially-explicit farm-based susceptible-exposed-infected-removed (SEIR) model of bTB transmission dynamics that includes local farm-farm transmission (such as shared equipment, shared feed, etc.) and long-distance transmission parameterized by the observed cattle transport network. We developed a non-spatial white-tailed deer age-structured SEI model parameterized by disease surveillance and demographic data in Michigan. We coupled the models by defining wildlife-farm contact using GIS layers to quantify deer habitat surrounding each farm.

Results/Conclusions

We found that observed patterns of spatially clustered low-prevalence bTB can be generated by separate transmission cycles in cattle and wildlife. However, relatively modest reductions in transmission among farms results in disease extinction without transmission from the wildlife reservoir. In addition, transmission to spatially dispersed areas is unlikely under scenarios with only local-scale transmission among livestock farms and wildlife; a mechanism for dispersal of the disease agent (such as truck transports of animals) is necessary to generate spatial dispersion of the disease. We conclude that bTB in Michigan is best described by a system with bi-directional transmission among multiple hosts species.