Directional effects in spatial models of infectious disease outbreaks.

Music City Center 

In many epidemic systems, the disease can be prone to spread in some directions more than others. This can be due to migration and behavior patterns or due to prevailing wind patterns. Individual-level models (ILMs) are commonly used for modelling spatial risk in infectious disease transmission but have not traditionally considered these directional tendencies. A class of ILMs that allow for the directional dynamics of disease transmission is introduced. In these directionally dependent ILMs, the probability of an individual being infected depends on both the direction and distance between susceptible and infectious individuals. The characteristics of these directionally dependent ILMs are discussed, and how they can be fitted in a Bayesian Markov chain Monte Carlo (MCMC) framework are shown. Results will be illustrated with both simulated data and real data from outbreaks of seasonal influenza and foot-and-mouth disease in livestock.

Individual-level models

Epidemic models

Circular distributions

Bayesian Markov chain Monte Carlo (MCMC)