One of the most important routes for the spread of infectious diseases in animal populations is movement of live animals. European Union legislation requires all bovine animals to be registered in national databases. Moreover, the databases must also contain date of birth, identification number of the holding were the animal was born, identification numbers of all holdings where the animal has been kept and dates of each change of holding and date of death and slaughter.

The use of real population and event data enables realistic disease spread modeling of the spatio-temporal dynamics due to age structures, trade patterns, population size and slaughter. However, performing detailed data driven simulations requires efficient algorithms and data structures to handle the complex network of herds and events.

Verotoxinogenic Escherichia coli O157:H7 (VTEC O157:H7) is a zoonotic pathogen of great public health interest. VTEC O157:H7 is capable of causing severe enteroheamorrhagic colitis in humans, notably children. Cattle is considered to be a major reservoir of the pathogen and shed the bacteria in feces without showing any clinical signs of disease.

In this work we study spatio-temporal models of the spread of VTEC O157:H7 in the Swedish cattle population where epidemiological processes are formulated as continuous-time Markov-chains. To accelerate computations in our discrete-event simulator SimInf we divide work among cores of shared memory multiprocessors by using the models inherent parallelism in between transport events gathered from actual data. We will present the epidemiological model and study the overall parallel efficiency of the simulation.