Understanding African Swine Fever infection dynamics in Sardinia using a spatially explicit transmission model in domestic pig farms

African swine fever virus (ASFV) has been endemic in Sardinia since 1978, resulting in severe losses for local pig producers and creating important problems for the island's veterinary authorities. This study used a spatially explicit stochastic transmission model followed by two regression models to investigate the dynamics of ASFV spread amongst domestic pig farms, to identify geographic areas at highest risk and determine the role of different susceptible pig populations (registered domestic pigs, non-registered domestic pigs [brado] and wild boar) in ASF occurrence. We simulated transmission within and between farms using an adapted version of the previously described model known as Be-FAST. Results from the model revealed a generally low diffusion of ASF in Sardinia, with only 24% of the simulations resulting in disease spread, and for each simulated outbreak on average only four farms and 66 pigs were affected. Overall, local spread (indirect transmission between farms within a 2 km radius through fomites) was the most common route of transmission, being responsible for 98.6% of secondary cases. The risk of ASF occurrence for each domestic pig farm was estimated from the spread model results and integrated in two regression models together with available data for brado and wild boar populations. There was a significant association between the density of all three populations (domestic pigs, brado, and wild boar) and ASF occurrence in Sardinia. The most significant risk factors were the high densities of brado (OR = 2.2) and wild boar (OR = 2.1). The results of both analyses demonstrated that ASF epidemiology and infection dynamics in Sardinia create a complex and multifactorial disease situation, where all susceptible populations play an important role. To stop ASF transmission in Sardinia, three main factors (improving biosecurity on domestic pig farms, eliminating brado practices and better management of wild boars) need to be addressed.