Risk factors analysis of ASF in wild boar

Data on African Swine Fever (ASF) positive and negative reports in wild boar (Sus scrofa) were obtained for Latvia, Lithuania and Italy as the three target countries identified by EFSA. Additional data from Belgium and Sweden were also obtained. All these data were considered of suitable quality in terms of geo-location, but only data from Latvia and Lithuania were suitable for a full risk assessment, since data from the other locations were very tightly clustered, with data from Sweden and Italy to be used only in a single analysis. A list of potential risk factors was examined, but for several of them the spatial variability in the target countries was small, so they had to be excluded. Two statistical approaches were used to examine ASF occurrence and ASF persistence, respectively. The final models to explain the factors associated with ASF were reasonably good but could be hardly extrapolated across Europe, since some factors had limited variability within the target countries. For the occurrence model, the important factors to predict ASF were climatic (temperature and rainfall), with domestic pig density and wild boar density of only moderate importance in this scenario (Latvia and Lithuania). For ASF persistence similar climatic factors, along with various habitats were significant. Habitat features are difficult to summarise, as mosaic cropland (>50%), forest fragmentation and mosaic tree cover increase risk, but too much tree cover, cropland, water bodies and urban areas decrease risk. Domestic pig density and wild boar density were not important driving factors, although the latter could be due to the limited variation in the study area. Under the modelled scenario, our findings suggest that ASF in wild boar was not driven by a simple relationship with host presence or density, but in combination with habitat features that promote wild boar connectivity (such as a mosaic habitat) and a climate that promotes infectivity (of individuals or carcasses). We therefore strongly recommend that all ASF and in general all wildlife disease investigations are reported together with precise spatial data and that all negative reports are included, as this would vastly improve the ability to perform future analyses