3.5 Risk analysis
A total 150 respondents answered this section, thus 70% of the respondents produced risk assessments and used the available tools. The Poisson model obtained (with the outcome being the Linkert coded score and the explanatory variable being the type of risk assessment approach) showed that the qualitative risk assessment approach was the one most used followed by quantitative and semi-qualitative (Figure 8a) . Quantitative and semi-qualitative risk assessment had no significant difference between them. When asked about the type of risk assessment they generally worked on, there was no difference among release, exposure and consequence assessment. The Poisson model with the type of risk assessments as the explanatory variable suggested absence of significance for all the factors. The percentages in all the five categories of the Linkert scale were similar for all the three types of risk assessment (Figure 8b) . When asked which was the primary objective of producing a risk assessment, in decreasing order of frequency of the always category were: “ Risk assessment and preparedness in areas involving origin and spread of animal diseases, including zoonoses”, “Provide stakeholders with relevant information and expert advice on issues related to disease preparedness and surveillance of animal diseases and zoonoses”, “identify key questions for targeted research”, “provide veterinary diagnostic laboratory services for zoonotic, epizootic and other animal notifiable diseases”, “evaluate the need for action to support policy changes”, “identify plausible future scenarios to be prepared to future animal incursions”(Figure 8c) . The Poisson model applied to determine the primary objective of respondents’ risk assessment suggested that the categories “to identify key questions for targeted research”, “to provide veterinary diagnostic laboratory services for zoonotic, epizootic and other animal notifiable diseases” and “identify plausible future scenarios to be prepared to future animal incursions” were of significantly higher importance than the reference category “risk assessment and preparedness in areas involving origin and spread of animal diseases”. The other two primary objectives: “evaluate the need for action to support policy changes”, and “provide stakeholders with relevant information and expert advice on issues related to disease preparedness and surveillance of animal diseases and zoonoses” were not significant with regards to the reference category. Therefore, they had the same importance of being the primary objective as risk assessment and preparedness in areas involving origin and spread of animal diseases.
A Poisson model was applied to study the most important feature required by experts when using a risk analysis platform tool. Significance for each category was suggested when using the category “Produce a risk assessment for two diseases for comparison” as the reference category. In decreasing order of importance: “spread assessment”, ” pathways of introduction of a disease until the border”, “produce a quick risk assessment”, “Produce a report using the system”, “Produce a risk assessment detailed for a single disease”, “Produce a risk assessment for two diseases for comparison”, with the reference category being the least important (Figure 9).
To assess the risk of threat, 85% of the respondents used different data sources and tools. To report a risk assessment result however, 67% answered that they did not combine the outputs of several types of methods.