on pig premises relevant for disease transmission between farms is low. Although farmers selling their products on the premises may often receive visits from external visitors, these are not likely to visit other pig farms on the same day. An intermediate frequency was assigned to wild boar as porcine disease pathway, even though specifically considering their role for between-farm transmission only and not for disease introduction. Indeed, in Switzerland areas densely populated with wild boar overlapped with regions with high density of farms with outdoor access for pigs , which may facilitate farm-to-farm spread of pathogens via wild boar movements.

Factors modifying the relevance of disease pathways

The aforementioned role of outdoor access for pigs as modifying factor for the exposure to wild boar is one example of how farm characteristics should be considered when interpreting our findings. Production type is another important factor. While upstream farms in the production chain, such as breeding farms, are traditionally considered a bigger threat for disease spread via pig trade, among our sample of farmers we observed that also fattening and closed-cycle production farms could potentially infect other farms via indirect disease pathways (Table 2, quotes from interviews 1, 4, 5, 6, 13, 17 and 19). When modelling the transmission of PRRSV, indeed found that fattening farms were involved in local disease spread. Another example of production type influencing contact pathways is that of carcass collection on the farm. This pathway was found by the experts to be frequent, but since carcass collection centers mainly offer this service for carcasses weighing 200 kilograms or more, it may more often occur on farms owning old breeding animals.
Following biosecurity protocols effectively mitigates the risk of disease transmission, especially for the indirect pathways that cannot be eliminated even during outbreaks, such as veterinarian visits . Trust might be an important aspect, when considering indirect pathways. A higher trust in colleagues or external collaborators might be associated with lower biosecurity (e.g., if biosecurity measures are not observed by the trusted party). Oppositely, the trust or distrust placed in colleagues and external collaborators might serve as a protective factor for disease spread. Among the interviewed farmers, this could be observed in several situations involving both colleagues (e.g., trusting colleagues to dispose of pig carcasses from their farm, or not trusting them to use their equipment) and external collaborators (e.g., trusting feed advisors to enter the stables, or not trusting carcass collection services to come on the premises). In future research it might be advantageous to consider psychological factors, such as trust, as vulnerability or protective factors.

The Mental Models Approach

The use of an adapted MMA in this context was found to present multiple advantages. First, the MMA provided the flexibility to be able to include further pathways in every step of data collection. Thanks to that, we gained knowledge on pathways that had not been considered in past works. Second, accessing the combined knowledge of farmers and experts proved to be a powerful tool for evaluating the importance of indirect disease pathways, as both deal with issues related to pig production and health, but from different perspectives. Third, interviewing farmers by means of qualitative interviews allowed us to simultaneously identify pathways and learn about farmers’ attitudes and perceptions towards them. Open face-to-face interviews were very effective in this process, as they greatly facilitated the building of trust towards interviewers and the sharing of information on sometimes frowned upon farming practices . Finally, expert elicitation enabled the generalization of findings from the limited sample of interviewed farmers to the country context, and to assess their relevance for the spread of three diseases of national and international concern.

Study limitations

The limitations of our study are mainly those associated with qualitative assessments of the pathways that are typically undertaken relying on limited samples. The number of elicited experts is small, but justifiable given the size of the country and the limited number of available experts in the field, as previously reported by another expert elicitation study in Switzerland . We coped with this limitation by selecting an expert sample as diversified as possible, and we provided the counts of single expert answers given during the workshop, which can be used to interpret with caution the relevance of pathways with low agreement among experts. The small sample of farmers did not allow for the investigation of the association between the responses provided by farmers and farm characteristics. A further quantitative study may be considered to draw further conclusions in these regards. Nonetheless, some of the associations are relatively straightforward, as in the case of the wild boar pathway and outdoor access for pigs and can be extrapolated by using national agricultural and other databases. Finally, the presented frequencies of disease pathway occurrence are based on expert opinions given in the absence of outbreaks in the country. This may significantly vary through behavior change of farmers and professionals in reaction to disease incursion. This shift in behavior must be considered when modelling disease spread after first detection.