In both MCA solutions the 1st dimension was primarily related to whether a herd was open or closed with respect to replacement animals, and the 2nd dimension described how replacement animals are handled in open herds. The absence of discrete patterns on these plots indicates broad variation in how different protocols are applied
. The 3rd dimension may provide more insight into these relationships, however we can obtain useful information based on the general differences between the extremes. In both MCA solutions, better practices are grouped in one quadrant, poorer practices are grouped in another quadrant, and practices that are associated with herds being closed with respect to replacement animals are grouped together on the left side of the plot. In previous work by this research group, two-step cluster analysis identified three external biosecurity groups using a different subset of variables from the same dataset
. Herds belonging to the high biosecurity group that was closed with respect to replacement animals generally did not receive replacements from outside the production system, and the movement of animals was usually via dedicated trucks. Herds belonging to the high biosecurity group that was open with respect to replacement animals generally had higher trucking standards for the movement of animals and feed, and higher entrance sanitation requirements. Herds belonging to the low biosecurity group tended to rely on replacement animals from outside the production system, and did not have strict policies regarding the trucking of live animals or feed
. In the current study, the locations of the biosecurity groups on both MCA plots agree well with previous work. The high biosecurity group that was closed with respect to replacement animals was closely associated with responses of “not applicable” for the introduction and transportation of replacement animals. In both MCA plots, the high biosecurity group that was open with respect to replacement animals was located in the quadrant with the least risky strategies for the handling of replacement animals, whereas the low biosecurity group was located in the quadrant with the most risky set of strategies.
The most important finding of this study is the understanding of how individual biosecurity practices form biosecurity strategies, particularly with respect to the introduction of replacement gilts into sow herds. Within these strategies, we can expect to find some practices that are generally considered high-risk, accompanied by other biosecurity practices that mitigate the risk. For example, in the MCA solution concerning the introduction of replacement gilts, the practice of having 4 or more replacement sources in the previous 2 years was closely associated with biosecurity practices that mitigated the associated risk, such as moderate isolation and acclimation periods that occurred in facilities located off-site, and blood testing of all replacements for PRRSV upon exit from the isolation/acclimation facilities. In order to ensure replacements are not viremic, it is advised that all replacement gilts are tested prior to entry into the main herd
. The use of multiple sources for replacement animals may be a necessity created by modern pork production systems, in order to facilitate improvements in the breeding program; the strategy described above mitigates the risk posed by this practice. At the other extreme are herds that were obtaining gilts from within the production system, some of which were introducing them after an isolation/acclimation period of between 1 and 60 days. Regardless of the source, isolation and acclimation of replacements is essential. Incoming pigs may appear healthy but be incubating infection or acting as carriers of a pathogen, and the likelihood of transmission to susceptible pigs is increased by the stress inherent to loading, mixing, and transportation of these animals
[12, 13]. Even when using PRRSV-negative suppliers, pigs may come into contact with the virus during transportation
. In the MCA solution concerning the transportation of replacement animals, some regions of the plot contain both good and poor practices in close proximity. Although the 3rd dimension may provide more insight into these relationships, generally these inconsistencies indicate broad variation in application of biosecurity protocols in this particular quadrant of the MCA plot. It is important to consider that vehicles that transport livestock are known to play a role in the spread of PRRSV from contaminated premises, and vehicle cleanliness is important in mitigating the associated risk
The complex interrelationships between these biosecurity variables cannot be easily examined by correlation coefficients alone, and MCA has proved useful in that respect. The findings of this study could have important implications for the assessment of biosecurity practices, since our results suggest that infection control should not rely exclusively on the benchmarking of individual practices against an ideal standard. Additionally, the entire strategy should be assessed simultaneously; the implementation of such strategies is likely driven by their feasibility, cost, and effectiveness. The practical application of this finding is that standard-setting agencies should not only look at promoting specific individual biosecurity practices. Groups of practices that form strategies should be examined in order to determine whether the strategy is designed to effectively reduce the risk of introducing pathogens. This idea aligns with the principle of equivalence set forward by the World Organization for Animal Health, which recognizes that different approaches to animal health and production systems can provide equivalent animal and human protection for the purposes of international trade
. Thus, encouraging producers to apply the same biosecurity standard for every individual practice may be an over-simplified approach for some aspects of biosecurity, such as the introduction of replacement gilts.
The results obtained from this study are subject to some limitations. The herds used in this study represent a convenience sample that was recruited without a formal selection process; this may be the biggest limitation of the study. We are unable to provide response rates, as this study was an industry-based project, and those statistics were not available to us. Nonetheless, the variation in herd size within our sample indicates that a broad variety of management styles were included in the study. Additionally, participation was voluntary, and although not a specific requirement, herds were more likely to be selected for participation if their veterinarian was a member of OASV. Our study group may differ from the source population as a result of these potential selection biases
. As with any survey data, the use of closed-ended questions may mean that some information regarding gilt replacement strategies could have been misclassified, resulting in a potential bias
. Our decision to include the external biosecurity groups from previous work may also serve as a limitation of this study. Some of the variables used in the two-step cluster analysis were also used in the MCA solutions presented here. For the introduction strategies, 4 variables were used in both methods; for the transportation strategies, 2 variables were used in both methods. This was done to allow a more complete assessment of biosecurity, as it relates to these two important areas. Additionally, the biosecurity groups were not used in the MCA solutions; they were supplementary variables that aided in our interpretation of the plots.