Discussion
Although there was an initial rapid increase in the number of SOAs when they were introduced, their numbers have only been increasing gradually in recent years. Furthermore, the volume of movements attributed to SOAs over recent years remains relatively constant. However, the analysis did reveal a number of points of potential concern. There appear to be very few intra-SOA movements recorded in the AMLS database despite a requirement that all such movements be reported. This could be explained by the previous 5 mile rule for sheep and goats, whereby movements under 5 miles between land under the same ownership and control did not need to be reported. As holdings within SOAs tend to be located very near each other, and are under the same control, this could well explain the lack of intra-SOA movements; we note that the 5 mile rule for sheep and goats has recently been disbanded [11] and now all movements between different CPHs must be reported. An alternative explanation for the lack of intra-SOA movement data is that there could be confusion on the reporting requirements with different rules governing SOAs and CTS Links combined with the 5 mile rule for sheep and goats.
Our modelling work showed that SOAs do pose a potential risk of increasing the number of holdings infected in an epidemic considerably, if the worst case scenario of “one infected – all infected” is assumed. However, the majority of the effect of increased epidemic size is the “book-keeping” increase in infected holdings within SOAs themselves. Furthermore, SOAs did not greatly effect the geographical spread of the epidemic, which is to be expected given that the component holdings within an SOA appear to be located near each other, typically within 10 km. Overall, the effect in terms of epidemic size of including SOAs into simulations appears to be mainly due to intra-SOA spread, in that infection events as a result of the disease spreading within the SOA contributes a large proportion to the overall epidemic size. There is currently insufficient information on stock levels and intra-SOA movements to determine whether or not this worst case is relevant, and so it is an indicator of a need for future data collection.
If the reported stock levels are correct, in that typically an SOA only has stock located at one holding, then SOAs would not appear to be a major risk in terms of epidemic size, as holdings without stock can not contract and therefore pass on the disease via local and movement based spread. This is consistent with additional registered CPHs within an SOA representing grazing land with no permanent stock. For example, within an SOA of five holdings that are used for grazing, at any one time only one holding may have stock on it, thus limiting the possibilities for disease spread. However, Figure 7 shows that there is a considerable difference if premises are stocked more evenly, suggesting that a more detailed understanding of intra-SOA movements and stock levels would be useful. The simulation results presented here are in terms of infected holdings, whereby intra-SOA spread results in more holdings becoming infected. However, if the majority of SOA holdings have no stock, then the number of livestock slaughtered within the SOA will increase little as a result of intra-SOA spread. It should also be noted that all valid SOAs are included in the above simulations. The data analysis did suggest that there could be a redundant accumulation of SOAs, as few SOAs seem to leave the scheme, and a fair amount of SOAs appear to be inactive in terms of movement into or out of the SOA on AMLS. Given the potential risk of SOAs in Figure 7, this could suggest that a clean up of inactive SOAs would be valuable in assessing the true risk associated with SOAs.
The number of CTS Linked holdings, especially those linked via Shared Facilities, appears to be increasing rapidly and the volume of movements from such CTS linked holdings is also increasing. This is in contrast to AMLS movements from SOAs, which remained relatively constant over time, and suggests that there is little redundant accumulation of inactive holdings within the CTS Links. Currently, there is neither a limit on the number of CTS Links an individual holding can be in, nor a limit on the distance between CTS linked holdings via the Additional Land category. Combined, these features can lead to large chains of linked holdings that can also be geographically dispersed throughout the country; however, the vast majority of chains only contain 2 holdings. These chains can then grow even further as there is currently no restriction on being within an SOA and in CTS Links at the same time.
In the worst case scenario (all members of a CTS Chain potentially directly linked, or linked within a short time), large and geographically dispersed CTS Chains are potentially a risk in terms of disease spread. Applying a 50 km distance limit on transmission amongst CTS Links markedly reduces epidemic size, suggesting that allowing long-range links or a short series of links covering a large distance may be an important risk. However, further investigations should be made as to whether or not the CTS Chains are also transmission chains – if multiple links in a chain are not used in the timeframe of FMD spread (at most a few weeks for single holdings), then they are not important in this context. Even if all links in the chain were active, it would take time for the disease to spread via cattle movements through each binary link in the largest chain – this was not considered here as we used a “one infected – all infected” approach to intra-CTS Chain spread similar to intra-SOA spread. Furthermore, while movements are not recorded, standstills are still expected to be in effect, and investigation into standstill compliance within chains may be merited.
The scenarios presented here examine worst-case scenarios for disease spread, with members of SOAs assumed to have the same infection status throughout. In reality, this would only be the case in a limited number of instances in a real outbreak, and although our data analysis does highlight a number of points for further investigation, it is important to balance the need for adequate disease prevention and control with commercial activity so that holdings can perform competitively. Shared Facility links where holdings share milking parlours are clearly essential, and would likely involve neighbouring farms. In terms of FMD spread, if one holding were to become infected, any neighbour would be at high risk of infection, irrespective of whether an SOA or CTS Link were present. One could question the need to link two holdings that are very geographically dispersed, as it would be unlikely that cattle movements between the two holdings would be frequent. Therefore, distance limits on Additional Land links may well be considered. We note that there was previously a 50 km distance limit between holdings within an SOA, although this has now been disbanded. Furthermore, a limit on the number of CTS Links that an individual holding can be in may well reduce the observed CTS Chains to smaller less risky units. The increase in the use of CTS Links as well as SOAs could simply reflect changes within the farming industry. If there are fewer farms, those that remain may become bigger through expansion into old farm land, but distinct CPHs must still remain due to distance rules on how far farm land can be from the main steading.
Broadly speaking, our analyses suggest that while SOAs and CTS Links are important, they are important because of the number of holdings so registered, rather than because of the activity associated with them. Under a worst case scenario of “one infected – all infected” both SOAs and CTS Links pose a risk of increasing epidemic sizes, but further investigations are needed into whether CTS Chains are transmission chains, and also investigations into intra-SOA movements and livestock distributions, in order to develop realistic alternative scenarios.