Now showing 1 - 5 of 5
  • Publication
    Combining expert knowledge and machine-learning to classify herd types in livestock systems
    A detailed understanding of herd types is needed for animal disease control and surveillance activities, to inform epidemiological study design and interpretation, and to guide effective policy decision-making. In this paper, we present a new approach to classify herd types in livestock systems by combining expert knowledge and a machine-learning algorithm called self-organising-maps (SOMs). This approach is applied to the cattle sector in Ireland, where a detailed understanding of herd types can assist with on-going discussions on control and surveillance for endemic cattle diseases. To our knowledge, this is the first time that the SOM algorithm has been used to differentiate livestock systems. In compliance with European Union (EU) requirements, relevant data in the Irish livestock register includes the birth, movements and disposal of each individual bovine, and also the sex and breed of each bovine and its dam. In total, 17 herd types were identified in Ireland using 9 variables. We provide a data-driven classification tree using decisions derived from the Irish livestock registration data. Because of the visual capabilities of the SOM algorithm, the interpretation of results is relatively straightforward and we believe our approach, with adaptation, can be used to classify herd type in any other livestock system.
      70Scopus© Citations 8
  • Publication
    Decision support beyond total savings—Eligibility and potential savings for individual participants from changes in the national surveillance strategy for bovine viral diarrhoea (BVD) in Ireland
    Surveillance and management of livestock diseases is often evaluated with reference to expected sector-wide costs. In contrast, we calculate losses or savings for individual herd owners of a change in monitoring strategy during a national cattle disease eradication programme: bovine viral diarrhoea (BVD) in Ireland. The alternative strategy differs in how the disease is identified; by its sample- rather than census-based approach; and by its greater cost per test. We examined the costs faced by each breeding herd if testing were conducted using serology on a sample of young stock, in contrast to the current method of tissue-tag testing of all newborn calves. Following best knowledge of the likely costs, the following input values were used: i) €2.50 per test for tissue-tag testing and €7.66 for serology, ii) serology conducted on a sample of 10 young stock per management group from either the 6–12 month or 9–18 month cohorts; iii) 3 scenarios for the number of management groups: one per herd (M∞), one per 100 cows (M100) and one per 50 cows (M50). We found that many herds would often not be able to supply a suitable sample of young stock for serology or would face higher testing costs than when using tissue tag testing. The largest number (25%) of herds would benefit from participating in the change if sampling were done in October. These could annually save between €2.1 million under M∞ and €0.8 million under M50 (€108 - €49 per herd). However, analysing herd-level data we found that 90% of all Irish breeding herds would save less than €1.42 per cow or €99 in total per annum under M∞ and €0.59 per cow or €36 in total under M50. In a sensitivity analysis, we allowed serology costs to vary between €2 and €10 per animal. Herds at the 10 t h percentile of most savings made from switching would save at most €155 (M∞ at €2 per serology test) but would not save anything under M50 at costs ≥ €10. We conclude that, under these assumptions, the expected reduction in testing costs for the majority of beneficiaries would barely outweigh the practical implications of the strategy switch or the risks to the eradication programme associated with sample based surveillance. This study does not assess the cost-effectiveness of alternatives post-eradication.
      286Scopus© Citations 3
  • Publication
    Epidemiology of age-dependent prevalence of Bovine Herpes Virus Type 1 (BoHV-1) in dairy herds with and without vaccination
    Many studies report age as a risk factor for BoHV-1 infection or seropositivity. However, it is unclear whether this pattern reflects true epidemiological causation or is a consequence of study design and other issues. Here, we seek to understand the age-related dynamics of BoHV-1 seroprevalence in seasonal calving Irish dairy herds and provide decision support for the design and implementation of effective BoHV-1 testing strategies. We analysed seroprevalence data from dairy herds taken during two Irish seroprevalence surveys conducted between 2010 and 2017. Age-dependent seroprevalence profiles were constructed for herds that were seropositive and unvaccinated. Some of these profiles revealed a sudden increase in seroprevalence between adjacent age-cohorts, from absent or low to close to 100% of seropositive animals. By coupling the outcome of our data analysis with simulation output of an individual-based model at the herd scale, we have shown that these sudden increases are related to extensive virus circulation within a herd for a limited time, which may then subsequently remain latent over the following years. BoHV-1 outbreaks in dairy cattle herds affect animals independent of age and lead to almost 100% seroconversion in all age groups, or at least in all animals within a single epidemiological unit. In the absence of circulating infection, there is a year-on-year increase in the age-cohort at which seroprevalence changes from low to high. The findings of this study inform recommendations regarding testing regimes in the context of contingency planning or an eradication programme in seasonal calving dairy herds.
      53Scopus© Citations 5
  • Publication
    The Irish cattle population structured by enterprise type: overview, trade & trends
    Background: The cattle sector is the most important economic production unit of the Irish farming and agri-food sector. Despite its relevance, there has been limited quantitative information about the structure of differing cattle production types and of the connections between them. This paper addresses this gap by providing, for the first time, an overview of the Irish cattle population structured by enterprise type. Methods & Results: We collected data from the cattle register for the period 2015 to 2019 and assigned registered herds to one of 18 different herd types using a recently published herd type classification approach. This allows, for the first time, to exploring changes in enterprise types and subtypes over time, and describing the movements between these subtypes and from these subtypes to slaughter. Conclusions: The overview and associated classification presented in this study will form the basis for a number of future comparative studies, including cross-sectoral assessments of profitability, estimation of the extent of animal health losses on Irish cattle farms or structural analysis of Greenhouse Gas (GHG) emissions across production systems.
      48
  • Publication
    Risk factors for detection of bovine viral diarrhoea virus in low-risk herds during the latter stages of Ireland’s eradication programme
    Background: A national programme to eradicate bovine viral diarrhoea (BVD) has been in place in Ireland since 2013. To inform decision making in the end stages of eradication, and support the development of posteradication surveillance strategies, an understanding of risks of infection in a low prevalence system is required. Methods: A case-control study design was implemented. The study population comprised bovine herds that had calves born and tested negative for BVD virus (BVDV) every year from 2013 to 2019 (n = 46,219 herds). We defined cases as herds which had one or more test positive calves for the first time in 2019 (n = 204). Controls (n = 816) were randomly sampled from the herds which remained test negative in 2019. The effects of herd size, management system, inward movements, including those of potential trojan dams (pregnant animals brought into the herd that could potentially be carrying infected calves in utero), and proximity to herds testing positive in the preceding year, were investigated. Network analysis approaches were used to generate variables measuring connections with test positive herds through inward cattle movements. A generalised linear mixed model, including a county-level random effect, was used to explore these risk factors. Results: Our final model retained ln (herd size) (Odds Ratio (95% CI): 1.72 (1.40, 2.12)), distance from test positive herds (0.54 (0.44, 0.66) for each extra land-parcel boundary crossed to reach the closest herd which tested positive the preceding year), and ln (potential trojan dams + 1) (1.29 (1.05, 1.60)). The same variables were retained in the model where herds with confirmed transient infections only (n = 25) were excluded. Conclusions: Our findings suggest that care with biosecurity at farm boundaries and visitors and equipment entering the farm, and avoidance or careful risk assessment of purchasing potentially pregnant animals, may help prevent introduction of BVDV to low-risk herds. At policy level, consideration of herd size, proximity to test positive herds and purchasing patterns of potentially pregnant cattle may help target surveillance measures towards the end of the eradication programme.
      42