Spatial and risk factor analysis of bovine viral diarrhoea (BVD) virus after the first-year compulsory phase of BVD eradication programme in Northern Ireland

Abstract

Bovine viral diarrhoea virus (BVDV) causes bovine viral diarrhoea (BVD), which is a contagious pathogen that can have a significant economic impact on cattle industries. In Northern Ireland (NI), the compulsory phase of a BVD eradication programme was implemented in 2016. The aim of this retrospective population based study was to utilize herd-level data after the first year of the compulsory phase (March 2016–March 2017) to determine the spatial distribution and variation of BVDV, to identify clusters of infection, and to quantify some risk factors associated with BVD in NI. Global spatial clustering (autocorrelation) and local spatial hot-spot analyses were used to specify the clustering areas (hot- and cold-spot). A suite of multivariable logistic analyses was performed to estimate the associations of spatial and non-spatial factors (relating to herd characteristics) with the risk of being a BVDV positive herd. Final models were compared by evaluating the model fit and the ability to account for spatial autocorrelation in the study area. There were 17,186 herds included in the analysis. The herd-level prevalence of BVDV was 11.31%. Significant spatial clustering of BVDV positive herds was presented in the central region of NI. A mixed effects logistic model, with a spatial random effect term, was considered the best model. The final model showed that a positive BVDV status during the voluntary phase prior to the compulsory phase started (OR = 2.25; CI 95% = 1.85–2.73), larger herd size (OR = 6.19; CI 95% = 5.22–7.34 for herd size > 100 animals) and a larger number of positive nearest neighbours within 4 km radius (OR = 1.24; CI 95% = 1.05–1.47 for 8–9 neighbours and OR = 1.41; CI 95% = 1.20–1.65 for 10–12 neighbours) were significantly related to the risk of a herd being tested positive for BVDV. The clear spatial pattern from the local spatial clustering analyses could be used for targeted surveillance and control measures by focusing on the central region of NI.