Veterinary decision-making in infectious disease control often involves interacting factors across individual characteristics, professional practices, contextual, and regulatory levels. Traditional analytical approaches rarely capture this complexity. To address this challenge, we applied a regularized multiblock generalized canonical correlation analysis (rGCCA) to field survey data from European swine veterinarians in the context of porcine reproductive and respiratory syndrome virus (PRRSV) management to identify key factors associated with perceived success in PRRSV stabilization programs. A cross-sectional survey was conducted among 108 veterinarians from 22 European countries. A total of 123 variables were organized into six conceptual blocks: socio-professional characteristics, professional exchanges, evidence-based veterinary medicine (EBVM) engagement, regulation perceptions, PRRSV epidemiological context and management practices, and country. After univariate selection (p < 0.15), 71 variables were retained as potentially associated with perceived PRRSV stabilization success. A rGCCA model with two components per block was applied, followed by bootstrap resampling (500 iterations) to assess variable stability and block-to-block associations. The first component captured partial success in PRRSV stabilization programs, mainly related to insufficient cooperation and technical constraints. The second component reflected complete success and was strongly associated with country, professional exchanges, and contextual practices blocks. Notably, veterinarians reporting higher EBVM integration, greater teamwork frequency, and easier access to diagnostic and vaccination tools reported higher perceived stabilization success. Cross-block correlations highlighted significant interdependencies between geographical context, regulatory perception, and professional networking. This study demonstrates the potential of multiblock analysis to straighten out complex field data with cross-relationships between domains in veterinary epidemiology. The rGCCA framework provides an interpretable, replicable approach to explore multidimensional relationships and identify practitioner profiles associated with perceived disease stabilization success. These findings should be interpreted in light of the subjective nature of the perceived success outcome and do not necessarily reflect objective virological or production outcomes. Our methodological contribution supports the integration of data-driven, evidence-based strategies into veterinary epidemiology and disease control decision-making.
Grazing is widely promoted for its potential welfare and environmental benefits, yet its economic implications for animal health remain unclear. Therefore, this study quantified annual disease-related costs in confinement and grazing systems on dairy farms using a deterministic bio-economic model, providing system-level disease cost estimates. Four diseases were included: ketosis, lameness, mastitis, and parasitic infections. A typical Dutch dairy farm was simulated over one year under full confinement and under grazing comprising six months pasture access and six months indoor housing. Total disease-related costs were estimated at €349 per cow per year under confinement and €317 under grazing, a difference of €32. In confinement, ketosis, lameness, mastitis, and parasitic infections accounted for €20, €113, €215, and €0 per cow per year, respectively, whereas the corresponding costs under grazing were €19, €93, €196, and €9 per cow per year. Sensitivity analyses of market prices, disease effect parameters, and system-based milk yield differences affected the magnitude, but not the direction of this finding. Across all sensitivity analyses, disease-related costs in the grazing systems were estimated to be lower than in the confinement system. Accounting for uncertainty in incidence rates did not alter the relative difference between systems, although confinement showed slightly greater cost dispersion. Disease-related costs represented approximately 11.1% and 10.1% of annual gross margins under confinement and grazing, respectively. Overall, the results indicate that, under the assumptions of the model, grazing can offer moderate economic advantages from an animal health perspective.
Foot and mouth disease (FMD) re-emerged and lumpy skin disease (LSD) emerged in Indonesia in early 2022, posing significant economic threats to the feedlot industry, which imports Australian-bred Bos indicus cattle. While the industry rapidly improved biosecurity measures, many were leading to unnecessary overspend with minimal effectiveness. This participatory project aimed to empower Indonesian feedlots and associated abattoirs to implement cost-effective biosecurity. The project engaged 22 feedlots and 11 abattoirs through workshops, site-specific risk assessments, and tailored recommendations. Within six months, adoption rates reached 65% for feedlots and 75% for abattoirs, with feedlot adoption further increasing to 75% by the project's end at 12 months. This initiative enhanced biosecurity practices, improved understanding of disease transmission, and instilled greater confidence in biosecurity measures across the industry. The changes will ultimately contribute to a more sustainable industry and national food security for Indonesia.
A cross-sectional study was conducted in 2022 to estimate the true prevalence of Leptospira spp. exposed cattle herds in Ho Chi Minh City, Viet Nam and to identify characteristics associated with herd level Leptospira spp. exposure. Herd managers completed a questionnaire about herd management and whole blood samples were collected from their cattle and tested for Leptospira spp. serovars using the microscopic agglutination test (MAT). Multivariable logistic regression analyses were used to identify characteristics associated with seropositive herd status. Sixty of 99 dairy herds and 28 of 61 beef herds had at least one animal with Leptospira spp. antibodies detected using the MAT. For the total of 160 herds, the estimated true prevalence was 27 (95% CI 5-47) Leptospira spp. exposure positive herds per 100 herds at risk. The odds of being Leptospira spp. exposure positive for dairy herds with a median cow age of greater than 3 years was 3.8 (95% CrI 1.7-15) times that of dairy herds with a median cow age of less than or equal to 3 years. Given the high herd-level prevalence of Leptospira spp. exposure we conclude that leptospirosis is a risk not only to animal health and productivity, but also to herd managers and those who work with their stock in this area of Viet Nam. Action to minimise water pooling in the farm environment and to implement rodent control and cattle leptospirosis vaccination programs are strongly recommended as key measures to reduce infection pressure, pre….
In 2022, Australia experienced an unprecedented outbreak of Japanese encephalitis virus (JEV) across four states, resulting in seven human deaths. Occurring during a significant La Niña period, the event posed major challenges for detection and response across human and animal health sectors. This study explores the experiences of frontline veterinary and industry professionals to understand the operational and biological challenges encountered. We conducted a qualitative epidemiological study with 15 swine veterinarians, industry professionals and medical entomologists directly involved in the response, using thematic analysis to explore outbreak dynamics, perceived risk factors, and decision-making under uncertainty. Participants described marked heterogeneity in disease presentation, severity, and duration. Transmission was perceived to be vector-mediated and shaped by environmental and farm-level factors, with no field observations supporting direct pig to pig transmission. A key biological challenge was the delay between infection and clinical recognition in swine. On-farm mitigation focused primarily on mosquito control to reduce exposure in high-risk animals. Constraints included uncertainty regarding local transmission ecology, systemic surveillance gaps, limitations of diagnostic tests and the absence of a swine vaccine. Knowledge sharing and the need for adaptive, evidence-informed government responses were seen as important. These findings illustrate how climate-sensitive, vector-borne zoonoses challenge outbreak response frameworks, and demonstrate how integrating veterinary field observations and industry experience into One Health surveillance and response planning can strengthen preparedness for emerging vector-borne diseases in regions with commercial swine industries.
Tests detecting pregnancy-associated glycoproteins (PAG) are widely used for pregnancy diagnosis in dairy cows. Recently, a days-since-last-breeding (DSLB)-specific interpretation algorithm was proposed to account for known fluctuations in PAG concentrations during early gestation. Our objective was to evaluate the diagnostic accuracy and predictive values of this DSLB-specific interpretation for a milk-based commercially available pregnancy test (milk PAG-ELISA) compared with the manufacturer's uniform interpretation thresholds. Data were assembled from two published diagnostic validation studies conducted in Canada and South Africa, including 1260 Holstein cows from 24 commercial dairy herds tested between 28 and 50 DSLB. Milk PAG-ELISA results from manually collected samples were compared with veterinary ultrasonography examinations. Diagnostic accuracy was estimated using a Bayesian latent class model accounting for the imperfect nature of both tests and allowing accuracy to vary across DSLB intervals. When compared with the manufacturer's interpretative thresholds, the DSLB-specific grid improved PAG-ELISA specificity and positive predictive value for cows tested between 41 and 50 DSLB (+12 and +7 %age-point increases, respectively), while yielding similar accuracy for cows tested between 28 and 40 DSLB. However, pregnancy diagnosis in the 41-50 DSLB interval showed lower overall accuracy and predictive values regardless of the interpretation approach used, likely reflecting the physiological decline in PAG concentrations during that stage of gestation. These findings support the use of the novel interpretation grid to improve test performance between 41 and 50 DSLB and indicate that the accuracy of pregnancy diagnosis using milk PAG-ELISA was greater between 28 and 40 DSLB.
In hyperendemic settings, a reduction in notification curves may reflect that surveillance capacity is deteriorating rather than indicating epidemiological improvement; this distinction has direct consequences for disease control. Rio de Janeiro, Brazil, where cat-transmitted sporotrichosis has persisted for decades, provides a suitable setting to examine whether a decrease in reported cases is accompanied by persistent spatiotemporal clustering. In this ecological study, laboratory-confirmed notifications of feline sporotrichosis from 2017 to 2024 and stratified by administrative region were analyzed. Notification rates were calculated using estimated feline populations. Temporal patterns were investigated using generalized additive models, spatiotemporal clusters were identified using SaTScan, and ecological associations with the Social Progress Index were assessed using geographically weighted regression. In total, 14,885 feline cases were reported, with an average notification of 13.07 cases per 1000 estimated cats per year. Although the number of notifications decreased at the municipal level, the highest average rates remained concentrated in Santa Cruz, Guaratiba, São Cristóvão, and Campo Grande, and model-identified high-risk clusters persisted in the western, northern, and central regions. Lower values of the Social Progress Index were generally associated with a larger number of notifications, although local associations were spatially heterogeneous and do not reflect causal relationships. These findings demonstrate that declining municipal notifications coexisted with hotspots and heterogeneous reported risk patterns, suggesting that aggregate trends may obscure local disease persistence and surveillance gaps. In long-standing hyperendemic settings, territorial surveillance, sustained diagnostic access, and spatially targeted One Health strategies are necessary to translate notification data into actionable epidemiological intelligence.
This study assessed antimicrobial prescribing practices and perceptions of antimicrobial resistance (AMR) among veterinarians working in companion animal practice in Peru. A cross-sectional study was conducted using an open online survey administered between 2022 and 2023. Responses from 442 veterinarians were analyzed descriptively, and multivariable logistic regression models were used to identify factors associated with selected prescribing practices. Fever (82.2%) was the most frequently reported clinical indicator for initiating antimicrobial treatment. Approximately half of respondents reported always performing culture and antimicrobial susceptibility testing before prescribing antimicrobials. Overall, 82.9% of respondents reported being unfamiliar with the WHO and/or WOAH classification of critically important antimicrobials (CIA). Regarding last-resort antimicrobial use, 29.0% reported having used these agents, including carbapenems, for severe clinical conditions. Prescribing patterns varied across clinical systems, with penicillins combined with β-lactamase inhibitors exhibiting the highest overall reported frequency of use, particularly in respiratory, dermatological, and urogenital conditions. In multivariable logistic regression analysis, frequent use of culture and susceptibility testing was associated with higher odds of last-resort antimicrobial use (OR = 4.30; 95% CI: 1.94-9.54), whereas postgraduate training was associated with greater odds of knowledge of CIA (OR = 3.29; 95% CI: 1.65-6.54). These findings reveal heterogeneity in prescribing patterns, substantial knowledge gaps regarding CIA, and professional factors associated with practices relevant to resistance selection, providing evidence to support the strengthening of antimicrobial stewardship strategies in companion animal practice in Peru.
Porcine reproductive and respiratory syndrome virus is associated with substantial costs for the global pig industry. We conducted a systematic review and meta-analysis of 71 cross-sectional surveys (307,863 pig specimens from 27 provinces, 2004-2025) to quantify long-term spatiotemporal trends and associated factors of sample-level PRRSV RT-PCR positivity in mainland China. The overall pooled PRRSV RT-PCR positivity was 32% (95% CI: 29-34%) with high between-study heterogeneity (I² = 99%, P < 0.001); this estimate summarizes diverse study populations and does not represent uniform national prevalence. Pooled prevalence declined from 36% (95% CI: 31-41%) in 2004-2015-26% (95% CI: 23-29%) in 2016-2025, potentially driven by shifts in study design and sampling populations. Higher prevalence was associated with South China, small-scale farms, finishing pigs and summer sampling. Funnel plot asymmetry and Egger's test indicated significant small-study effects, likely reflecting publication bias and study heterogeneity. These findings confirm widespread PRRSV detection in sampled pig populations, with a temporal decline and identifiable putative risk factors associated with heterogeneity. Results provide a reference to optimize future PRRSV surveillance and control.
Porcine reproductive and respiratory syndrome virus (PRRSV) infections in growing pigs are common and costly. We evaluated risk factors related to production practices and herd characteristics associated to wild-type PRRSV (WT-PRRSV) introduction into 63 wean-to-finish (WTF) herds through the implementation of surveys, active recording of activities and monthly oral fluid sampling during the growing phase. Herds were located in the Midwest region of the USA and observed prospectively. Logistic regression analysis of risk factors indicated that WTF herds that outsourced trucks for mortality removal (OR 7.08, p = 0.03) and herds with external maintenance crews (OR 7.14, p = 0.03) had higher risk of WT-PRRSV infections. These findings highlight the importance of mechanical transmission of PRRSV and the need to develop stricter biosecurity protocols in growing pigs. Overall, this study describes most common production management practices and characteristics of WTF sites, and potential risk factors that if addressed, can help improve biosecurity, control and elimination of PRRSV in growing pigs.
Gastrointestinal nematodes (GIN) are a major health concern in dogs, particularly in high-density environments such as shelters. Toxocara canis, Ancylostoma spp., and Trichuris vulpis are widely distributed, with recognised veterinary and zoonotic relevance. Effective prevention requires improved understanding of environmental and management-related risk factors driving transmission. This study investigated determinants of GIN infections in shelter dogs across Portugal. Faecal samples (n = 189), collected between November 2021 and October 2022 from 19 shelters, were analysed using coprological (mini-FLOTAC) and molecular (targeted PCR) methods. Infection risk was assessed using Boosted Regression Trees, Random Forest models, and ecological niche modelling (ENM), integrating climate, land use, shelter management, and host-related variables. Key risk factors included housing density (number of dogs per enclosure), age group, and deworming regimen, with parasite-specific patterns. Environmental predictors varied across species, with solar radiation, temperature, and soil characteristics influencing T. canis distribution, while land cover and precipitation were more relevant for T. vulpis and A. caninum. Marked spatial differences in the ecological suitability of the three GIN were evident. Spatial predictions indicated widespread risk of T. canis and T. vulpis across mainland Portugal, whereas higher-risk areas for A. caninum were concentrated in southern regions. Altogether, our findings highlight the combined influence of environmental conditions and shelter management on GIN transmission. The integration of eco-epidemiological and spatial modelling approaches using machine learning and ENM provide a useful framework for understanding local and regional nematode transmission dynamics and identifying high-risk settings, supporting targeted surveillance and more effective, shelter- and region-specific parasite control strategies.
Highly pathogenic avian influenza (HPAI) H5 viruses are a major threat to the French poultry sector. All sectors are affected and virus transmission between sectors has been reported. The objective of this work was to identify the risk of virus transmission to the broiler and layer sectors from any other poultry sector in France through nine indirect transmission pathways. We used scenario tree modelling of data collected from the literature or expert knowledge. Three questionnaires were produced, one for each field of expertise: avian influenza, the broiler and layer sectors. The experts' estimates were combined using a beta distribution weighted by their confidence level. Monte Carlo iteration was used to combine the different probabilities of the scenario tree and to assess the risk of transmission. In the broiler sector, the probability of viral transmission was the highest when indoor broiler farms were exposed farms and duck farms were source farms and high when turkey and guinea fowl farms were source farms. In the layer sector, probabilities of transmission were similar between all types of farms as source farms and layer farms as exposed farms. Probabilities were very low between breeder farms but similar in all the sectors. This work highlights the role of some pathways, especially management of manure and dead birds in the probability of introduction into broiler or layer farms. These results can be used to improve management of the risk of HPAI H5 transfer between different poultry sectors in France.
Veterinary medicine is integral to the response to the global health crisis of antimicrobial resistance (AMR). Livestock and companion animals frequently harbor and potentially transmit resistant bacteria to humans, yet sometimes veterinarians prescribe antimicrobials to animals without confirming bacterial infection with culture and susceptibility testing. Lack of culture sample submission also limits the ability of diagnostic laboratories to create informative, unbiased antibiograms, also called cumulative susceptibility reports, impairing decision-making for empirical antimicrobial drug prescribing. We hypothesized that the cost of cultures, which are borne by animal owners, is the primary barrier to culture and susceptibility testing in companion animal practice. This study evaluated the impact of a subsidized diagnostic testing program at a veterinary medical teaching hospital on antibiogram reports and testing proportions. The subsidy was implemented in two phases: an initial 16-month period requiring clinicians to complete a form to access free testing, followed by a 2-month period with automatic application of the subsidy without forms. Significant changes in susceptibility (pre vs post subsidy) were observed for several organism-drug combinations, including decreased fluoroquinolone susceptibility in Streptococcus canis and improved beta-lactam susceptibility in Staphylococcus pseudintermedius. Culture submission proportions showed mixed results: canine submissions significantly decreased during the form-required period but showed no change during the automatic subsidy period, while feline submissions showed no significant changes in either period, though there was a non-significant trend toward increased submissions during the automatic subsidy period. This initiative highlights both the complexity of implementing diagnostic subsidies in veterinary practice and the potential for improving the representativeness and utility of diagnostic antimicrobial susceptibility testing (AST) data, which may support broader AMR monitoring efforts when appropriately analyzed.
High pathogenicity avian influenza (HPAI) has caused severe disruptions to poultry production worldwide and in Sweden in the last decade. During 2020 and 2021, 18 outbreaks were confirmed in poultry due to viral introduction from wild birds. The aim of the study was to investigate farm-level risk factors for HPAI in Swedish poultry, to guide prevention measures. A case-control study based on 15 cases from 2020 and 2021 and 33 controls, representing four poultry categories and three counties in southern Sweden was conducted. Data were collected through interviews with farmers/ farmworkers and on-farm observations using a questionnaire and Biocheck.UGent™. Predictor variables were selected based on a hypothesised causal diagram and included in a case-control model using stability selection based on three regularisation models. In addition, a cross-sectional analysis was done based on poultry registry data, and six years of outbreak data. The case-control results showed an association between HPAI outbreaks and presence of waterfowl near the poultry farms. A higher number of birds per farm was positively associated with outbreaks in both parts of the study. Poultry category was associated with HPAI status where turkeys had higher risk and broilers lower risk of outbreaks in the cross-sectional study. The results didn't show a significant association between biosecurity score or other biosecurity covariates, but trends were identified that may deserve further research.
BACKGROUND:Unseat events in steeplechase racing, including falls and lost-rider incidents, affect equine welfare and jockey safety. Timber races are often perceived as higher risk than hurdle/fence races, but are also longer. METHODS:This retrospective observational study analyzed 3070 U.S. National Steeplechase Association hurdle/fence and timber starts from 2023 to 2025. Flat, steeplethon, training, and apprentice/amateur-apprentice starts were excluded. Crude rates of falls, lost-rider incidents, and combined unseat events were compared between race types. Multivariable Poisson regression examined race type, race distance, field size, track condition, and year. Approximate distance at dislodgement was evaluated using Equibase charts. RESULTS:Across 3070 starts, 187 unseat events occurred, including 110 falls and 77 lost-rider incidents. Crude unseat incidence was higher in timber than hurdle/fence races (103.9 vs. 51.8 per 1000 starts; IRR = 2.01, 95% CI 1.47-2.74). Timber races were longer than hurdle/fence races (25.70 vs. 17.79 furlongs; p < 0.001). After adjustment, race type was not independently associated with unseat events (adjusted IRR = 0.98, 95% CI 0.50-1.93), whereas each additional furlong was associated with an 11% increase in unseat incidence (adjusted IRR = 1.11, 95% CI 1.03-1.19; p = 0.005) and each additional starter was associated with a 9% increase (adjusted IRR = 1.09, 95% CI 1.01-1.18; p = 0.025). CONCLUSIONS:Higher crude unseat rates in timber races appear largely attributable to longer race distances rather than race type itself. Race length and field size may be important considerations in jump-racing safety evaluation.
