Using a probabilistic partial-budget model built from literature-derived epidemiological parameters and Korean livestock-industry data, researchers estimated that a 50-head Hanwoo beef farm facing BVDV-related production impacts would lose a mean of KRW 12.906 million per year, with a 90% uncertainty interval running from KRW 8.782 million to KRW 18.056 million. Expressed per animal, that amounts to KRW 258,113 per head, with an interval of KRW 175,646 to KRW 361,121. For a 100-head dairy farm with 80 lactating cows, the corresponding mean annual loss was KRW 11.102 million, spanning KRW 7.126 million to KRW 16.211 million, or KRW 111,023 per head between KRW 71,261 and KRW 162,119.
The study, published in Frontiers in Veterinary Science, was designed explicitly as a modeling exercise rather than a measurement of actual losses on Korean farms. Its authors stress that the outputs are modeled scenarios grounded in literature-informed assumptions, providing a quantitative baseline rather than direct causal estimates of what BVDV costs the Korean cattle industry. That distinction matters because Korea lacks the long-term farm-level data comparing BVDV-affected and unaffected herds that would allow a definitive national loss figure. Instead, the model assembles the best available parameter estimates, propagates their uncertainty through simulation, and reports a distribution of plausible outcomes.
BVDV is a globally distributed cattle pathogen with a wide clinical reach. Infection can cause enteric and respiratory disease, reproductive disorders, fetal damage, immunosuppression, and reduced productivity. At the farm level, the virus may erode income through reduced milk yield, impaired reproductive performance, calf mortality, and increased treatment and control costs. A particularly important feature of BVDV biology is that infection during early pregnancy can produce persistently infected animals, which shed virus throughout their lives and serve as a major reservoir for maintenance and transmission within herds. Early detection and removal of these persistently infected cattle, along with prevention of infection in pregnant animals, are central principles of control programs worldwide.
The global evidence base shows how heterogeneous BVDV’s occurrence and economic effects are. A systematic review and meta-analysis of studies published between 2010 and 2021 estimated a pooled antigen prevalence of 15.74% and a pooled antibody prevalence of 42.77%, with wide variation by region, breed, diagnostic method, sample type, and production system. Reported annual direct monetary losses in a worldwide review ranged from USD 0.50 to 687.80 per animal. Several countries have responded with control programs combining identification and removal of persistently infected animals, routine testing, movement restrictions, farmer education, vaccination, and biosecurity. In Germany, after a nationwide compulsory program was implemented, the proportion of farms with persistently infected cattle fell from 3.44% in 2011 to 0.16% in 2016, while individual-level prevalence declined from 0.48% to 0.02%.
BVDV is well documented in Korea, but the domestic studies report different epidemiological measures that cannot be read interchangeably. In one study of 319 samples from calves with clinical diarrhea in Gyeongnam Province, BVDV was detected in 12.9% of samples. Another study reported an antibody-positive rate of 72.2% among 4,260 Hanwoo cattle and identified 27 persistently infected animals, 40.7% of which were calves younger than six months. A review of 81 Korean studies on infectious-disease agents in cattle published from 2010 to 2019 found a BVDV detection rate of 14.8% in calves and growing cattle, 10.3 percentage points higher than the 4.5% rate in adult cattle. Virus detection, antibody positivity, and persistently infected animal identification each capture a distinct aspect of the virus’s epidemiology, and the model’s authors deliberately treated prevalence as a scenario input rather than a national estimate.
The model’s architecture reflects careful attention to avoiding double counting. Each loss component was applied to an event-specific population at risk rather than to the entire herd. Abortion and postnatal calf mortality arising from the same pregnancy were treated as mutually exclusive events, so a single fetal or calf death is never counted twice. The abortion and calf-mortality parameters represent conditional excess probabilities attributable to the modeled BVDV effect among infected animals, not absolute probabilities from all causes, and background abortion and mortality rates were therefore not added again. Reproductive losses were applied to pregnant breeding cows, calf mortality only to live-born calves from pregnancies that did not end in abortion, milk-yield losses only to infected dairy cows that survived and remained in milk, and growth-delay losses to infected growing cattle based on reduced average daily gain.
Uncertainty was handled through probability distributions and 10,000 Monte Carlo iterations per production system. Triangular distributions, whose minimum, most-likely, and maximum values the authors describe as literature-informed scenario anchors rather than empirical quantiles, were assigned to the BVDV infection proportion (5%, 10%, and 15%), excess abortion probability (2%, 5%, and 10%), excess calf-mortality probability (2%, 5%, and 10%), average daily gain reduction (0.2, 0.35, and 0.5 kg/day), body-weight loss rate (2%, 5%, and 10%), milk-yield reduction (5%, 10%, and 20%), duration of milk-yield loss (5, 10, and 20 days), and premature culling rate (0.5%, 1.0%, and 2.0%). Calf value was modeled with lognormal distributions reflecting right-skewed market prices: a mean of KRW 2.5 million per head for Hanwoo calves and KRW 0.8 million for dairy calves. Fixed inputs included replacement-animal purchase price, cull salvage value, carcass price, feed cost, baseline growth rates, daily milk yield, and raw milk price drawn from Korean official statistics.
The simulation results reveal where the economic burden concentrates. In the Hanwoo model, abortion accounted for 44.3% of the modeled loss and calf mortality for 38.8%, together 83.1% of the total, followed by body-weight loss at 10.1% and premature culling at 6.8%. The dairy model was even more dominated by reproductive and calf losses: abortion contributed 47.9% and calf mortality 47.6%, roughly 95.5% combined, with milk-yield loss at 3.3% and premature culling at 1.2%. Sensitivity analysis using Spearman rank correlations showed that calf value had the strongest association with simulated total loss in both systems, at 0.662 for Hanwoo and 0.640 for dairy, followed by excess abortion probability at 0.598 and 0.518 respectively, and excess calf-mortality probability at 0.348 and 0.503. The authors note that rank correlation measures monotonic association rather than decomposing variance, so the coefficients should not be read as causal effects or proportions of variance explained.
The authors are candid about the model’s limitations. Stochastic inputs were sampled independently, although in biological settings infection prevalence, reproductive failure, calf mortality, and production effects may be correlated, meaning the reported uncertainty intervals may understate or overstate joint uncertainty, particularly in the tails of the distribution. The same sampled prevalence was applied to breeding cows and calves because production-group-specific estimates were unavailable, and the losses are conditional outcomes of hypothetical infection scenarios ranging from 5% to 15%, not figures reflecting actual national prevalence. Some Korean unit prices and management costs were treated as fixed values, and the static annual partial-budget structure did not capture within-herd transmission, repeated infection, temporal changes in prevalence, replacement dynamics, market feedback, or industry-level price effects. The uncertainty intervals reflect only the variables and distributions included in the model.
Despite these constraints, the study offers Korean policymakers and veterinarians something previously missing: an integrated, uncertainty-aware estimate of what BVDV-related production impacts could cost representative domestic farms. The authors position the model as a baseline for future evaluation of surveillance, early detection and removal of persistently infected cattle, vaccination, biosecurity, and movement-control strategies, while noting the work was not designed to assess the cost-benefit of any specific control program. They call for longitudinal Korean farm data on infection, reproductive outcomes, calf survival, milk yield, growth, culling, and control expenditures, integrated with dynamic transmission models, so that intervention effects and the cost-effectiveness of BVDV control in Korea can eventually be quantified explicitly over time.
Subject of Research: Scenario-based probabilistic estimation of farm-level economic losses from bovine viral diarrhea virus in Korean Hanwoo beef and dairy cattle
Article Title: Scenario-based estimation of economic losses associated with bovine viral diarrhea virus infection in Hanwoo and dairy cattle in Korea
Article References: Moon, S.-H., Choi, K.-S., Jeong, C.-G., & Oem, J.-K. (2026). Scenario-based estimation of economic losses associated with bovine viral diarrhea virus infection in Hanwoo and dairy cattle in Korea. Frontiers in Veterinary Science, 13, Article 1973189. https://doi.org/10.3389/fvets.2026.1973189
Image Credits: AI Generated
DOI: 10.3389/fvets.2026.1973189
Keywords: bovine viral diarrhea virus, BVDV, Hanwoo, dairy cattle, economic loss, Monte Carlo simulation, partial budget model, Korea, abortion, calf mortality, sensitivity analysis, veterinary economics
