Bulk tank milk (BTM) provides a practical, non-invasive matrix for repeated herd-level surveillance in dairy farms. This study evaluated a four-season BTM framework for bovine brucellosis, influenza A virus, and Q fever in Korean dairy farms. In 2025, 602 BTM samples were collected from 167 farms in spring and summer and 134 farms in autumn and winter. Seasonal comparisons were interpreted cautiously because farm composition changed after summer. Bovine brucellosis was tested by milk ring test, influenza A virus by M-gene real-time RT-PCR, and Q fever antibodies by ELISA. An exploratory avian influenza virus (AIV) antibody ELISA was also applied to whey. Coxiella burnetii PCR testing was restricted to nine participating farms in Daegu Metropolitan City, and one BTM real-time PCR-positive farm was further investigated using sera from 105 lactating cows. All BTM samples were negative for bovine brucellosis, no influenza A virus M-gene RNA was detected, and no samples exceeded the positivity threshold in the exploratory AIV antibody ELISA. Q fever antibodies were detected in 343 of 602 BTM samples (57.0%). C. burnetii DNA was detected by real-time PCR only in spring and summer BTM samples from one of the nine Daegu farms. In this farm, all serum samples were PCR-negative, whereas 46 of 105 cows were antibody-positive (43.8%). These findings support repeated BTM testing as a complementary herd-level screening approach, particularly for Q fever monitoring.

Spatio-temporal epidemic forecasting aims to predict future outbreak trajectories across interconnected regions from historical epidemiological observations and meteorological covariates. However, existing approaches often fail to preserve historically salient epidemic states or to fully exploit delayed and region-varying meteorological associations, leading to unstable temporal representations and insufficient meteorological-context-aware spatio-temporal context for prediction at later forecast horizons. In this paper, we propose CASRL, a Climate-Aware Self-Retrospective Representation Learning network for stable and meteorological-context-aware spatio-temporal epidemic forecasting. CASRL first employs a Self-Retrospective Epidemic Encoder (SREE) to retrospectively aggregate historically salient epidemic states through query-guided weighting and adaptive gating, thereby preserving informative historical epidemic states within the look-back window. It then introduces a Climate-Adaptive Graph Message Passing (CAGMP) module that breaks away from traditional passive feature concatenation. Instead, it constructs a separate meteorological-view predictive graph conditioned on the static spatial prior and adaptively fuses it with the incidence-associated topology to model complex cross-regional predictive associations. By integrating self-retrospective epidemic representations with meteorological-view spatio-temporal interactions, CASRL produces forecasts with improved predictive stability at later forecast horizons. Extensive experiments on two public influenza benchmarks show that CASRL is competitive at shorter forecast horizons and provides clearer advantages at later forecast horizons, particularly in phase-alignment-related evaluation and 15-week-ahead forecasting.

Post-pandemic influenza dynamics have shown unusual patterns globally. We investigated the re-emergence and recurrent seasonal dominance of influenza B in Korea across three post-pandemic seasons (2022-2026). We analyzed 157 consecutive weeks of national sentinel surveillance data (K-SenS, Korea Disease Control and Prevention Agency), covering September 2022 to August 2025, plus preliminary data through Week 10 of 2026. Influenza B predominance was defined as weeks with > 50% influenza B among total influenza detections; a sensitivity analysis required overall influenza activity > 5%. Influenza B activity remained minimal during the 2022-23 season (mean proportion 3.3%; peak 25.3%) and never reached predominance. Two complete predominance episodes followed influenza A epidemics: after the 2023-24 epidemic (co-circulating A/H1N1pdm09 and A/H3N2; peak ILI 61.3/1000), B predominance began January 2024 and persisted 23 weeks (peak 100%); after the 2024-25A/H1N1pdm09-dominant epidemic (peak ILI 99.8/1000), it began February 2025 and accumulated 20 weeks over two periods with a brief interruption (peak 100%). Onset occurred 5-8 weeks after the influenza A peak by ILI (6-8 weeks virologically) in all three episodes, unaffected by threshold choice. A third episode following the 2025-26A/H3N2 epidemic was ongoing as of March 2026 (onset 7 weeks after the A peak). In post-pandemic Korea, influenza B repeatedly established seasonal dominance 5-8 weeks after influenza A epidemics across three consecutive seasons. This recurring temporal pattern may have implications for influenza surveillance, clinical preparedness, and vaccine strategy, though its underlying mechanism remains unestablished.

We examined inequalities in weekly hospitalization rates related to COVID-19, influenza, and respiratory syncytial virus based on neighbourhood-level material deprivation, neighbourhood-level ethno-racial diversity, rurality, and sex. Hospitalization rates were measured as the weekly number of hospitalizations per 10 000 people. Local regression models were used to fit observed weekly hospitalization rates within each study year and each population subgroup, and statistical significance of between-subgroup differences in peak hospitalization rate was assessed based on the overlap of 95% confidence intervals. We also provided comparisons of pre-pandemic (July 2017-June 2020) versus post-pandemic (July 2020-June 2023) trends. More materially deprived neighbourhoods showed up to a 2.5-times increase in peak weekly hospitalization rates compared to less materially deprived neighbourhoods. Peak hospitalization rates were up to 2.3-times higher among more versus less diverse neighbourhoods. Some age groups showed peak hospitalization rates up to 2.8-times higher among urban residents. Peak hospitalization rates were up to 1.3-times higher in younger males versus younger females and up to 1.6-times higher in older females versus older males. Variations due to neighbourhood-level material deprivation, neighbourhood-level ethno-racial diversity, rurality, and sex should be considered when designing and implementing public health policies aiming to reduce respiratory virus burden.

Compared to developed countries, influenza vaccines have lower popularity in developing countries and are generally not given much attention. This paper analyzes the effect of China's policy of providing free influenza vaccine for the population aged 60 and above on vaccination take-up and health outcomes.

Human seasonal influenza evolves rapidly, necessitating twice yearly decisions about whether to update the strains in the vaccine. To help inform this decision, we have been using high-throughput sequencing-based neutralization assays to make twice yearly measurements of how recent human sera neutralize current human H3N2 and H1N1 strains. Here we provide the third installment in this series of measurements by reporting 52,268 titers representing neutralization of 148 viral strains by 355 human sera collected between April and August of 2026. Our measurements show that new H3N2 subclade K strains with mutations in antigenic region D and new H1N1 subclade D.3.1.1 strains with mutations in antigenic region Sa (such as G155E) have reduced neutralization by human sera, with notable heterogeneity in the impact of some of these mutations across sera from different individuals. This paper is accompanied by an interactive summary (https://jbloomlab.github.io/flu-seqneut-2026/summary.html) that enables detailed exploration of the results, and all titer data are publicly available for further analysis to aid vaccine antigen selection and studies of viral evolution.