Simulation and prediction of human brucellosis epidemic trends in Changji city Authors: Pengwei Lou, Yuting Huang, Xiaolin Fan, Nannan Chen, Yifan Hu, Honggang Liu, Feng Zhao, Jiabo Xu, Enguang Gao, Kai Wang Published: 2026 ABSTRACT AbstractThe number of new human brucellosis cases has shown a periodic upward trend in Changji city, Xinjiang. It is essential to estimate the transmission intensity of brucellosis and put forward effective prevention and control measures. According to the mechanistic characteristics of brucellosis transmission, we established a seasonal dynamic model to describe the transmission of brucellosis among sheep/cattle, environment and humans. The constructed seasonal dynamic model SEIVW with periodic transmission rates has perfectly simulated the periodic upward trend of new human brucellosis cases in Changji city from 2020 to 2023. The mean absolute percentage error (MAPE) between the model-predicted and actual values was 33.20% for new cases and 10.56% for cumulative new cases. Long-term prediction over the next 25 years suggested that the number of new human brucellosis cases will first show an upward trend, reaching a peak of 3688 cases in July 2032, followed by a gradual decline and stabilization after 2038. Furthermore, we estimated the basic reproduction number R0 for brucellosis transmission is 2.8228 (95% CI: 2.8161-2.8295), indicating that the disease transmission remains severe and persistent. Parameter sensitivity analysis quantitatively identified key intervention strategies: increasing the culling rate of infected sheep/cattle, and improving the vaccination rate of susceptible sheep/cattle are the primary control measures. Additionally, enhancing environmental disinfection frequency can be served as a supplementary measures.Keywords: Brucellosis, Dynamic model, Epidemic trends, Basic reproduction number, Control measures Introduction Brucellosis is a chronic infectious-allergic disease caused by brucella species, which invade host organisms and cause zoonotic infections (Dean et al., 2012). The World Health Organization (WHO) classifies brucella into six species: B. melitensis, B. abortus, B. suis, B. neotomae, B. ovis, and B. canis. Among these, B. melitensis, B. abortus, and B. suis are the species that most commonly infect humans, with B. melitensis exhibiting the highest pathogenicity (Dadar et al., 2021; Hull & Schumaker, 2018). Recognized as one of the most widespread and harmful zoonoses globally, brucellosis has been reported in over 170 countries and regions, with approximately 500,000 new human cases annually (Li et al., 2020). Since the first two human cases were reported in Chongqing, China in 1905, outbreaks of brucellosis have recurred without effective control (Deb et al., 2023; Ran et al., 2019). From 2020 to 2023, China reported 47,245, 69,767, 71,437, and 70,439 new human brucellosis cases respectively, accounting for approximately 12.94% of global total number of cases. Brucellosis is classified as a Class B notifiable infectious disease in accordance with the Law of the People's Republic of China on the Prevention and Treatment of Infectious Diseases (Sun et al., 2020). The primary sources of brucellosis infection is infected animals, including approximately 60 species such as cattle, sheep, pigs, horses, dogs, and wildlife (Almuzaini, 2023), while human-to-human transmission is extremely rare, with only sporadic documented cases (Mesner et al., 2007). Infected animals shed the brucella into the environment through excretions and secretions, particularly during parturition, when large quantities of the bacteria are released via fetuses, placental tissues, and other birth materials. Abortions also result in further shedding of brucella via vaginal secretions and milk. Additionally, other infectious secretions include saliva, urine, and feces (Çiftci et al., 2019). This causes to environmental contamination and increases the risk of infection for susceptible animals and humans. Brucella is transmitted through multiple routes, primarily via respiratory inhalation of aerosolized bacteria in contaminated environments. Other routes include gastrointestinal and reproductive system exposure (Lyamkin et al., 2016). In addition, direct contact with infected animals or their contaminated secretions and excretions also poses significant infection risks (Yuan et al., 2020). Human brucellosis infection demonstrates distinct occupational characteristics, showing higher prevalence among people who have frequent contact with livestock. High-risk occupations include veterinarians, slaughterhouse workers, livestock farmers, hide processors, and breeding technicians, whose infection risks is significantly higher than that of people in other occupations (Escobar et al., 2013). Cases occur year-round but exhibit obvious seasonal patterns, the incidence peaks during June and July, and falls to its lowest level from October to December (Nyerere et al., 2020). The incubation period for brucellosis in livestock varies from one week to six months. Infections in the early stage are predominantly subclinical, and affected animals exhibit similar clinical signs, primarily reproductive disorders. In the early stage of infection, affected animals develop elevated body temperature and conjunctivitis. Pregnant females show decreased appetite, polydipsia, restlessness, mammary gland induration, and discharge of foul-smelling brownish-yellow viscous fluid from the vaginal during gestation. In severe cases, this can lead to abortion, weak offspring or stillbirth (Aloufi et al., 2016). Infected males typically present with anorexia, general debilitation, and orchitis, some cases may develop arthralgia, joint swelling or bursitis, which can lead to lameness and impaired mobility (Byndloss et al., 2010). Human brucellosis manifests after an incubation period of approximately 2 weeks, with main symptoms including fever, hyperhidrosis, fatigue, arthralgia, anorexia, and anemia. The clinical course is often protracted and recurrent, which can cause severe damage to the urinary system, reproductive system and multiple organs (Battikh et al., 2021). Some patients may develop complications such as encephalitis, myelitis, endocarditis, orchitis, cough, or cutaneous manifestations (Hasanjani Roushan et al., 2016). If the disease is not treated during the critical acute phase, it can progress to the chronic phase, causing long-term morbidity; meanwhile, its nonspecific initial symptoms often lead to misdiagnosis as influenza or rheumatic fever (Buzgan et al., 2010). As one of China's major livestock production bases, Xinjiang Uygur Autonomous Region has experienced severe brucellosis transmission, with the number of reported human cases being 3,167, 4,780, 6,737, and 9992 during 2020-2023, respectively. As a prefecture-level city under the jurisdiction of Xinjiang, Changji city maintained an average cattle and sheep inventory of 349,420 during the four-year period, with 506 new human brucellosis cases reported, indicating sustained endemic transmission. Given the current epidemiological trajectory in Changji city, existing prevention and control measures have effectively interrupted disease transmission, and there is a risk of outbreaks. Therefore, there is an urgent need to establish predictive models for brucellosis fluctuation patterns and conduct quantitative evaluations of intervention strategies. Infectious disease dynamics is an effective methodology for elucidating the transmission patterns of infectious diseases. By establishing epidemiological models combined with qualitative and quantitative analyses and numerical simulations, this approach provides a crucial theoretical foundation for better understanding intrinsic transmission mechanisms, predicting development trends, analyzing epidemic determinants, and evaluating intervention strategies (Aïnseba et al., 2010). To investigate the transmission characteristics and epidemic patterns of brucellosis and identify effective prevention and control measures, researchers have established deterministic models for brucellosis transmission and achieved favorable research results (Hou & Qin, 2019; Hou & Wang, 2017; Lolika & Mushayabasa, 2018). Hou et al. (Hou et al., 2013) constructed a human-sheep coupled brucellosis model for the Inner Mongolia Autonomous Region, incorporating sheep populations, human populations, and environmental brucella contamination, demonstrating that combined vaccination and environmental disinfection could effectively control brucellosis in this region. Zinsstag et al. (Zinsstag et al., 2005) established a tripartite human-sheep-cattle dynamic model based on surveillance data from Mongolia, predicting local brucellosis trends and proposing risk management measures for disease control. Li et al. (Li et al., 2014) developed a human-sheep coupled brucellosis dynamic model analyzing epidemic trends in China's high-incidence provinces including Hinggan League of Inner Mongolia, with findings indicating that vaccination, testing, and culling constitute effective measures for controlling human transmission. Ma et al. (Ma, Sun, et al., 2022) formulated a discrete-time dynamic model coupling humans and sheep to study the impact of disinfection and health education on the spread of brucellosis. While previous studies have comprehensively addressed the transmission patterns of brucellosis among sheep, cattle, and humans, they only used annual aggregated data and failed to incorporate seasonal transmission dynamics, thus making it difficult to fully reveal the cyclical epidemic characteristics of the disease. In recent years, infectious disease dynamics has been applied to analyze seasonally recurring infectious diseases, including rabies, hand, foot and mouth disease (HFMD), schistosomiasis, and tuberculosis. It enables ideal simulation and prediction of seasonal patterns and supports the development of effective control measures (Li, Teng, et al., 2017; Liu, 2011; Liu et al., 2010; Zhang et al., 2012). Brucellosis has been confirmed to exhibit a distinct seasonal incidence pattern, which satisfies the data characteristics required for the application of seasonal dynamic models. Lou et al. (Lou et al., 2016) developed a brucellosis dynamic model incorporating periodic transmission rates to investigate seasonal transmission dynamics in the Bayingolin Mongol Autonomous Prefecture of Xinjiang. However, this study has a limitation in that the model did not consider the impact of brucella surviving in the environment on disease transmission, which may compromise the accuracy of assessing the transmission intensity of brucellosis. The primary objective of this study is to establish a seasonal dynamic model by comprehensively analyzing the transmission patterns among sheep/cattle, the environment, and humans, incorporating the characteristic mechanisms of brucellosis transmission and introducing a periodic transmission rate. The model was fitted to the epidemic trends of new human brucellosis cases in Changji city and used for long-term prediction. Transmission intensity was assessed by estimating the basic reproduction number R0. Meanwhile, sensitivity analysis was performed on model parameters related to new human brucellosis cases and R0, so as to quantitatively propose effective prevention and control measures. Materials and methods Data source The study data were obtained from monthly reports of new human brucellosis cases between 2020 and 2023 provided by the Changji City Center for Disease Control and Prevention, comprising 506 cases, with annual case numbers of 63, 108, 164, and 171 respectively, demonstrating a distinct yearly cyclical increasing trend as shown in Fig. 1. Monthly number of new human brucellosis cases in Changji city, 2020-2023. Seasonal index The seasonal index was employed to quantify periodic fluctuations in time-series data across annual cycles, calculated as Sj=x¯j/x¯, where x¯j represents the monthly average incidence of new human brucellosis cases and x¯ denotes the overall mean incidence (Lou et al., 2016). As demonstrated in Fig. 2, the seasonal index for new human brucellosis cases in Changji city reached 2.0398 in July, indicating the occurrence of a transmission peak, the value remained consistently exceeding 1 from April to August, confirming elevated case incidence during summer and autumn, and validating the characteristic seasonal epidemic pattern of brucellosis. Seasonal index of new human brucellosis cases. Brucellosis transmission model construction Based on mechanistic analysis of brucellosis transmission among humans, sheep/cattle, and the environment, we developed a seasonal dynamic model of brucellosis transmission as shown in Fig. 3. The sheep/cattle population was divided into four subpopulations: susceptible S(t), exposed E(t), infected I(t), and vaccinated V(t); humans were categorized into three subpopulations: susceptible Sh(t), acutely infected Iha(t), and chronically infected Ihc(t); while environmental brucella contamination was modeled as a separate compartment W(t). The primary transmission processes in the model include: both susceptible S(t) and vaccinated V(t) sheep/cattle losing immunity may become exposed E(t) upon contact with brucella, subsequently progressing to infected status I(t), with both exposed E(t) and infected I(t) sheep/cattle capable of shedding brucella into the environment. Susceptible humans Sh(t) acquire acute infection Iha(t) through contact with exposed E(t) or infected I(t) animals or environmental brucella, potentially progressing to chronic infection Ihc(t) if untreated. Key model assumptions include (1): Human-to-human transmission is negligible based on rare documented cases (2); Infected individuals progress directly to acute infection, which is consistent with the approximately two-week incubation period and frequent diagnostic delays. Transmission flowchart of brucellosis. Based on the transmission patterns of brucellosis among humans, sheep/cattle, and the environment, the following system of ordinary differential equation (1) can be formulated: In model (1), only the first five equations for the transmission of brucellosis by sheep/cattle need to be considered, as shown in model (2). It is clear that model (2) has a unique positive disease-free equilibrium P0=(S∗,0,0,V∗,0), where Consider the following auxiliary equations: In fact, the Jacobian matrix of model (4) at equilibrium (A (μ+δ)/μ(μ+v+δ), Av/μ(μ+v+δ)) is and then the corresponding characteristic equation is By simple calculation, it is easy to obtain that the two roots of Φ(λ) are λ1=-μ andλ2=-(μ+ν+δ). Hence, we obtain that the unique positive equilibrium (A (μ+δ)/μ(μ+v+δ), Av/μ(μ+v+δ)) is locally asymptotically stable. In addition, since model (4) is linear, by the theorems of stability of the differential equations, we obtain that the equilibrium (A (μ+δ)/μ(μ+v+δ), Av/μ(μ+v+δ)) is globally asymptotically stable, which completes the proof. Now, we compute the basic reproduction number R0 of model (2) by applying the way given in (Wang & Zhao, 2008; van den Driessche & Watmough, 2002) by Wang et al. Let where x=(E,I,W,S,V)T, and then model (2) takes the following form Obviously, model (8) has a disease-free equilibrium x∗(t) = (0, 0, 0, S∗, V∗). Next, we set two 2 × 2 matrices as follows: where fi(t,x(t)) and vi(t,x(t)) are the ith component of fi(t,x(t)) and vi(t,x(t)), respectively. Then, by simple computation, it follows that Hence, we easily check that conditions (A1) - (A7) given in (Wang & Zhao, 2008) are satisfied. Let Y (t, s) be the 2 × 2 matrix solution of the following initial value problem: Let Cω be the ordered Banach space of all ω-periodic continuous function form R to R2 with the maximum norm ‖·‖. The positive cone Cω+={ϕ∈Cω:ϕ(t)≥0forallt∈R}. Suppose ϕ(s)∈Cω is the initial distribution of infectious individuals in this periodic environment, then F(s)ϕ(s) is the rate of new infectious individuals produced by the infected individuals who were introduced at time s, and Y(t,s)F(s)ϕ(s) represents the distributions of those infected individuals who were newly infected at time s and remain in the infected compartment at time t for t ≥ s. Hence, we define a linear operator L:Cω→Cω as follows: The operator L is positive, continuous, and compact on Cω. Thus, R0 can be characterized by the existence of a nonnegative and nonzero ϕ∈Cω+ such that Now, we define basic reproduction number R0 for model (8) by where ρ(L) is the spectral radius of L. Model parameters and values The parameters in model (Dean et al., 2012), along with their definitions and values, are summarized in Table 1, with most parameter values obtained from published literature. The periodic transmission rates in this model are represented by sin functions (Yang et al., 2013). In this study, β(t)=α[1+bsin(ωt+c)] represents the transmission rate among sheep/cattle; βe(t)=αe[1+besin(ωt+ce)] denotes the transmission rate from environmental brucella to susceptible sheep/cattle; βh(t)=αh[1+bhsin(ωt+ch)] indicates the transmission rate from exposed/infected sheep/cattle to humans; and βhe(t)=αhe[1+bhesin(ωt+che)] signifies the transmission rate from environmental brucella to humans. To reflect the annual seasonal characteristics of brucellosis transmission, we set the transmission cycle to one year. Mathematically, the period T of a sinusoidal function is related to its angular frequency ω by T = 2π/ω. With monthly time units, setting ω = π/6 yields a period T = 12, corresponding exactly to 12 months. This parameter ensures that the transmission rates β(t), βe(t), βh(t), and βhe(t) exhibit a complete annual cycle. For the twelve parameters a, b, c, ae, be, ce, ah, bh, ch, ahe, bhe and che included in the periodic transmission rates β(t), βe(t), βh(t), and βhe(t), reasonable estimates were derived through model fitting to the 2020-2023 new human brucellosis cases in Changji city using Markov Chain Monte Carlo (MCMC) method. Since the Poisson distribution is generally used to describe sparsely occurring events, it is assumed that new human brucellosis cases follow a Poisson distribution. The parameter vector is θ|z={α,b,c,αe,be,ce,αh,bh,ch,αhe,bhe,che}|z, where vector z represents the number of new human brucellosis cases, and Z(t) = Y(t) - Y(t-1), ddtY(t)=αh(1+bhsin(ωt+ch))Sh(E+I)+αhe(1+bhesin(ωt+che))ShW. We employed the Metropolis-Hastings algorithm within the MCMC methodology for parameter estimation. Assume that the prior distributions of the parameters a, ae, ah and ahe follow a uniform distribution U (0, 1), and the prior distributions of all other parameters in the periodic transmission rate follow the uniform distribution U (-30, 30). The constructed posterior distribution is fΘ|zθ|z=ΠTLYt|θfΘθ. Code implementation was performed using the deBInfer package in R software (Version 4.5.1). The estimated values of the posterior distributions these parameters are presented in Table 1, and their probability density curves are illustrated in Fig. 4. Accordingly, the full expressions for the periodic transmission rates are β(t)=3.5028×10−7[1+7.8239sin(π6t−0.4270)], βe(t)=8.8660×10−9[1+15.6067sin(π6t−1.2826)], βh(t)=6.3677×10−7[1+0.9667sin(π6t−7.9381)], and βhe(t)=1.6095×10−8[1+12.2079sin(π6t−24.1251)], respectively. Table 1Definitions and values of model parameters.Table 1ParameterUnitValueCommentsRefAMonth−18470Monthly new recruitment of sheep/cattleaμMonth−10.0183The natural mortality rate of sheep/cattleajMonth−10.0313The culling rate of infected sheep/cattle[ (Lou et al., 2016)]δMonth−10.0333The immune loss rate of sheep/cattle[ (Liu et al., 2025)]νMonth−10.0209The vaccination rate of susceptible sheep/cattle[ (Lou et al., 2016)]εMonth−10.0150The ineffective vaccination rate[ (Hou et al., 2013)]λMonth−12The transition rate from exposed to infected sheep/cattle[ (Lou et al., 2016)]a-3.5028 × 10−7 (95% CI: 3.4971 × 10−7, 3.5086 × 10−7)Baseline contact ratesFittingb-7.8239 (95% CI: 7.8105, 7.8374)AmplitudesFittingc-−0.4270 (95%CI: −0.4273, −0.4260)PhasesFittingae-8.8660 × 10−9 (95% CI: 8.8521 × 10−9, 8.8799 × 10−9)Baseline contact ratesFittingbe-15.6067 (95% CI: 15.5795, 15.6338)AmplitudesFittingce-−1.2826 (95% CI: −1.2848, −1.2803)PhasesFittingBMonth−1382Monthly new recruitment of humanbμ1Month−10.0005The natural mortality rate of humansbηMonth−10.0500The cure rate from acute infected individuals to susceptible individuals[ (Qin et al., 2022)]ωhMonth−10.1667The removal rate from acute infected individuals[ (Hou et al., 2013)]kMonth−11.2500brucella shedding rate by exposed or infected Sheep/Cattle[ (Li et al., 2014)]dMonth−10.2981Decay rate of brucella in the environment[ (Hou et al., 2013)]nMonth−10The frequency of disinfection[ (Li, Sun, et al., 2017)]τMonth−10.9500Effective disinfection rate[ (Ni et al., 2024)]ah-6.3677 × 10−7 (95% CI: 6.3568 × 10−7, 6.3785 × 10−7)Baseline contact ratesFittingbh-0.9667 (95%CI: 0.9652, 0.9682)AmplitudesFittingch-−7.9381 (95%CI: −7.9511, −7.9251)PhasesFittingahe-1.6095 × 10−8 (95%CI: 1.6069 × 10−8, 1.6121 × 10−8)Baseline contact ratesFittingbhe-12.2079 (95%CI: 12.1863, 12.2294)AmplitudesFittingche-−24.1251 (95%CI: −24.1646, −24.0856)PhasesFittingaThe Xinjiang Livestock Yearbook provides records of the stock numbers and natural mortality rates for sheep and cattle in Changji city from 2020 to 2023, enabling the derivation of monthly parameter values for A and μ.bThe Xinjiang Statistical Yearbook documents the total population, birth rate, and mortality rate in Changji city during 2020-2023, from which the monthly values of parameters B and μ1 can be derived. Definitions and values of model parameters. The Xinjiang Livestock Yearbook provides records of the stock numbers and natural mortality rates for sheep and cattle in Changji city from 2020 to 2023, enabling the derivation of monthly parameter values for A and μ. The Xinjiang Statistical Yearbook documents the total population, birth rate, and mortality rate in Changji city during 2020-2023, from which the monthly values of parameters B and μ1 can be derived. Probability density curves of the twelve parameters in periodic transmission rates. (A)-(C) represent parameters a, b, and c in the periodic transmission rate β(t), respectively. (D)-(F) represent parameters ae, be, and ce in βe(t), respectively. (G)-(I) represent parameters ah, bh, and ch in βh(t), respectively. (J)-(L) represent parameters ahe, bhe, and che in βhe(t), respectively. Initial values need to be set for perform the numerical simulations of model (Dean et al., 2012). According to the Xinjiang Statistical Yearbook, the number of sheep/cattle in stock at the end of 2020 was N (0) = 375,000. The monthly vaccination rate for sheep/cattle in Changji city is 0.0209, with an efficacy of 0.0683; therefore, V (0) = 375,000 × 0.0209 × 0.0683 = 535. The incidence rate of brucellosis in sheep/cattle is approximately 0.0019, so E (0) +I (0) = 375,000 × 0.0019 = 712, where E (0) = 237 and I (0) = 475. S (0) = N (0) - V (0) - E (0) - I (0) = 373,753. The Xinjiang Statistical Yearbook records the population at the end of 2020 as Sh(0) = 410,000. The initial value Iha (0) = 4 is obtained from the Changji City Center for Disease Control and Prevention. The initial value Ihc (0) is derived from Iha (0) using the parameter ωh (1-η), which is approximated to 1. We use data fitting to obtain W (0) = 500. Results Model fitting The constructed seasonal dynamic model was applied to fit both the monthly new and cumulative new human brucellosis cases in Changji city from 2020 to 2023, as shown in Fig. 5A and B, respectively. The model effectively simulated the epidemic trend of human brucellosis in Changji city, demonstrating good agreement between the fitted and observed values, with distinct periodic fluctuations. The mean absolute percentage error (MAPE), calculated as follows, was used to evaluate the model's accuracy: The seasonal dynamic model fitting results for the monthly new and cumulative new human brucellosis cases in Changji city, 2020-2023. (A) Monthly new human brucellosis cases. (B) Monthly cumulative new human brucellosis cases. Here, Wt denotes the actual number of new human brucellosis cases, Qt represents the model-estimated number of new human brucellosis cases, and n indicates the number of observations (Zhang et al., 2015). A smaller MAPE value corresponds to better agreement between simulated and observed values, indicating higher model fitting accuracy. The MAPE values for fitting both the monthly new and cumulative new human brucellosis cases in Changji city were 33.20% and 10.56%, respectively, demonstrating that the constructed seasonal dynamic model can reasonably simulate the fluctuating trends of human brucellosis in Changji city. Meanwhile, we also constructed a Seasonal Autoregressive Integrated Moving Average (SARIMA) time series model to simulate the epidemic trend of new human brucellosis cases in Changji city. The estimated MAPE value was 34.38%, which is slightly higher than the error value of the seasonal dynamic model constructed in this study, detailed results are provided in the Supplementary Material. This also verifies that the seasonal dynamic model constructed in this study exhibits favorable predictive performance. Model prediction Based on the satisfactory model performance, we predicted the monthly new and cumulative new human brucellosis cases in Changji city from 2024 to 2050, as shown in Fig. 6A and B, respectively. Under current prevention and control measures, the monthly number of new human brucellosis cases in Changji city is predicted to exhibit a periodic upward trend from 2024 to 2032, reaching a peak of 3688 cases in July 2032, followed by a gradual decline and stabilization after 2038. The model-predicted monthly new human brucellosis cases from 2024 to 2049 are detailed in Table 2. Predictions of monthly new and cumulative new human brucellosis cases in Changji city from 2024 to 2049. (A) Monthly new human brucellosis cases. (B) Cumulative new human brucellosis cases. Dark yellow denotes the predicted values, light yellow denotes the 95% confidence interval, and dark red denotes the observed values. Table 2Predicted monthly new human brucellosis cases from 2024 to 2038.Table 2YearsMonths12345678910111220241413121629495841168132020252119182446831027228122132202635312838791491851324920365620276153486614127234524589356510020281089483116256504642455163611171812029195169148208465921116582029210920732220303452982603668211604198313784951863495402031575496433608134525272983203174929252680620328577386459051922335636882442942394668100820331067919812112822353519353922779344306811010203410679258291133204828702652167373537956982920358807727069431550196417101075506290425616203666259255472411031307110670335122132347020375134704495788389598065202711822693942038436407395502708797671438235164244359 Predicted monthly new human brucellosis cases from 2024 to 2038. Concurrently, we have separately predicted the trends in the numbers of exposed, infected, and vaccinated sheep/cattle. As shown in Fig. 7, the number of exposed and infected sheep/cattle exhibited a rapid increase between 2020 and 2032, followed by a gradual decline, eventually stabilising into a cyclical pattern. The model's prediction regarding the number of exposed and infected sheep/cattle align perfectly with the trends of new human brucellosis cases, indicating that human infection mainly results from contact with exposed or infected sheep/cattle. The number of vaccinated sheep/cattle gradually increased initially before stabilising, which remains markedly smaller than the number of the exposed and infected sheep/cattle, indicating the transmission situation of brucellosis among sheep/cattle is rather severe. The model predicted the trends in the number of exposed, infected, and vaccinated sheep and cattle, respectively. Purple denotes exposed sheep/cattle, blue denotes infected sheep/cattle, and green denotes vaccinated sheep/cattle. Basic reproduction number R0 Using the methodology described by Wang et al. (Wang & Zhao, 2008), we calculated the basic reproduction number R0 for brucellosis transmission in Changji city as 2.8228 (95% CI: 2.8161-2.8295), with R0 = 2.3668 attributed to transmission through contact with infected sheep/cattle and R0 = 0.4560 resulting from exposure to viable brucella in the environment. When model (Dadar et al., 2021) degenerates into an autonomous case with β¯=(1/12)∫012β(t)dt, β¯e=(1/12)∫012βe(t)dt we obtain The formula for calculating the mean basic reproduction number R˜0 can be obtained using the method established by van den Driessche and Watmough (van den Driessche & Watmough, 2002): The estimated mean basic reproduction number was R˜0 = 2.8282 (95% CI: 2.8215-2.8349), slightly higher than the periodically varying R0, indicating that the mean-based calculation method may lead to a modest overestimation of brucellosis transmission intensity. The boxplot comparing the basic reproduction number R0 and mean basic reproduction number R˜0 is shown in Fig. 8. With R0 > 1, these results demonstrate that brucellosis transmission in Changji city has not been effectively controlled, cannot yet be eliminated at present, and is likely to persist as an endemic disease for a long time. Boxplot of the basic reproduction number R0 and mean basic reproduction number R˜0. Parameter screening We present Partial Rank Correlation Coefficients (PRCC) values based on Latin Hypercube Sampling (LHS) with a sample size of N = 1000. Parameters from model (Dean et al., 2012) were selected as input variables, with new human brucellosis cases serving as the output variable. Assuming all parameters follow a uniform distribution and a significance level of 0.05, a P-value below 0.05 indicates the selected parameter is statistically significant. The PRCC values and P-values for model parameters of new human brucellosis cases are presented in Table 3 and Fig. 9. Asterisks denote statistically significant parameters. A larger absolute PRCC value indicates a stronger influence on the model, with positive and negative values representing positive and negative effects respectively. Parameters j, A, ν, n and δ exert significant influence on new human brucellosis cases. Parameters j, ν and n exhibit negative correlations, while A and δ demonstrate positive correlations. Consequently, controlling the values of these influential parameters in the model within reasonable ranges may reduce the new human brucellosis cases. Table 3Partial rank correlation coefficients between new human brucellosis cases and each input parameter variable.Table 3ParameterAjδνεηBnωhλkPRCC0.7851−0.99080.0813−0.47520.02240.07300.0320−0.0896−0.0298−0.0221−0.0238P valueP < 0.01P < 0.010.0399P < 0.010.48430.05270.31730.02460.35150.48940.4560 Partial rank correlation coefficients between new human brucellosis cases and each input parameter variable. The PRCC values of new human brucellosis cases versus to different parameters. The asterisk indicates that the parameter is significant. Sensitivity analysis of model parameters on new human brucellosis cases The effects of five model parameters on new human brucellosis cases in Changji city are presented in Fig. 10. As the new recruitment of sheep/cattle A decreases, the peak magnitude of new human brucellosis cases declines, and the peak time is accordingly delayed. When A < 2,995, new human brucellosis cases demonstrate a sustained decline approaching zero, as shown in Fig. 10A. Higher culling rates of infected sheep/cattle j result in fewer new human brucellosis cases and delayed peak time. When j > 0.1287, new human brucellosis cases demonstrate a continuous decreasing trend toward elimination, as presented in Fig. 10B. Similarly, increased vaccination rates in susceptible sheep/cattle v correlate with fewer new human brucellosis cases and postponed peak occurrence. When v > 0.1618, new human brucellosis cases maintain a continuous decreasing trend toward elimination, as illustrated in Fig. 10C. Reducing the immune loss rate of vaccinated sheep/cattle δ reduces the peak magnitude and delays the peak time. However, even when δ approaches zero, new human brucellosis cases still exhibit an initial increase followed by a subsequent decline, as shown in Fig. 10D. Increasing the disinfection frequency n amplifies the fluctuation amplitude of new human brucellosis cases fluctuations while reducing trough values to zero, even at n = 10, the reduction in new human brucellosis cases remains minimal, as illustrated in Fig. 10E. The basic reproduction number R0 quantifies the transmission intensity of brucellosis. When R0 > 1, higher values correlate with an earlier peak time and greater peak magnitude of new human brucellosis cases, indicating strong disease transmissibility. Conversely, when R0 < 1, new human brucellosis cases demonstrates a sustained decline approaching zero, signifying gradual disease elimination. Impacts of five model parameters on new human brucellosis cases. (A) New recruitment of sheep/cattle A. (B) Culling rate of infected sheep/cattle f. (C) Vaccination rate of susceptible sheep/cattle v. (D) Immune loss rate of vaccinated sheep/cattle δ. (E) Disinfection frequency n. Sensitivity analysis of model parameters on the basic reproduction number R0 The respective effects of parameters A, j, v, δ, and n on the basic reproduction number R0 of brucellosis transmission are illustrated in Fig. 11. The analysis reveals three critical thresholds for disease control: when the new recruitment of sheep/cattle A < 2,995, the culling rate of infected sheep/cattle j > 0.1287, or the vaccination rate of susceptible sheep/cattle v > 0.1618, R0 can be reduced to below 1 (Fig. 11A–C). This threshold indicates effective disease control, characterized by a sustained decline in incidence and eventual elimination of the disease. The immune loss rate of vaccinated sheep/cattle δ shows a positive proportional relationship with R0. However, even when δ = 0, R0 remains at 1.8769 (Fig. 11D), demonstrating that this parameter alone cannot interrupt brucellosis transmission. Environmental disinfection n exhibits diminishing returns, while n = 1 produces a notable reduction in R0, additional disinfection with n > 1 yields progressively smaller benefits (Fig. 11E). Importantly, even extreme disinfection (n = 10) fails to achieve R0 < 1. Our analysis demonstrates that reducing the new recruitment of sheep/cattle A, increasing the culling rate of infected sheep/cattle j, and increasing the vaccination rate of susceptible sheep/cattle v are effective measures to control the spread of brucellosis. In addition, reducing the immune loss rate in vaccinated sheep/cattle δ and increasing environmental disinfection frequency n can serve as supplementary measures to effectively control brucellosis transmission in Changji city. Effects of five model parameters on the basic reproduction number R0.(A) New recruitment of sheep/cattle. (B) Culling rate of infected sheep/cattle f. (C) Vaccination rate of susceptible sheep/cattle v. (D) Immune loss rate of vaccinated sheep/cattle δ. (E) Frequency of disinfection n. Optimal control measures analysis Among measures for controlling brucellosis transmission, reducing the new recruitment of sheep/cattle A will have a significant impact on pastoralists' economic income. Lowering the immune loss rate of vaccinated sheep/cattle δ primarily relies on continuous monitoring of antibody levels to assess immunogenicity, which requires substantial investment of funds and manpower. Therefore, this study analyses the culling rate of infected sheep/cattle j, the vaccination rate of susceptible sheep/cattle ν, and the frequency of environmental disinfection n to comprehensively control the transmission of brucellosis. The contour lines for the parameter on R0 are shown in Fig. 12. When j > 0.0946 and n > 1.25, R0 can be reduced below 1, thereby controlling the spread of brucellosis (Fig. 12A). When ν > 0.1124 and n > 1.25, R0 < 1 can be achieved, thereby effectively controlling brucellosis (Fig. 12B). When ν > 0.0612, j > 0.0745 and n = 0, R0 < 1 can be achieved, leading to the gradual eradication of brucellosis (Fig. 12C). We can calculate that when n = 1, R0 = 2.4804, and when n = 2, R0 = 2.4333, differing by merely 0.0471. Increasing the disinfection frequency beyond this point has negligible impact on R0. Therefore, n = 1 is deemed the most appropriate disinfection frequency. When ν > 0.0548, j > 0.0660, and n = 1, R0 < 1 can be achieved, which is the optimal prevention and control strategy for brucellosis transmission (Fig. 12D). Contour plots of R0 on j, v and n in Changji city. (A) The impact of j and n on R0. (B) The impact of v and n on R0. (C) The impact of j and v on R0. (D) The impact of j, v and n on R0. The orange curve represents the R0 contour lines. The red dot indicates the threshold for the selected control measure. Discussion Brucellosis remains one of China's most pressing public health challenges. Despite recognition of its significant health and economic impacts and the implementation of certain control measures, the disease continues to exhibit relatively high outbreak frequency and case incidence (Huang & Li, 2020). Consequently, both central and local governments have actively implemented region-specific prevention and control strategies to significantly reduce the transmission risk of brucellosis, ultimately achieving the goals of disease control and elimination. Changji city in Xinjiang maintains substantial sheep/cattle populations, and the number of emerging human brucellosis cases in the area shows an increasing annual trend. Therefore, predicting case fluctuations and implementing targeted prevention and control measures are critical for curbing the transmission of brucellosis. Through mechanistic analysis of brucella transmission among sheep/cattle, the environment, and humans, we developed a dynamic model in which infection is spreads via exposed/infected animals and environmental contamination. The model incorporating periodic transmission rates to reflect seasonal patterns, effectively fitted the rising trend of cases during 2020-2023. It also predicts the epidemic trends for the subsequent 25 years, showing that the number of cases will first rise and then stabilize. Finally, sensitivity analysis of parameters affecting the number of new human brucellosis cases and the basic reproduction number R0 in Changji city. Quantitative results demonstrate that reducing the new recruitment of A, increasing the culling rate of infected sheep/cattle j, improving the vaccination rate of susceptible sheep/cattle v are the most effective control measures. Decreasing the immune loss rate δ and increasing environmental disinfection frequency n play supplementary roles. Taken together, these measures provide a feasible strategy for effectively controlling and gradually eliminating brucellosis in the region. This study demonstrates that reducing the new recruitment of A is one effective approach to controlling the brucellosis transmission in Changji city. The model indicates that eliminating brucellosis would require reducing the current new recruitment of sheep/cattle by approximately two-thirds (A < 2995). While epidemiologically effective, this intervention would significantly impact local livestock production and economic income, rendering its practical implementation value is limited after comprehensive assessment. Conversely, vaccination can effectively reduce human brucellosis cases, underscoring the critical role of vaccination rates in brucellosis prevention and control strategies. Our findings indicate that vaccination alone can effectively control the spread of brucellosis only when v > 0.1618, which means the vaccination rate needs to be increased by approximately 87% compared to current levels. This suggests that the current vaccination coverage in Changji city is still insufficient and urgently needs to be improved. Under current conditions, approximately 65% of breeding ewes/cows in the annual stock remain unvaccinated, making it challenging to achieve the prevention and control targets even if all other sheep/cattle are fully vaccinated. These findings align with previous research reports (Ma, Li, et al., 2022; Zhou et al., 2018). To enhance vaccination effectiveness, future efforts should prioritize vaccine improvement, novel technology development, and strategic vaccination campaigns, while accounting for the complex social network structures and higher-order interactions influencing disease transmission (Aznar et al., 2017; Malik et al., 2023). Targeted culling of infected sheep/cattle serves as an effective intervention measure, as it can eliminate the primary infection sources. Simply increasing the culling rate (f > 0.1287, representing approximately 75% increase over current levels) can significantly reduce the incidence of human brucellosis, particularly during early control phases, ultimately achieve disease elimination. In practice, a substantial number of infected sheep/cattle remain unculled and continue to transmit brucella to humans and other livestock, likely due to the substantial economic losses incurred by livestock owners. Environmental disinfection measures alone demonstrate limited effectiveness in brucellosis control, as evidenced by previous studies, showed that only marginal reductions in R0 despite increased disinfection frequency (Liu et al., 2025; Peng et al., 2020). Therefore, environmental disinfection should be implemented in combination with other interventions (such as culling and vaccination) to control brucellosis more effectively. In Changji city, the brucella suis strain S2 vaccine has been used for nearly 50 years to prevent brucellosis transmission among sheep and cattle, playing a crucial role in disease control by effectively reducing livestock abortion rates and minimizing economic losses for farmers. The S2 vaccine provides protection lasting approximately 2-3 years, demonstrating an 82% efficacy rate against brucella infection in vaccinated sheep/cattle (Lou et al., 2016). In this study, the immune loss rate δ was 0.0333. Even if reduced to zero, it remained insufficient to control the brucellosis transmission in Changji city, as evidenced by the continued gradual increase in new human brucellosis cases, indicating the necessity of implementing other control measures in combination. Ultimately, we comprehensively considered vaccination rates of susceptible sheep/cattle ν > 0.0548, culling rates of infected sheep/cattle j > 0.0660, and environmental disinfection frequency n = 1 as the optimal measures for controlling brucellosis transmission in Changji city. The gap between the target levels and the current actual levels of each control measure is not significant, ensuring the feasibility of implementation. The transmission intensity of brucellosis can be quantified by the basic reproduction number R0. When R0 > 1, with higher values indicating stronger transmission, brucellosis will persist and cannot be eliminated. Conversely, when R0 < 1, the smaller the value, the weaker the transmission capacity, the disease can be effectively controlled and gradually fade out. Our study estimates the basic reproduction number for brucellosis transmission in Changji city as R0 = 2.8228, indicating a serious epidemic situation characterized by an increasing trend in new human brucellosis cases. Lou et al. (Lou et al., 2016) estimated the basic reproduction number R0 of brucellosis transmission as 2.5524 in Bayingolin Mongol Autonomous Prefecture, Xinjiang, by using an SEIV dynamic model, a result that highly consistent with our findings. Hou et al. (Hou et al., 2013) developed an SEIVB dynamic model for brucellosis transmission in Inner Mongolia, estimating an R0 of 1.8. Liu et al. (Liu et al., 2025) developed a dynamic SEIVWShIahIch model to simulate the transmission of brucellosis in Inner Mongolia, and calculated the R0 = 2.86, which closely aligns with the results of our study, suggesting a concerning epidemic trend with increasing incidence in this region. Li et al. (Li et al., 2014) proposed a deterministic SoEoIoVoSfEfIfVfWShIhYh model estimating an R0 as 1.9789 for Hinggan League, Inner Mongolia. Qin et al. (Qin et al., 2022) established a patch dynamics model, revealed R0 values of 0.7497 and 0.5022 for Shanxi and Hebei provinces respectively, that the disease had been effectively controlled. Gong et al. (Gong et al., 2023) proposed a dynamical model of sheep-human-environment, which calculated an R0 of 1.47 for Ningxia, suggesting persistent brucellosis transmission in this area. Li et al. (Li, Sun, et al., 2017) developed a dynamic model to fit annual new human brucellosis cases across 11 provinces in mainland China from 2004 to 2014, estimating R0 values ranging from 1.0227 to 1.8348. In summary, the brucellosis epidemic in Changji city, Xinjiang demonstrates comparable severity to Inner Mongolia and exceeds that of other provinces, aligning with the actual transmission patterns observed across the country. This phenomenon may be attributed to two key factors (Dean et al., 2012): the substantial population of sheep/cattle in Changji city, where brucella infection has led to a gradual increase in the number of human cases (Dadar et al., 2021); the utilization of monthly human brucellosis case data in our study, which captures detailed transmission dynamics, whereas previous studies relying on annual data may have overlooked critical temporal variations in disease spread. This study has several limitations (1): The seasonal model constructed in this study fails to incorporate temperature variation, a critical environmental factor, which may reduce the scientificity and practicality of its prediction results. Temperature directly affects the transmission of brucellosis and significantly alters the survival time of brucella in the environment, thereby influencing the transmission intensity and epidemic cycle of the disease. The absence of this factor in the model will lead to deviations between the simulated seasonal transmission pattern of the disease and the actual situation. In addition, it may also result in the underestimation or overestimation of transmission risks, thereby affecting the prediction of epidemic trends (2). The model relies on the number of confirmed cases, but there is a time lag between symptom onset and the cases confirmation, which may compromise the accuracy of predictions (3). The model does not include several livestock species that can transmit brucellosis as infection sources, such as pigs, horses, and dogs. We will conduct further in-depth research on the transmission patterns of brucellosis, optimize the application of the model, and develop more accurate and effective strategies for disease prevention and control. Conclusion The developed seasonal dynamic model SEIVW effectively simulated and predicted the fluctuating patterns of new human brucellosis cases in Changji city, demonstrating a periodic yearly increasing trend. Meanwhile the estimated basic reproduction number R0 indicates persistent brucellosis transmission that could not be eliminated under the current conditions. Sensitivity analysis of parameters quantitatively identified effective intervention measures to curb the brucellosis transmission, providing scientific theoretical support for the Changji City Center for Disease Control and Prevention in formulating prevention and control strategies. CRediT authorship contribution statement Pengwei Lou: Writing – original draft, Validation, Software, Resources, Project administration, Investigation, Formal analysis, Data curation, Conceptualization. Yuting Huang: Visualization, Resources, Formal analysis, Conceptualization. Xiaolin Fan: Writing – original draft, Visualization, Software, Methodology, Formal analysis, Conceptualization. Nannan Chen: Validation, Data curation. Yifan Hu: Validation, Data curation. Honggang Liu: Validation. Feng Zhao: Writing – review & editing. Jiabo Xu: Writing – review & editing, Project administration, Funding acquisition. Enguang Gao: Writing – review & editing, Supervision, Investigation, Data curation. Kai Wang: Writing – review & editing, Supervision, Project administration, Methodology, Funding acquisition. Funding This study was supported by the National High-Level Talent Development Program (XYD2024GR04), the Regional Collaborative Innovation Program of the Xinjiang Uygur Autonomous Region (Shanghai Cooperation Organization Science and Technology Partnership Program and International Science and Technology Cooperation Program) (2025E01039), the National Natural Science Foundation of China (52564025). Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Handling Editor: Dr Daihai He Footnotes Peer review under the responsibility of KeAi Communications Co., Ltd. Supplementary data to this article can be found online at https://doi.org/10.1016/j.idm.2026.05.002. Contributor Information Pengwei Lou, Email: loupwei@163.com. Yuting Huang, Email: 953061361@qq.com. Xiaolin Fan, Email: mxjhfxl@163.com. Nannan Chen, Email: xuelvyi@foxmail.com. Yifan Hu, Email: 2725181852@qq.com. Honggang Liu, Email: 2740312889@qq.com. Feng Zhao, Email: zhaofeng1016@126.com. Jiabo Xu, Email: xujiabo_math@aliyun.com. Enguang Gao, Email: 1521913654@qq.com. Kai Wang, Email: wangkaimath@sina.com. Supplementary data The following is the Supplementary data to this article. Multimedia component 1 References Associated Data Supplementary Materials Multimedia component 1