Estimating Dengue Transmission Intensity In China Using Catalytic Model Based On Serological Data

Objectives:The southeast coast and southwestern border of China are areas with a high incidence of dengue fever.In view of the large number of recessive infections of dengue fever,the use of serological data to construct a catalytic model will help us clarify the epidemic characteristics of dengue fever in China and take targeted prevention and control measures.This study collects the age-specific serological data of dengue fever released in China in recent years,reviews the characteristics and historical changes of dengue fever transmission;constructs and optimizes catalytic models to estimate the sero-epidemiological parameters of dengue fever in China(force of infection,basic reproduction number,transmission intensity);Compare the development trend of dengue fever transmission intensity in the study area in the 30 years under different transmission mechanisms,and propose targeted prevention and control measures.Methods:1.Through literature search,look for research on dengue fever serology in China,and extract serological age stratified data.2.Construct catalytic models based on the extracted serological age stratified data,taking into account the influence of different transmission mechanisms(constant force of infection,antibody protection decay,critical age and constant force of infection,critical age and antibody protection decay).3.Use four catalytic models to fit the extracted serological age stratified data to estimate the force of infection and basic reproduction number.4.When carrying out parameter estimation,the idea of Bayesian statistics is used,priori information is given through literature review and past experience judgments,and the posterior distribution of parameters is calculated based on data extracted from literature research.Using the normal distribution random walk method in the Markov chain Monte Carlo method(MCMC),the R statistical software is used to independently sample the posterior distribution to obtain the expectation and 95%confidence interval of the parameter posterior distribution.5.The deviation information criterion(DIC)combines the fit and complexity of the model,and can be used to compare the fit effect of any structural model.Use DIC to compare the fitting effects of different models and select the best model for each dataset.6,Explore the transmission characteristics of dengue fever in different populations by comparing the parameter information under different transmission mechanisms of each research population.7.Comparative analysis of the current situation and prevention and control measures of dengue fever in several cities and regions in China,combined with the transmission intensity estimated by the model,and put forward targeted dengue prevention and control recommendations.Results:1.Through literature search,we included 23 studies from the 515 selected articles.These studies reported the transmission of dengue fever in 14 regions in 8 provinces of China from 1981 to 2017.2.Through reviewing historical documents,it is found that there are relatively few studies on dengue fever serology in China,especially the lack of serological age stratification data and specific serological data.3.We considered the different transmission mechanisms of dengue fever and constructed four catalytic models.A.Constant force of infection model,which assumes that FOI remains constant in all age groups;B.Antibody protection decay model,which assumes that antibodies in seropositive people will decay at a constant rate over time.After finally falling to a critical level,it turns into a seronegative population(susceptible population);C.Critical age and constant force of infection model,this model assumes that the risk of dengue infection will change around a certain age in the study population,leading to the FOI of dengue fever has changed around the critical age;D.Critical age and antibody protection decay model,this model still assumes that there is a change in the risk of dengue infection in the study population around a certain age,and this change may be due to the effect of antibody protection decay.4.The DIC compares the fitting effects of the four models and selects the best model for each dataset.The results show that the best model of 4 studies is model A;the best model of 7 studies is model B;the best model of 8 studies is model C,and the best model of 4 studies is model D.5.The parameter estimation results of the best model of the dataset show that the estimated value of the basic reproduction number(R0)is 1.04～2.24.The largest R0 is the study conducted in Guangzhou in 1981,R0=2.24(95%CI:1.71～3.00);the smallest R0 are conducted in Zhuhai in 1998,R0＝1.04(95%CI:1.01～1.15),and the research conducted in Qinzhou City,Fangchenggang City and Hepu County from 2010 to 2012,R0=1.04(95%CI:1.02～1.19).The best model is that in the four studies of Model A,the estimated value of R0 is 1.04～1.14,and the estimated value of FOI is 0.0011～0.0083;The best model is that in the four studies of Model B,the estimated value of R0 is 1.05～2.24,and the estimated value of FOI is 0.0037～0.0150;and the average antibody protection decay rate is 0.07 per year;the best model is model C in 8 studies,the critical age(Acrit)is estimated to be 15.2～67.2 years,and the estimated FOI is 0.0010-0.0228;In 4 studies,the best model is Model D,Acrit is estimated to be between 13.0～54.5 years,the estimated value of R0 is 1.06～1.54,and the estimated value of FOI is 0.0113～0.1238.6.The estimated FOI of Model A is less than the estimated FOI of Model B,which indicates that if there is an effect of antibody protection decay,the risk of dengue infection in the population is significantly increased,which has important guiding significance in the prevention and control of dengue fever.7.For the 12 datasets where the best models are model C and model D,there are different FOIs before and after Acrit.The results of the correlation analysis between the FOI difference(λdif=|λ1-λ2|)and Acrit show that there is a significant negative correlation with the critical age(P<0.05),and the correlation coefficient is-0.69.This indicates that the greater the population’s risk of dengue infection changes,the lower the critical age for this change to occur.Conclusion:1.The study found that there are differences in the transmission mechanism of dengue fever in different regions of China.This is confirmed by comparing the DIC of different catalytic models.This study provides a new idea for distinguishing the transmission characteristics of dengue fever in different regions.2.The spread of dengue fever under different mechanisms presents different epidemic characteristics and requires targeted prevention and control measures.This is conducive to our rational allocation of health resources and cost savings.3.Age is closely related to the spread of dengue fever.There may be a special age window(Acrit)in the population,during which the risk of dengue infection has changed significantly,leading to different epidemic characteristics in the population based on this age.People before and after Acrit present different risks of dengue infection.For people whose FOI before Acrit larger than that after Acrit,the key population of dengue prevention and control measures should be the young people before Acrit;for the population before Acrit,the FOI is smaller than the population after the key population of dengue prevention and control measures should be the population after Acrit.The greater the change in the risk of dengue infection,the lower Acrit when the change occurs.4.Antibody protection decay is also closely related to the spread of dengue fever.Antibody protection decay will reduce the level of individual antibodies,resulting in a population whose FOI before Acrit is less than that after Acrit.5.The advantage of using serological data for dengue research is that it is not affected by infectious disease surveillance systems and case reporting systems.Serological testing can determine past dengue infections and is of great significance for us to estimate the potential spread of dengue fever.