| BackgroundDengue fever is a mosquito-borne infectious disease caused by dengue virus and transmitted by Aedes mosquitoes.In recent years,the incidence of dengue fever in China has shown an upward trend with expanding distribution.Therefore,description of the epidemic characteristics of dengue fever in China and identification of the period and area with high incidence is important to delimit the key areas of dengue fever prevention and improve the dengue fever control strategies.Although previous studies have analyzed the influencing factors of dengue fever incidence,most of them focused on meteorological factors,lacking comprehensive exploration of meteorological,natural environmental,and socioeconomic factors.Meanwhile,there is no research on the spatial heterogeneity of the correlation intensity and direction at the district/county level using spatial regression models.Dengue fever,as a climate-sensitive disease,is closely related to meteorological factors,however,the conclusions among relevant research were not consistent due to the different study periods,areas,and statistical methods.In addition,most studies paid attention to the conventional meteorological monitoring indicators such as temperature,precipitation,relative humidity,and wind speed,while few studies have focused on the impact of local hydrometeorological conditions on dengue fever,especially in China.Although some studies have analyzed the effects of socioeconomic factors on the incidence of dengue fever,few studies have systematically evaluated the quality of city development from multiple dimensions,and explored its modification effects on the association between the meteorological factors and the incidence of dengue fever.Therefore,it is necessary to explore the nonlinear lag relationship between the key meteorological factors and the dengue fever incidence in the high-risk period and region in China,and evaluate the modification effects of city development quality.Climate change has become a major threat and challenge to the world,which will affect the survival,development and reproduction of Aedes,and hence affect the dengue fever incidence.So far,few studies have predicted the excess risk of dengue fever due to climate change in the future.Therefore,it is necessary to project the excess risk of temperature and hydrometeorological conditions on the distribution of Aedes and the incidence of dengue fever in China under different climate change scenarios.The results have important public health significance for developing and adjusting future relevant policies and curbing the spread of dengue fever.Objectives1.To describe the distribution of dengue fever in China,explore the spatial distribution pattern and spatiotemporal aggregation characteristics,and identify the period and region with high incidence of dengue fever.The findings can provide a basis for the selection of the study period and area in the next part of the study.2.To explore the impact of meteorological,natural environmental,and socioeconomic factors on the incidence of dengue fever in Guangdong province in 2014 and the spatial heterogeneity of the intensity and direction of the effects,identify the main contributors,and analyze the interaction effect between them.3.To evaluate the nonlinear lag association between temperature,hydrometeorological conditions,and the incidence of dengue fever,and explore the modification effect of varying dimensions of city development quality on this relationship.4.To project the distribution of area with dengue fever risk in China in different periods under climate change scenarios in the future,predict the excess risk of dengue fever associated with hydrometeorological conditions by cities,geographical regions,climate zones,and areas with different population density and GDP per capita levels,and quantify the excess risk attributed to wet and dry condition,respectively.Methods1.Data collectionDengue fever cases in China during 2005-2019 were from the national notifiable infectious disease surveillance system of the Chinese Center for Disease Control and Prevention.The meteorological data of Guangdong province in 2014 were obtained from the National Earth System Science Data Center of the National Science&Technology Infrastructure of China,the natural environment data and population data were collected from the Resource and Environment Science and Data Center of the Institute of Geographic Sciences and Natural Resources Research,CAS,and the socioeconomic data came from the city statistical yearbooks and health statistical yearbooks.The meteorological data of Zhejiang,Fujian,Guangdong,and Guangxi during 2013-2019 were from the CN05.1 dataset of the National Climate Center of the China Meteorological Administration,and the city development quality indicators were obtained from the city statistical yearbooks and TapData dataset.The future temperature and precipitation data were obtained from the simulations under the three representative concentration pathways developed by the Intergovernmental Panel on Climate Change.The future population data came from the gridded population projection dataset for China,and the population under the low,medium,and high fertility scenarios were generated respectively.All the gridded data were then processed into vector data.The hydrometeorological indicators were calculated using monthly average temperature and precipitation.2.Statistical analysis(1)The distribution of dengue fever cases in China during 2005-2019 was described,the spatial pattern of the number of cases at city level was determined using spatial autocorrelation analysis.Space-time scan analysis was used to identify the spatiotemporal clusters of dengue fever in China during the study period.(2)The cumulative incidence(CI)of dengue fever in all districts/counties of Guangdong province from July to December 2014 was adjusted using the Besag-York-Mollie Bayesian spatial model.Spearman correlation analysis and variance inflation factor were used to test the multicollinearity of meteorological,natural environment,and socioeconomic factors.The association between the variables included and dengue fever CI was assessed using ordinary least square(OLS)regression model and global spatial regression models,expressing as odds ratio(OR)and 95%confidence interval(CI).Then,geographic weighted regression(GWR)model was used to analyze the spatial heterogeneity of the intensity and direction of their effects.Finally,the factors affecting the spatial stratified heterogeneity of dengue fever CI and their interactions were identified using geographical detector technology.(3)The nonlinear lag association between the temperature,hydrometeorological conditions and the incidence of dengue fever in Zhejiang,Fujian,Guangdong,and Guangxi during 2013-2019 was fitted using the hierarchical Bayesian spatiotemporal model and distributed lag nonlinear model.The goodness-of-fit of the models was evaluated by deviation information criterion(DIC),and the optimal model was selected with the lowest DIC value.The modification effect of different city development quality levels on the association between temperature or hydrometeorological conditions and the dengue fever incidence was evaluated by incorporating the linear interaction terms of the cross-basis function between the city development indicator and the meteorological variables,and centering it to the value of the 10th,50th,and 90th percentiles among all cities.Sensitivity analysis was conducted by adjusting the parameters of the cross-basis of meteorological variables or replacing the included variables.(4)The linear terms of city development indicators were included in the model built in Part 3,and the DIC was used to identify the optimal model to assess the association between hydrometeorological conditions and dengue fever incidence in the historical period(20132019).A biological model of dengue fever risk was constructed,considering the temperature threshold for Aedes development and dengue virus transmission,as well as the potential transmission index of dengue fever,to predict the dengue fever risky area and period in the future.Finally,according to the historical association model and the biological model,the monthly excess risk(MER)and the population-weighted monthly excess risk(PWMER)of dengue fever incidence attributable to hydrometeorological conditions at city level in China during 2021-2040,2051-2070,and 2081-2100 under RCP2.6,RCP4.5,and RCP8.5 scenarios were projected,assuming a constant and dynamic population in the future,respectively.Meanwhile,the average MER of each geographical region,climate zone,population density and GDP per capita level was calculated,the MER attributable to wet and dry condition was also computed,respectively.Monte Carlo simulation was used to estimate the 95%empirical credible interval(eCI)of MER.Sensitivity analysis was conducted by adjusting the parameters of the cross-basis of hydrometeorological variables or replacing it with suboptimal models.Results1.A total of 96 540 confirmed dengue fever cases were reported in China during 20052019,and the number of cases increased sharply in 2013-2019.The incidence showed obvious seasonality,with high incidence in summer and autumn.The spatial distribution of the dengue fever incidence showed a decreasing trend with the increase of latitude.The distribution of dengue fever in China presented a positive spatial correlation.The high-high concentration areas and hot spots were mainly distributed in Guangdong,Yunnan,Fujian,Zhejiang,etc.The results of spatiotemporal clustering analysis showed that the first cluster was the central Guangdong from September to October 2014,the second cluster was the southwest China from July to November 2019,and the third cluster was the East China from July to November 2019.2.From July to December 2014,45 141 local dengue fever cases were reported in 123 districts/counties in Guangdong province.The adjusted dengue fever CI ranged from 13.37 to 6 548.33 per million.The high-high concentration area was located in the central and southern Guangdong.OLS results showed that the average temperature(OR=1.202,95%CI:1.0571.366),PDSI(OR=1.186,95%CI:1.096-1.284),proportion of water area(OR=1.107,95%CI:1.014-1.207),population density(OR=1.323,95%CI:1.134-1.543),and GDP per capita(OR=1.152,95%CI:1.026-1.293)increased the dengue fever CI,while the relative humidity(OR=0.899,95%CI:0.815-0.991)and sunshine hours(OR=-0.779,95%CI:0.677-0.897)reduced the dengue fever CI.The results of spatial Dubin model,the best global spatial model,demonstrated that the proportion of farmland(OR=1.236,95%CI:1.033-1.479)and population density(OR=1.546,95%CI:1.086-2.199)in the neighboring areas increased the local dengue fever CI.The GWR results showed that the strength and direction of the associations between the independent variables and the dengue fever CI were spatially heterogeneous,and the goodness-of-fit of GWR model was better than that of the global regression models.The℃eographical detector results showed that PDSI(q=0.463,P<0.001),relative humidity(q=0.417,P<0.001),GDP per capita(q=0.373,P<0.001),and average temperature(q=0.270,P<0.001)were the most important explanatory variables of dengue fever CI,and there were two-factor enhanced interaction between them.3.From 2013 to 2019,70 053 dengue fever cases were reported in Zhejiang,Fujian,Guangdong,and Guangxi.There was nonlinear and delayed effect of monthly average temperature on the dengue fever incidence.The effect was nonsignificant when the average temperature below 20℃.With the increase of temperature,the risk of dengue fever increased rapidly,and the largest risk was found in lag 2 month at 31.5℃[relative risk(RR)=2.31,95%Cl:1.60-3.35].Compared with the median temperature(21.3℃);the highest cumulative RR within 6 months was occurred at 27.1℃(RR=8.27,95%CI:4.18-16.40).The effect of wet condition on the dengue fever incidence existed throughout the whole lag period,with the maximum effect occurring at the three-month standardized precipitation evapotranspiration index(SPEI3)equating to 2.7 in lag 2 month(RR=1.71,95%CI:1.33-2.20).The effect of dry condition occurred after lag 3 month,reaching the maximum when SPEI-3=-2.4 in lag 6 month(RR=1.91,95%CI:1.40-2.61).Extreme wet and extreme dry condition can both increase the risk of dengue fever,and the cumulative RR was 3.27(95%CI:1.83-5.86)and 2.93(95%CI:1.40-6.17),respectively.The city development quality had modification effects on the association between temperature,hydrometeorological conditions and dengue fever incidence.The increase of GDP per capita,number of doctors per capita,and disposable income per capita reduced the risk of high temperature and extreme hydrometeorological conditions on the dengue fever incidence.With the increase of green area per capita,the risk of high temperature on dengue fever was reduced,while the risk of extreme dry condition was increased.The risk of high temperature on dengue fever decreased with the increase of urbanization rate.The risk of dengue fever was higher in rural areas after high temperature and extreme dry condition,but higher in highly urbanized areas after extreme wet condition.Sensitivity analysis showed that the model results were robust.4.The national SPEI-3 was expected to decrease in the future,especially under RCP8.5 scenario.The results of association between hydrometeorological conditions and dengue fever in the historical period showed that both extreme wet and extreme dry conditions increased the risk of dengue fever after controlling the socioeconomic variables.In the future,the risk areas of dengue fever in China would gradually expand under all RCP scenarios,especially under RCP8.5 scenario.Under RCP2.6 scenario,the MER due to SPEI-3 was projected to be higher in the low-latitude coastal cities,and the peak would occur in 2051-2070,with the national average MER=12.87(95%eCI:11.12%-15.80%).Under RCP4.5 scenario,the MER increased significantly in 2081-2010 with the national average MER=1 7.00%(95%eCI:14.69%-20.86%).The MER was expected to be further increased under RCP8.5 scenario,with the majority of coastal cities exceeding 50%in 2081-2100,and the national average MER=41.82%(95%eCI:36.46%-51.16%).Compared to MER,PWMER was expected to show different degrees of decrease in different periods.South China and tropical monsoon climate zone were predicted to have the highest MER in China in the future.Additionally,the MER was expected to be higher with the increase of population density and GDP per capita.The MER was expected to be mainly attributed to dry condition under all scenarios.The projections were proved to be robust by sensitivity analysis.Conclusions1.There was an upward trend in the dengue fever incidence in China from 2005 to 2019.The incidence of dengue fever is concentrated in summer and autumn,with a high incidence in the southeast coastal areas.The distribution of dengue incidence was spatially clustered in China,concentrated in Guangdong,Yunnan,Fujian,and Zhejiang.There were three spatiotemporal clusters,namely the central part of Guangdong province in 2014 and the southwest and east China in 2019.2.The outbreak of dengue fever in Guangdong in 2014 was associated with the average temperature,PDSI,proportion of water area,population density,GDP per capita,relative humidity,and sunshine hours.There was a positive correlation between the population density and GDP per capita of surrounding districts/counties and the dengue fever incidence.The intensity and direction of their effects had obvious spatial heterogeneity.PDSI,relative humidity,GDP per capita and average temperature were the main factors driving the spatial stratified heterogeneity of dengue fever,and there were two-factor enhanced interactions between them.This study proved that the incidence of dengue fever was jointly affected by meteorological,natural environment and socioeconomic factors,and the main meteorological factors were temperature and hydrometeorological conditions.3.There was a nonlinear lag relationship between temperature,hydrometeorological conditions,and the incidence of dengue fever.High temperature,extreme wet and dry was associated with the increase of dengue fever incidence.The city development quality in each dimension modified the associations in different patterns.4.With the rise of temperature in the future,the areas with high risk of dengue fever in China would further expand to the north.The excess risk of dengue fever associated with hydrometeorological conditions would show obvious spatiotemporal heterogeneity.The excess risk in South China,the tropical monsoon climate zone,and the areas with higher population density and GDP per capita would be higher,and the excess risk due to dry condition was greater than that due to wet condition.It suggests that effective mosquito and dengue fever prevention strategies should be developed to reduce the burden of dengue fever caused by future temperature and hydrometeorological changes.Innovations1.Hydrometeorology is a comprehensive meteorological indicator considering temperature,precipitation,potential evapotranspiration,and so forth simultaneously.Few studies focusing on the relationship between hydrometeorological conditions and the incidence of dengue fever in China.This study used several hydrometeorological indicators and statistical models to analyze the effects of hydrometeorological conditions on the incidence of dengue fever.The results enrich the epidemiological evidence of the effect of meteorological factors on the incidence of dengue fever,and provide further scientific basis for the development of dengue fever prevention and control strategies.2.Based on the latest national standard,this study systematically evaluated the city development quality from five dimensions,and explored their modification effects on the association between meteorological factors and dengue fever incidence.The results make up for the limitations of previous studies on the modification effect of a single socioeconomic indicator,and have important significance for guiding different regions to conduct targeted dengue fever prevention measures based on the city development quality context.3.This study combined the biological model considering the impact of temperature on Aedes and dengue virus with the spatiotemporal Bayesian hierarchical model considering the impact of hydrometeorological conditions on the incidence of dengue fever,to comprehensively predict the excess risk of dengue fever in China under the change of temperature and hydrometeorological conditions in the future.This research framework provides a methodological reference for the future risk projection of other vector-borne diseases with limited geographical distribution. |
PDF Full Text Request