Multi-scale Spatiotemporal Influencing Factors Of Dengue Fever In China

BackgroundThe viral disease Dengue fever is transmitted by infected female Aedes,mainly Aedes aegypti and Aedes albopictus.The virus has four serotypes:DENV1,DENV2,DENV3 and DENV4.Affected individuals have influenza-like symptoms,3-17 days(average 5-7 days)after infection.The disease incidence had a 30-fold increase over the past 50 years,with an estimate of 50 million to 100 million dengue infections in over 100 countries annually,putting half of the world’s population at risk.After World War II,dengue fever was mainly prevalent in the southeastern coast areas and Taiwan of China,but there were few cases reported from 1949 to 1977.Until 1978,the first outbreak of dengue fever occurred in Foshan city of Guangdong Province.Infection spread from Guangdong and Hainan Provinces to other coastal and inland areas.At present,in the studies of macro-influencing factors on dengue fever in domestic and abroad,dengue fever was mainly affected by climate factors,ecological environment and socio-economic factors.However,on the one hand,many studies on the influencing factors of dengue fever in China concentrated on the time models,on the other hand,mainly studies concentrated on the high-risk areas of Guangdong Province.Moreover,dengue fever is widely distributed in tropical and subtropical areas,and spreading in China with varying risk.Therefore,it is necessary to comprehensively analyze the influencing factors on dengue fever from the perspective of time and space and different scales.For example,from the different views of the whole of China and different high risk areas of dengue fever,it will be better to understand the spatial distribution of dengue fever in China and the risk factors affecting the occurrence of dengue fever,and to find out the driving factors on dengue fever in different high risk areas.This study examines the temporal and spatial characteristics of dengue fever disease and its influencing factors.It ultimately will provide recommendations to decision makers to effectively allocate health resources for the control and prevention of the disease.MethodsDengue cases from 2005 to 2017 were collected from the Information System of the Chinese Center for Disease Control and Prevention.The number of dengue fever at county level in China from 1978 to 2004 was sorted out from historical records.Meteorological data From the Chinese Meteorological Data Sharing Service System from 1978 to 2017 were downloaded,the raster data of average value of every meteorological variable from 1970 to 2000 were collected from WorldClim,and the raster data of the world population density of 2007 were collected from Netherlands Environmental Assessment Bureau.The East Asian Summer Monsoon Index,Normalized Difference Vegetation Index,land cover and altitude were collected from the Institute of Global Change and Earth System Science of Beijing Normal University,the Institute of Geospatial Data Cloud,the Institute of Geospatial Science and Resources of the Chinese Academy of Sciences and the European Space Agency respectively.Surveillance data of Aedes larvae BI of high-risk areas of Yunnan Province were collected from the "China Key Infectious Diseases and Vector Biological Detection Report" from 2006 to 2017.Aedes larvae(BI)data in high risk areas of Guangdong Province were collected from the national vector biological surveillance data.Firstly,ArcGIS 10.3.0 was used to analyze the spatial correlation of the disease from 2005 to 2017 and SatScan9.4.4 to measure spatial and temporal aggregation.In addition,we analyzed the impact of meteorological and environmental factors on the spatial distribution of dengue fever on a large scale and evaluate the high-risk areas of dengue fever using ecological niche model.The generalized additive model was used to analyze the relationship between regional climate of East Asian Summer Monsoon and dengue fever outbreak.Based on the single factor analysis method,the epidemiological characteristics of dengue fever cases in Guangdong and Yunnan Provinces were compared,such as sex,age,occupation,onset time,interval between onset and diagnosis.The effect factors on dengue fever occurrence were compared between Guangdong and Yunnan Provinces using generalized additive model,structural equation model and random forest method.ResultsFrom 2005 to 2017,a total of 60 448 cases occurred in China with a continuous increasing.High incidence areas were mainly in the Guangdong and Yunnan Provinces.Spatial-temporal clustering analysis revealed 7 main clustering areas.The highest clustering was in the southeast coastal area of China,including 13 counties and districts of Guangdong Province.There were 36957 cases with a relative risk of 5453.03 from September to October 2014.The second clustering region included a few counties and districts in Yunnan Province,with a total of 5390 cases from August 2013 to November 2017.The global autocorrelation analysis of indigenous cases of dengue fever in China from 2005 to 2017 showed a correlation between the spatial distribution and the occurrence of the disease with significant years being in 2008,2010,2012,2014,2016,2017 and the total 13 years.The spatial correlation analysis showed that the hotspots were slightly different for every year,but there were mainly three hotspots,namely,Guangdong,Yunnan and Fujian Provinces.In the analysis of ecological niche model for every year from 2005 to 2016,it was found that the daily average minimum temperature,land cover and daily average rainfall were significantly correlated with the spatial distribution of dengue fever with a non-linear relationship between the factors and the spatial distribution of dengue fever.The high-risk areas were mainly concentrated in Guangdong,Yunnan,Fujian and Zhejiang Provinces.Further studies revealed that dengue fever involving areas were widespread in most of China,and population density played an important role in the occurrence of dengue fever,with a contribution rate of 68.7%.More than 11 Dengue cases for every county mainly existed in southern and central China with the contribution rate of population density factor has gradually decreased to 28.8%.The counties with more than 101 cases of dengue fever were mainly confined to the southeast coastal areas,including Guangdong,Guangxi and Taiwan Provinces and the contribution rate of population density was 21.6%,and meteorological factors was the most important role.Based on the large-scale analysis of the relationship between the East Asian Summer Monsoon and the dengue fever outbreak,there was a non-linear negative relationship between the East Asian summer monsoon index and dengue fever,a non-linear positive correlation between the monthly mean temperature and dengue fever and a non-linear positive correlation between the cumulative rainfall and dengue fever last month.There was a total of 44674 indigenous cases of dengue fever in the high-risk area of Guangdong Province and 3676 indigenous cases of dengue fever were found in the high-risk areas of Yunnan Province.There were significant different in occupational,age,sex and time interval between onset and diagnosis of dengue fever in the two study areas.The local cases of dengue fever in Guangdong high risk areas were mainly housework and unemployed,while those in Yunnan province were mainly commercial service providers.Dengue fever cases in high-risk areas of Guangdong were widely distributed in all age groups and Yunnan Province mainly concentrates on the young and middle-aged.About the gender distribution,the male proportion of dengue fever in Guangdong Province was lower than that in Yunnan province.In the analysis of different factors affecting the occurrence of dengue fever in different hot spots of Guangdong and Yunnan Provinces,there was similar effects between the imported cases,Bretton index,temperature,rainfall and dengue incidence in different high risk areas using generalized additive model analysis but the descriptive indexes of temperature and rainfall in the best models for different regions were different.Random forest results showed that the imported cases,Bretton index,temperature and rainfall in Guangdong Province,but the most important factors in Yunnan Province are the number of imported cases last month,the monthly average temperature and the monthly relative humidity.The analysis of structural equation model in different high-risk areas showed that the last month imported cases and the density index of mosquito larvae had direct effects on the occurrence of dengue fever,while rainfall had indirect effects on the occurrence of dengue fever.The direct effect intensity of the larvae density index on dengue fever in Guangdong Province was higher than Yunnan Province.In addition,the temperature in Guangdong had a significant direct effect on the occurrence of dengue fever but Yunnan was notConclusionDengue fever was widely distributed throughout China with an outbreak in 2014.High-risk areas included Guangdong,Yunnan,Zhejiang and Fujian provinces.The large-scale distribution of dengue fever in China was mainly determined by population density,while outbreaks were mainly related to temperature,rainfall and regional climate such as the East Asian Summer Monsoon.In the small-scale analysis,the epidemiological characteristics of dengue fever cases in different high-risk areas were different and temperature,rainfall and humidity and mosquito density were the key drivers on dengue fever incidence in different high-risk areas of China.