| Objectives This study aims to describe the spatial distribution characteristics of Japanese encephalitis(abbreviated as JE)in Gansu province from 2005 to 2018.Irregular spatial scanning technology was used to detect the concentrated areas of JE in Gansu province in each year at county level.At the same time,the spatial regression statistical method was used to explore the relationship between the incidence of JE and meteorological and socio-economic factors in Gansu province.This study will provide scientific evidences for the regional prevention and control of JE in Gansu province.Methods This study has collected clinically and laboratory-diagnosed cases of JE in Gansu province from 2005 to 2018.Then these cases were combined with the population of each county in the same year,meteorological factors,and socioeconomic factors to establish a geographic information database of JE in Gansu province.We have described the distribution characteristics of JE in Gansu province on spatial level,irregular spatial scanning was used to explore the clustering of JE in Gansu province and ordinary least squares regression model and global spatial regression model were used to explore the correlation between the incidence of JE in Gansu province and meteorology and socioeconomic factors.We selected the best model over the years according to the model evaluation indicators.Considering the spatial heterogeneity of the incidence of JE,this study has used the geographically weighted regression model to further explore the variability of influencing factors in different spaces.Results 1.The reported incidence of JE in Gansu province from 2005 to 2018 was between 0.02 and 1.91 per 100,000.Except the reported incidence of 0.61 per100,000 in 2006,the reported incidence of 2005-2015 was below 0.25 per 100,000.It has a decreasing trend year on year and is relatively stable.The reported incidence rate in 2012 was the lowest during the study period.During 2016-2018,the reported incidence rate rose from 0.32 to 1.91 per 100,000.2.In terms of spatial distribution,JE cases were mainly concentrated in the southeastern of Gansu province,including some counties in Longnan,Tianshui,Pingliang and Qingyang city.During the study period,those counties with a larger number of cases were Qinzhou district(150 cases),Wudu district(120 cases),Qin’an county(110 cases),Li county(103 cases)and Xihe county(93 cases).3.The scan results were basically consistent with the spatial distribution of the incidence of JE.The clustering areas in Gansu province were dominated by some counties in Longnan,Tianshui and Pingliang city in each year.The first-tier clustering area covered about 14 counties with its scanning radius ranging from 174km～256km.The risk of JE in the clustering area was more than 2.9 times higher than that outside the clustering area.The second-tier clustering area covered about 5 counties,and its scanning radius was between 75 km and 201 km.The risk of JE in the clustering area was more than 2.5 times higher than that outside the clustering area.The third-tier clustering area only covered Chongxin and Lingtai county with its scanning radius of41.85 km.The risk of JE in this cluster was 3.96 times higher than that outside the clustering area.4.The results of the global regression model showed that the fitting effect of the spatial lag model of the incidence of JE in Gansu province in 2006,2010,2011,and2018 was better than that of the ordinary least squares regression model.The spatial error model of the incidence of JE in Gansu province in 2007 and 2013 was better than that of the ordinary least squares regression model.The residuals of the least squares regression model in other years were independent,which was suitable to fit the ordinary least squares regression model.Geographically weighted regression model was better than ordinary least squares regression models.5.The analysis results of the influencing factors of the spatial regression model showed that the incidence of JE in Gansu province in 2005,2015 and 2016 was positively correlated with the annual average temperature(>0,P<0.05).In 2005,2007,2008,2009,2014 and 2016,the incidence of JE was inversely correlated with the duration of sunshine(<0,P<0.05).The incidence of JE in 2007,2008,2009,2013,and2014 was inversely correlated with the annual relative humidity(<0,P<0.05).The incidence of JE in 2009,2013 and 2018 was positively correlated with the average annual air pressure( >0,P<0.05).The incidence of JE in 2013 was positively correlated with the average annual precipitation(>0,P<0.05),the incidence of JE in2018 was inversely correlated with the average annual precipitation(<0,P<0.05).Correlations between the incidence of JE in each year and GDP per capita and the sown area of crops were not statistically significant.Conclusions 1.The reported incidence of JE in Gansu Province from 2005 to2015 was low and stable,but the reported incidence increased significantly from 2016 to 2018.The cases were mainly concentrated in the southeastern of Gansu province,including some counties in Longnan,Tianshui,Pingliang and Qingyang city.From2016 to 2018,the coverage of JE cases began to spread from the southeast of Gansu province to the central region,and the areas of high incidence and case reporting areas gradually expanded.The spatial scanning results were basically consistent with the spatial distribution of the incidence of JE.2.We found that there was spatial autocorrelation in the model residuals in 2006,2007,2010,2011,2013 and 2018,which was suitable to fit spatial regression model.The geographically weighted regression model was further optimized on the basis of the global spatial regression model,reflecting the difference of parameter estimates in different counties.3.The analysis of influencing factors in different years were different.The reported incidence of JE in Gansu province had correlation with temperature,air pressure,precipitation,duration of sunshine and relative humidity.Each influencing factor had different effects on the reported incidence of JE in different counties,which showed spatial variability. |
PDF Full Text Request