Study Of The Spatial Distribution Characteristics And Macroscopic Influencing Factors Of Poor Eyesight Among Primary And Secondary School Students

Objective This study based on the poor eyesight data from Chinese students’physique and health survey,established spatial regression model by analyzing the prevalent characteristics and spatial distribution of poor eyesight among Han students aged 7-18 from 2000 to 2014;this study explored the spatial clustering of poor eyesight and the application value of spatial regression model in screening macroscopic influencing factors of poor eyesight,providing theoretical basis for the scientific development of eyesight health management strategies and optimization of public resource allocation.Methods Collecting cross-sectional data of poor eyesight of Han students aged 7-18 in 2000,2005,2010 and 2014,and using software Excel 2019 and SPSS 19.0 for data entry,collation and analysis,to describe the prevalent characteristics of poor eyesight;using software Geoda to associate the detection rate of poor eyesight with the vector map of national administrative regions,and draw the statistical map of the detection rate of poor eyesight,to display the grade distribution of the mean annual detection rate of poor eyesight in different administrative regions and the dynamic trend with the year;adopting Moran’s I,Local Moran’s I and Moran scatter plot,and according to the geographical adjacent relationship of national administrative divisions,to explore the spatial clustering and correlation pattern;Stata 16.0 software was used to take the number of people with college education or above,per capita education years,per capita GDP,Gini coefficient,urban green space area,PM2.5,annual average temperature,annual average sunshine duration and urban population density as independent variables,and the detection rate of poor vision from 2000 to 2014 as dependent variables to construct a spatial regression model,and evaluate and compare the fitting effect of different models.Results1.The prevalent characteristics of the detection rate of poor eyesight among national primary and secondary school students1.1 From 2000 to 2014,the mean detection rate of poor eyesight in China was 52.49%with15 regions exceeding the mean level,and the top three regions were Zhejiang（66.24%）,Jiangsu（64.80%）and Shandong（63.49%）;While Xinjiang（43.71%）,Guizhou（39.74%）and Hainan（36.98%）ranked the bottom three among the 15 regions below the mean level.1.2 Compared with 2000,the detection rate of poor eyesight increased in all regions in 2014with imbalanced amplitude.And the top three regions of increasing amplitude were Tianjin（33.56%）,Inner Mongolia（29.74%）and Heilongjiang（27.42%）;The bottom three regions were Anhui（14.81%）,Qinghai（13.53%）and Xinjiang（8.05%）.In other areas,the amplitude ranged from 15.46%to 26.87%.1.3 In 2000,2005,2010 and 2014,the areas with high incidence of detection rate of poor eyesight were all including Shandong,Jiangsu,Shanghai and Zhejiang located in East China.2.Spatial correlation pattern of the detection rate of poor eyesight among national primary and secondary school students2.1 Results of global spatial autocorrelation analysis:in 2000,2005,2010 and 2014,the order of global Moran index is₂₀₁₄（0.070）>₂₀₀₀（0.064）>₂₀₁₀（0.038）>₂₀₀₅（0.018）;except for 2005,all other years were statistically significant（P<0.05）.It showed a downward trend first and then an upward trend.2.2 Results of local spatial autocorrelation analysis:in 2000,there were two spatial correlation patterns:high-high（Anhui,Jiangsu,Shandong,Shanghai,Zhejiang）and low-low（Guangxi）;In 2005,there were two spatial correlation patterns:high-high（Jiangsu,Shandong,Shanghai,Zhejiang）and low-high（Sichuan）.In 2010,there were two spatial correlation patterns:high-high（Jiangsu,Shandong,Shanghai,Zhejiang）and low-high（Tianjin）;In 2014,only high-high（Jiangsu,Shandong,Shanghai,Zhejiang）spatial correlation pattern existed,the high-high distribution mainly concentrated in Jiangsu,Shandong,Shanghai and Zhejiang,all tests were statistically significant（P<0.05）.3.Spatial regression model of poor eyesight among national primary and secondary school students3.1 In spatial lag model test,both the Wald test（x²==14.36,P<0.01）and LR test（x²==35.16,P<0.01）have statistical significance;In spatial error model test,both the Wald test（x²==39.42,P<0.01）and LR test（x²==37.33,P<0.01）have statistical significance;Hausman test（x²==8.70,P>0.1）has no statistical significance.3.2 Spatial Durbin model is better than OLS model（R²SDM=0.7442>R²OLS=0.6530）;Spatial Durbin model is better than spatial lag model and spatial error model（R²SDM=0.7442>R²SLM=0.6460>R²SEM=0.6090,Log-LSDM=195.8871>Log-LSLM=178.3073>Log-LSEM=178.2202）.3.3 Regression results showed that the spatial lag coefficient of the explained variables has statistical significance（ρ=-1.339,P<0.01）,this coefficient represents the spatial interaction term,that is,the influence of spatial factors on poor eyesight cannot be ignored.From the regression coefficients of explanatory variables,GDP per capita（β=0.083,P=0.000）,urban green space area（β=0.030,P=0.045）,PM 2.5（β=0.051,P=0.001）,number of people with college education or above（β=-0.021,P=0.004）and annual mean temperature（β=-0.059,P=0.005）,they all have statistical significance;Above are the main factors affecting the spatial and temporal pattern of poor eyesight among national primary and secondary school students.When GDP per capita and PM 2.5 increased by 1%,the detection rate of poor eyesight increased by 0.083%and 0.051%,respectively.When the number of people with college education and annual mean temperature increased by 1%,the detection rate of poor eyesight decreased by 0.021%and 0.059%,respectively.Conclusions1.From 2000 to 2014,the mean detection rate of poor eyesight in China was 52.49%with the most severe poor eyesight in Shandong,Jiangsu,Shanghai and Zhejiang.2.Compared with 2000,the detection rate of poor eyesight showed an increasing trend in2014 across the country,with the largest and smallest increase in Tianjin and Xinjiang respectively.3.The poor eyesight from 2000 to 2014 showed a clustered distribution,and the spatial correlation pattern included high-high,low-low and low-high;Among them,the high-high correlation pattern was mainly clustered in Jiangsu,Shandong,Shanghai and Zhejiang located in east China.4.The fitting degree of spatial Durbin model is better than that of OLS（ordinary least square）,SLM（spatial lag model）and SEM（spatial error model）.5.The direct effect of GDP per capita and PM 2.5 on poor eyesight in this region is positive,which will promote the development of poor eyesight to some extent;The direct effect of the number of people with college education or above and the annual mean temperature is negative,which will inhibit the development of poor eyesight to some extent;urban green space had little effect on poor eyesight.6.Macroscopic factors have obvious spatial spillover effect on poor eyesight in adjacent areas:the indirect effects of GDP per capita and Gini coefficient on the adjacent areas were negative,which inhibited the development of poor eyesight in the adjacent areas to some extent;urban green space had little effect on poor eyesight.