Application of spatial lag model in ecological study on influencing factors of human immunodeficiency virus/acquired immune deficiency syndrome reporting rate in Xinjiang
Objective To analyze the sociodemographic factors affecting the human immunodeficiency virus/acquired immunedeficiency syndrome(HIV/AIDS)reporting rate with spatial lag model,and to explore the application value of this model in HIV/AIDS epidemiology.Methods Geoda software was used for the spatial regression analysis of HIV/AIDS report rate and social demographic factors in 85 counties in Xinjiang during 2007 to 2015.Results Ordinary least square(OLS)showed the spatial autocorrelation of regression residuals(Moran's I=0.286 7,P＜0.001),and the spatial lag model(SLM)was selected based on lagrange multiplier(LM)tests.The SLM showed that the spatial lag terms were statistically significant(P＜0.001),which means there was a positive spatial autocorrelation of HIV/AIDS reporting rate."Population"was positively correlated to the HIV/AIDS reporting rate,while"proportion of agricultural population"was negatively correlated.Compared to traditional OLS model,the SLM had better fitting effect with increased adjusted R2 and decreased logarithmic likelihood ratio and Akaike information criterion(AIC)value.Conclusions HIV/AIDS reporting rate was higher in areas with large population and high proportion of urbanization.The SLM is valuable to popularize in ecological study on influencing factors of HIV/AIDS.
Human immunodeficiency virusAcquired immunodeficiency syndromeSpatial lag modelXinjiang
万方数据
维普
