The Applicational Study Of Bayesian Spatiotemporal Modeling In The Risk Of Smear Positive Pulmonary Tuberculosis In Xinjiang

Objective: Based on the epidemiological data of smear tuberculosis in 98 districts and counties of Xinjiang from 2011 to 2015,spatial epidemiological methods and Bayesian spatiotemporal regression models were used to analyze the temporal and spatial distribution characteristics of tuberculosis risk,as well as ecological population,economic and meteorological environment.The effect of factors on disease risk is expected to provide information and methodological reference for tuberculosis prevention and control.Methods: Collect data on tuberculosis epidemic from 2011 to 2015,as well as sociodemographic,economic,and meteorological data,describe the temporal and spatial distribution characteristics of smear-positive tuberculosis standardized risk ratio(SMR);use factor analysis to reduce many ecological indicators,based on Geographically weighted regression(GWR)and Bayesian spatiotemporal regression model to analyze the spatial and temporal correlation effects of ecological factors on SMR.Results:(1)From2011 to 2015,a total of 57,477 cases of smear-positive pulmonary tuberculosis were reported in Xinjiang.The annual average reporting rate was 47.87/100,000.The SMR distribution in southern Xinjiang was higher than that in east and northern.The high-risk areas were mainly concentrated in Yingjisha County,Zepu County and Awati County which belong Kashgar,Hotan and Aksu.(2)factor analysis constructed a total of five comprehensive ecological indicators of social population factors,economic factors,social welfare factors,rain factors and temperature factors.(3)The SMR spatial autocorrelation results show that the Moran’s I index is between 0.2341 and 0.4382(P<0.05)from 2011 to2015,and there is spatial clustering.(4)The results of GWR model constructed by SMR and ecological indicators show that there are regional differences in the effects of different ecological factors in different regions.The sociodemographic factors have a greater impact on the central region of Xinjiang.The economic factors mainly affect the northern Xinjiang region.Other ecological factors mainly affect the southern Xinjiang region.(5)Inthe Bayesian space-time modeling,the Bayesian model including various ecological factors and space-time interaction effects is the optimal model;the optimal model results show that there is a spatio-temporal effect on the risk of smear-positive tuberculosis in Xinjiang;High(that is,the proportion of agricultural population is high,the population is high,the proportion of ethnic minorities is high),the risk of tuberculosis is increasing;the higher the economic factor score(that is,the higher the economic level),the risk of disease is decreasing;the more the rain factor is scored High(ie,more precipitation and higher humidity),the risk is decreasing;the number of beds/1000 people is negatively correlated with SMR,suggesting that the risk of disease is low in areas with abundant medical resources.Conclusion: There is spatial clustering and regional differences in the risk of smear-positive tuberculosis in Xinjiang.The effects of ecological factors on the risk of tuberculosis are also different at different times and regions.Therefore,combined with the characteristics of the social environment in each region,targeted adjustments to the prevention and control strategies and means of tuberculosis have practical significance for the precise control of tuberculosis epidemic.