Temporal-spatial distribution characteristics and its influencing factors of tuberculosis epidemic in the Chinese mainland from 2014 to 2018

Objective The number of beds per thousand, gross domestic product (GDP) per capita, population density and annual sunshine hours were investigated to analyze the spatio-temporal characteristics of tuberculosis incidence of 31 provinces in the Chinese mainland from 2014 to 2018, so as to provide a theoretical basis for the prevention and control of the spread of tuberculosis in the future. Methods The spatial autocorrelation analysis was performed to explore the spatio-temporal agglomeration of the incidence of tuberculosis in the Chinese mainland in each year. A geographically and temporally weighted Logistic regression analysis model was established to reveal the spatio-temporal non-stationarity of tuberculosis incidence, and to compare the fitting results with the classic Logistic regression analysis model. Results The global spatial Moran's I index of tuberculosis incidence in the Chinese mainland from 2014 to 2018 were 0.43, 0.44, 0.47, 0.49, 0.37, showing a certain spatiotemporal correlation. The coefficient function of the geographically and temporally weighted Logistic regression analysis model fluctuated in the interval (-1, 1), and could reflect the spatio-temporal heterogeneity of tuberculosis well. The confidence interval of each coefficient function did not contain 0, which had spatio-temporal non-stationarity. The results of the goodness of fit test were better than the classic Logistic regression analysis model. Conclusions The incidence of tuberculosis in 31 provinces in the Chinese mainland was closely related to four macro factors, and appropriate strategies should be formulated for tuberculosis prevention and control according to the difference of regional social macro factors.