Spatial-temporal analysis of tuberculosis in the geriatric population of China: An analysis based on the Bayesian conditional autoregressive model

Abstract

Background: Tuberculosis (TB) remains a clinical and epidemiological challenge in the geriatric population. We aim to examine the spatial-temporal pattern of TB in the geriatric population and its relationship with meteorological & sociodemographic factors using the Bayesian conditional autoregressive (CAR) model.

Method: An ecological design was used in the geriatric (age > = 65 years) population from 2005 to 2015. Spatial autocorrelation and hot spots were explored using geographical information system (GIS) statistics. The Bayesian CAR model was used for modeling TB to estimate the parameters using the WinBUGS software. Deviance information criteria (DIC) were used to select the best performing model.

Results: Spatially, TB was clustered in Central China and southeast of China. Temporally, an increasing trend and high peak of TB was detected during the spring. TB was significantly associated with air temperature at the posterior mean: -0.165 (95%CI: -0.235, -0.108), and it was negatively associated with average wind speed: -0.028 (95%CI: -0.043, -0.018) and positively associated with rainfall: 0.095 (95%CI: 0.045, 0.163). TB was significantly and positively associated with population density: 0.088(95%CI: 0.031, 0.129) and sex ratio (M: F): 0.162 (95%CI: 0.091, 0.284) and was negatively related with gross domestic product (GDP): -0.046(95%CI: -0.156, -0.037). Out of 31 provinces, 17 provinces had a higher risk for TB.

Conclusion: TB shows a clear spatial and seasonal variation; it is geographically aggregated, and more men are affected than women. Areas with an underprivileged economy, high population density, high rainfall, low wind speed, and low temperature have a higher risk for TB.

Keywords: Bayesian; Elderly; Spatial; Spatial-temporal; TB.