The influence of meteorological factors on tuberculosis incidence in Southwest China from 2006 to 2015

Abstract

The influence of meteorological determinants on tuberculosis (TB) incidence remains severely under-discussed, especially through the perspective of time series analysis. In the current study, we used a distributed lag nonlinear model (DLNM) to analyze a 10-year series of consecutive surveillance data. We found that, after effectively controlling for autocorrelation, the changes in meteorological factors related to temperature, humidity, wind and sunshine were significantly associated with subsequent fluctuations in TB incidence: average temperature was inversely associated with TB incidence at a lag period of 2 months; total precipitation and minimum relative humidity were also inversely associated with TB incidence at lag periods of 3 and 4 months, respectively; average wind velocity and total sunshine hours exhibited an instant rather than lagged influence on TB incidence. Our study results suggest that preceding meteorological factors may have a noticeable effect on future TB incidence; informed prevention and preparedness measures for TB can therefore be constructed on the basis of meteorological variations.

Figure 1

Geographical location of study site (Created by ArcMap 10.2).

Figure 2

Monthly reported TB incidence in Jinghong, China, 2006–2015.

Figure 3

Monthly data on meteorological factors in Jinghong, China, 2006–2015.

Figure 4

Autocorrelation function (ACF) graph of differenced TB incidence sequence.

Figure 5

Distributed lagged nonlinear associations between meteorological factors and TB incidence. A represents minimum temperature (every 1 °C increase), B represents maximum temperature (every 1 °C increase), C represents average temperature (every 1 °C increase), D represents total precipitation (every 1 °C increase), E represents maximum precipitation (every 1 centimeter increase), F represents minimum relative humidity (every 1 percent increase), G represents average wind velocity (every 1 meter/second increase), H represents total sunshine hours (every 40 hours increase). The error bounds reflect 95% CIs.