The exposure-response association between humidex and bacillary dysentery: A two-stage time series analysis of 316 cities in mainland China.

Abstract

BACKGROUND\nMany studies have reported the interactive effects between relative humidity and temperature on infectious diseases. However, evidence regarding the combined effects of relative humidity and temperature on bacillary dysentery (BD) is limited, especially for large-scale studies. To address this research need, humidex was utilized as a comprehensive index of relative humidity and temperature. We aimed to estimate the effect of humidex on BD across mainland China, evaluate its heterogeneity, and identify potential effect modifiers.\n\n\nMETHODS\nDaily meteorological and BD surveillance data from 2014 to 2016 were obtained for 316 prefecture-level cities in mainland China. Humidex was calculated on the basis of relative humidity and temperature. A multicity, two-stage time series analysis was then performed. In the first stage, a common distributed lag non-linear model (DLNM) was established to obtain city-specific estimates. In the second stage, a multivariate meta-analysis was conducted to pool these estimates, assess the significance of heterogeneity, and explore potential effect modifiers.\n\n\nRESULTS\nThe pooled cumulative estimates showed that humidex could promote the transmission of BD. The exposure-response relationship was nearly linear, with a maximum cumulative relative risk (RR) of 1.45 [95% confidence interval (CI): 1.29-1.63] at a humidex value of 40.94. High humidex had an acute adverse effect on BD. The humidex-BD relationship could be modified by latitude, urbanization rate, the natural growth rate of population, and the number of primary school students per thousand persons.\n\n\nCONCLUSIONS\nHigh humidex could increase the risk of BD incidence. Thus, it is suitable to incorporate humidex as a predictor into the early warning system of BD and to inform the general public in advance to be cautious when humidex is high. This is especially true for regions with higher latitude, higher urbanization rates, lower natural growth rates of population, and lower numbers of primary school students per thousand persons.