Spatiotemporal or temporal index to assess the association between temperature variability and mortality in China?

Kejia Hu; Shanshan Li; Jieming Zhong; Xuchao Yang; Fangrong Fei; Feng Chen; Qi Zhao; Yunquan Zhang; Gongbo Chen; Qian Chen; Tingting Ye; Yuming Guo; Jiaguo Qi

doi:10.1016/j.envres.2018.12.037

Spatiotemporal or temporal index to assess the association between temperature variability and mortality in China?

Environ Res. 2019 Mar:170:344-350. doi: 10.1016/j.envres.2018.12.037. Epub 2018 Dec 19.

Authors

Kejia Hu¹, Shanshan Li², Jieming Zhong³, Xuchao Yang⁴, Fangrong Fei⁵, Feng Chen⁶, Qi Zhao², Yunquan Zhang⁷, Gongbo Chen², Qian Chen⁸, Tingting Ye⁸, Yuming Guo², Jiaguo Qi⁹

Affiliations

¹ Ocean College, Zhejiang University, Zhoushan 316021, China; Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Melbourne 3004, Australia.
² Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Melbourne 3004, Australia.
³ Zhejiang Provincial Center for Disease Control and Prevention, Hangzhou 310051, China. Electronic address: jmzhong@cdc.zj.cn.
⁴ Ocean College, Zhejiang University, Zhoushan 316021, China; Center for Global Change and Earth Observations, Michigan State University, East Lansing 48823, USA. Electronic address: yangxuchao@zju.edu.cn.
⁵ Zhejiang Provincial Center for Disease Control and Prevention, Hangzhou 310051, China.
⁶ Zhejiang Institute of Meteorological Sciences, Hangzhou 310008, China.
⁷ Department of Epidemiology and Biostatistics, School of Health Sciences, Wuhan University, Wuhan 430071, China; Hubei Province Key Laboratory of Occupational Hazard Identification and Control, School of Public Health, Wuhan University of Science and Technology, Wuhan 430065, China.
⁸ Ocean College, Zhejiang University, Zhoushan 316021, China.
⁹ Center for Global Change and Earth Observations, Michigan State University, East Lansing 48823, USA.

PMID: 30623880
DOI: 10.1016/j.envres.2018.12.037

Abstract

Epidemiological studies increasingly provide evidence about the adverse health effects of temperature variability (TV), which reflects short-term intra- and inter-day temperature change. However, calculation of TV only considers the temporal variability and lacks spatial variability. This study intends to investigate whether the lack of spatial variability in TV calculations has biased the health effect estimates. We collected daily data from the fine-gridded hourly temperatures and more than 2 million all-cause mortality counts in Zhejiang province in China from 2009 to 2015. A spatiotemporal TV index was developed by calculating the standard deviation of the hourly temperatures based on records from multiple sites. This new index could be compared to the two typical temporal TV indices that are calculated based on the hourly temperatures from one-site and area-average records. The three types of TV are compared using a three-stage analytical approach: district-specific time series Poisson regression, meta-analysis, and calculation of attributable mortality fraction. We observe that both spatiotemporal and temporal TVs produce very similar TV-mortality associations, attributable mortality fractions, and model fits at the district level. For example, the mortality increase associated for every increase of 1 °C during 0-7 exposure days is 1.53% (95% CI: 1.31, 1.73) in spatiotemporal TV, whereas it is 1.48% (95% CI: 1.27, 1.68) and 1.45% (95% CI: 1.24, 1.67) in the one-site and area-average temporal TV, respectively. Thus, time series models using temporal TV index are equally good at estimating the associations between TV and mortality as spatiotemporal TV at the district level in population-based epidemiological studies in China. Epidemiological studies using temperature from one site or the averages of multiple sites in TV calculation will not bias the effect estimates of TV. Our study could provide an important guidance method for future TV-related research in China and even in other countries.

Keywords: China; Mortality; Risk; Temperature change; Temperature variability.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

China
Data Collection
Environmental Exposure*
Mortality*
Seasons
Temperature*