Daily ambient temperature and mortality in Thailand: Estimated effects, attributable risks, and effect modifications by greenness

Abstract

Background: In recent years, many previous studies have examined the association between ambient temperature and mortality in different parts of the world. However, very few studies have explored the mortality burden attributable to temperature, especially those in developing countries. This study aimed to quantify the burden of mortality attributable to non-optimum temperature in Thailand and explore whether greenness, using normalized difference vegetation index (NDVI) as indicator, alleviates the mortality contributed by non-optimum ambient temperature. Methods: Daily number of mortality (i.e., all-cause, cardiovascular and respiratory diseases) and daily meteorological data were obtained over 65 provinces in Thailand during 2010 to 2017. The two-stage statistical approach was applied to estimate the association between temperature and mortality. First, the time-stratified case-crossover analysis was performed to examine province-specific temperature-mortality association. Second, province-specific association was pooled to derive national estimates using multivariate meta-regression. Mortality burden attributable to temperature was then estimated, and the association between attributed mortality and NDVI was explored using multivariate meta-regression models. Results: A total of 2,891,407 all-cause of death was included over the study period, in which 403,450 and 264,672 deaths were accounted for cardiovascular and respiratory diseases, respectively. The temperature-mortality association at cumulative lag 0–7 days was non-linear with J-shaped curve for all-cause and respiratory mortality, whereas V-shaped curve was observed for cardiovascular mortality. Using minimum mortality temperature (MMT) as optimum temperature, 3.72% (95% empirical CI: 2.18, 5.21) of all-cause, 2.92% (0.55, 5.10) of cardiovascular and 3.00% (0.27, 5.49) of respiratory mortality were attributable to non-optimum temperature (both hot and cold effects). Higher level of NDVI was associated with alleviated impacts of non-optimum temperature, especially hot temperature. Conclusion: Exposure to non-optimum temperature was associated with increased risks of mortality in Thailand. This finding is useful for planning the public health interventions to reduce health effects of non-optimum ambient temperature.