Abstract:
Background Heatwaves are a lethal natural disaster in many countries, posing a big threat to population health. However, there is no international consensus on how to properly define a heatwave.
Objective This study aims to understand the impact of heatwave definition on the assessment of the health effects of heatwaves, and to provide evidence for the development of heatwave early warning systems.
Methods Data on climatic variables and cause-specific deaths in 60 provinces of Thailand from 1st January 1999 to 31st December 2008 were collected. Climatic variables included daily maximum temperature, minimum temperature, mean temperature, and relative humidity. Deaths included total deaths and deaths from ischemic heart disease, pneumonia, and diabetes. Cause of death was classified according to the International Classification of Diseases, Tenth Version (ICD10). Quasi-Poisson regression combined with distributed lag non-linear model was used to assess the short-term effect of heatwaves on deaths in Thailand, and compare the risks of total and causespecific deaths under different heatwave definitions (intensity and duration). The cut-off percentiles used for heatwave definitions were the 90th, 95th, 97th, and 98th percentiles of daily mean temperature; the durations were ≥ 2 d, ≥ 3 d, and ≥ 4 d, respectively. The total effects of heatwaves on cause-specific deathsrelative risks (RR) and corresponding 95% confidence intervals (CI) were estimated by random-effect meta-analysis that combined the results of the 60 provinces.
Results The risks of total and cause-specific deaths increased significantly during heatwave periods, compared with non-heatwave periods. During mild heatwaves (> the 90th percentile of mean temperature and ≥ 2 d), the risk of total deaths increased by 12.8% (95% CI:10.6%-15.1%). The RR of deaths during heatwaves did not consistently increase with the increase of heatwave intensity or duration. When the heatwave duration was ≥ 2 d, deaths due to pneumoniaRR:1.42, 95% CI:1.27-1.59 and deaths due to diabetes (RR:1.34, 95% CI:1.17-1.52) were most vulnerable to the heatwaves defined using the 97th percentile of mean temperature distribution as the threshold; total deaths (RR:1.16, 95% CI:1.13-1.20) as well as deaths due to ischemic heart disease (RR:1.35, 95% CI:1.24-1.48) were most vulnerable to the heatwaves defined using the 95th percentile of mean temperature distribution as the threshold. When the heatwave cut-off percentile was the 98th percentile, with the extension of heatwave duration, the risk of deaths from diabetes increased, but the risks of total deaths and deaths due to pneumonia decreased.
Conclusion This study suggests that the impact of heatwaves on death risks changes under different intensity and duration definitions. Therefore, in the development of heatwave warning systems and other heat adaptation strategies, incorporating the characteristics of heatwave vulnerability of local population and adopting an optimal local heatwave definition may more effectively protect public health in the face of heatwaves.