Effect Modification By Prefecture Characteristics On The Decline Of Heat-Related Mortality In Japan, 1972 To 2010

Abstract

Introduction: The declining impact of heat on mortality has been reported previously, but the factors affecting this phenomenon vary by local environment. We examined whether the decline of heat-related mortality in Japan can be explained by the variation of prefecture-specific characteristics over time such as local temperature, prevalence of air conditioning, economic conditions, and health care coverage. Methods: We analyzed daily heat-related all-cause non-trauma mortality and weather data for the summers (June to August) of 1972 to 2010 in 47 prefectures. The trends of heat-related mortality in each prefecture were estimated using generalized linear model. Random-effects meta-regression was used to investigate the relationship between prefecture characteristics over time and the temporal changes of heat-related mortality across prefectures. Results: The daily number of excess heat-related mortality reduced from 27 (95% confidence interval [CI] = 25–30) to 4 per 1,000 deaths (95% CI = 3–6) in Japan over the 39-year period. Higher local average temperature was associated with larger temporal decline in heat-related mortality. This decline was also seen in prefectures with larger increase in population, average savings, and the improvement of fiscal health. The temporal variations for the prevalences of air conditioning and the number of nurses (per 10,000 population) were inversely associated with the trends of heat-related mortality. Conclusion: Excess mortality due to summer heat has declined over the years. Prefectures that experienced large population growth and improved economic condition appeared better at reducing mortality attributable to heat. Further study is needed to understand accessibility to health care and usage patterns of air conditioning to explain the inverse association observed in this study.