Effect modification of the association between short-term meteorological factors and mortality by urban heat islands in Hong Kong

Abstract

Background: Prior studies from around the world have indicated that very high temperatures tend to increase summertime mortality. However possible effect modification by urban micro heat islands has only been examined by a few studies in North America and Europe. This study examined whether daily mortality in micro heat island areas of Hong Kong was more sensitive to short term changes in meteorological conditions than in other areas. Method: An urban heat island index (UHII) was calculated for each of Hong Kong's 248 geographical tertiary planning units (TPU). Daily counts of all natural deaths among Hong Kong residents were stratified according to whether the place of residence of the decedent was in a TPU with high (above the median) or low UHII. Poisson Generalized Additive Models (GAMs) were used to estimate the association between meteorological variables and mortality while adjusting for trend, seasonality, pollutants and flu epidemics. Analyses were restricted to the hot season (June-September). Results: Mean temperatures (lags 0-4) above 29°C and low mean wind speeds (lags 0-4) were significantly associated with higher daily mortality and these associations were stronger in areas with high UHII. A 1°C rise above 29°C was associated with a 4.1% (95% confidence interval (CI): 0.7%, 7.6%) increase in natural mortality in areas with high UHII but only a 0.7% (95% CI: -2.4%, 3.9%) increase in low UHII areas. Lower mean wind speeds (5th percentile vs. 95th percentile) were associated with a 5.7% (95% CI: 2.7, 8.9) mortality increase in high UHII areas vs. a -0.3% (95% CI: -3.2%, 2.6%) change in low UHII areas. Conclusion: The results suggest that urban micro heat islands exacerbate the negative health consequences of high temperatures and low wind speeds. Urban planning measures designed to mitigate heat island effects may lessen the health effects of unfavorable summertime meteorological conditions. © 2012 Goggins et al.