International Scholarly Research Notices

Research Article

An Exploration of Temperature Metrics for Further Developing the Heat-Health Weather Warning System in Hong Kong

Table 1

Recent studies of temperature and mortality relationship in Hong Kong.


Reference	Study period	Exposure	Outcome	Methods	Summary of findings

Yan [1]	1995–2005	Daily maximum, minimum, dew point temperature; barometric pressure, cloud cover, and wind speed	Mortality from all-cause, CIR, CVD, and IHD	Generalized additive models (GAM) using a cubic smoothing spline	Temperature was found to be a major factor causing deaths from CIR and RD, but not mortality from all-cause. The elderly (age 65) had longer time lags for mortality compared with the younger age groups (age 65).

Yan [2]	1980–2005	Seasonal analysis	Mortality from all-cause, NEO, CIR, and RD	Cosinor analysis	Statistically significant seasonality was discovered for total and gender-specific deaths from all causes, CIR, and RD but not for NEO.

Yan [3]	1980–1994	Maximum and minimum temperature, and seasonal analysis	Mortality from all-cause, NEO, CIR, and RD	Regression analyses and ANOVA	Mortality is higher in winter months (November to February), and it is lower in summer months (June to August).

Leung et al. [4]	1995–2004	Max NET, Min NET	Mortality from all-cause, NEO CIR, RD, IHD, CBD, PIF,COP, thermal stress from excessive heat and cold	Poisson regression	Heat threshold: daily max. NET ≥ 26°C (heat stroke risk double per unit rise beyond 26°C; cold threshold: daily min. NET ≤ 14°C (hypothermia risk increases 1.3 fold per unit fall below 14°C).

Chan et al. [5]	1998–2006	Daily mean, dew point temperature, mean humidity, NO₂, SO₂, O₃, and PM₁₀	Nonaccidental mortality	Retrospective ecological study, using generalised additive (Poisson) models	Temperature threshold is 28.2°C (1.8% increase per 1°C temperature increase above this threshold).

Tam et al. [10]	1997–2002	DTR, NO₂, SO₂, PM₁₀, O₃	CVD, among persons aged ≥ 65 years	Generalized additive model to regress daily mortalities of the elderly	Significant associations between CVD and DTR at lag day 1 and at lag days 0–1 to 0–5 were discovered. The largest effect was at lag days 0–3 (RR = 1.017; 95% CI = 1.003–1.031).

Li and Chan [14]	1968–1995	Net effective temperature (NET)	Mortality from all-cause, excluding external causes	Correlation between WSI and daily mortality	There could be a relationship between low WSI and mortality rates in winter, no conclusive results are found for summer.

Xu [15]	1998–2009	Apparent temperature (AT), diurnal AT range	Mortality from all-cause, CVD and RD, among persons aged ≥ 65 years	A nested case—control approach	AT threshold: 20.8°C (cold days), 28.2°C (hot days). AT was not significantly associated with mortality on hot days. Effect of diurnal AT range on mortality was not observed.

 *CBD: cerebrovascular diseases; COP: chronic obstructive pulmonary disease; CIR: circulatory diseases; CVD: cardiovascular diseases; IHD: ischaemic heart diseases; NEO: Neoplasm; PIF: pneumonia and influenza; RD: respiratory diseases.