| |
|---|---|
 (819KB)
|
4,004,000Yen (FY1998; 2,004,000Yen)
Heat stroke, Air Pollutants, Heat Index, Generalized Linear Regression Models
The daily numbers of heat stroke emergency transport cases/million residents for Tokyo were obtained for males and females in three age groups, 0-14, 1 5-64 and 65+, for the months of July and August, 1980-1995. The daily numbers of heat stoke emergency transport cases/million residents for each gender and age group were merged with daily average heat index (HI) and concentrations of NO2 and 03. Lag times of O (same day) to 4 days in these variables were incorporated into the model as additional risk factors . Generalized linear models (GLM'S) assuming a Poisson error structure, were used to determine if HI, concentrations of NO2 and 03, age, gender and interactions among all of these variables were significant risk factors for heat stroke. To account for correlations among observed daily heat stroke cases from the same sub population group, GLM's were fit using the generalized estimating equation (GEE) method. Responses between males and females and responses by age groups were compared. Same day, or zero lag time HI and concentrations of NO2 and 03 were all significant risk factors for heat stroke in all age groups of males and females. A 1-day lag time for HI and NO2 concentrations and a 2-day lag time for 03 concentrations were also significant risk factors for all age groups of males and females. Interaction terms for 0- and 1-day lag times in HI and gender in all age groups, and 0-lag time HI and 0-lag time 03 concentrations and gender in all age groups were additional significant risk factors for heat stroke. The number of heat stroke emergency transport cases/million residents was greater in males than in females in the same age groups. The smallest number of heat stroke emergency transport cases/million residents occurred for 0-14 males and females and the greatest number of heat stroke emergency transport cases/million residents occurred for 65+ males and females. To evaluate the health effects caused by heat stress in summer, regression analysis was carried out between emergency trasport data, climatic data, and air pollution. Generalized linear regression models were used to examine heat stroke as functions of average daily temperature, relative humidity, heat index (HI) temperature (a combination of daily temperature and relative humidity) and air pollution concentrations. Best models were obtained when lag time between when exposure occured and when emergency transport was reported, was either zero or 1 day and when HI temperature was used. Using regression models that contained HI temperature and concentrationd of NO2 and 03, for males and females in all age groups, log likelihood tests indicated that HI temperature was a significant factor for heat stroke. For all males and females with a lag time of one day, NO2 was a significant contributing factor for heat stroke. For all females with a zero lag time, 03 was a significant contributing factor for heat stroke.