Abstract:
The impact of ambient heat in summer on human health is an issue of increasing concern, especially in light of climate change and the increased frequency and intensity of heat events. Previous studies have shown that mortality increased during heat events and is positively associated with ambient heat. However, the association with morbidity has seldom been examined. The goal of this study was to elucidate the regional difference in the strength of association.
Potential health impacts of heat on multiple morbidity indicators (hospital admissions, length of stay, charge per day and total charges) were investigated in four Washington State population centers: Greater Seattle (King, Pierce, Snohomish counties), Spokane County, Tri-Cities (Benton, Franklin counties) and Yakima County. summer (June to August) of 2004 was selected as a study period. Morbidity indicators for each hospital admission were matched to daily maximum humidex through residential zip codes. Day-of-the-week and holiday effects were eliminated by averaging daily values of morbidity indicators with today and six more days' values and by excluding days containing data of July 5 (federal holiday for Independence Day).
More than half of the associations formed a U-shaped curve, suggesting the existence of humidex threshold. Whereas the threshold was observed at 30oC in Spokane, Tri-Cities and Yakima, such a threshold was not observed in Greater Seattle. More than half of the modes of association were different between June and July-August, suggesting the existence of period-specific factors, such as seasonal diseases, that influenced the association. In particular, overall hospital admissions were higher in June than in July-August in Greater Seattle, Spokane and Tri-Cities, suggesting an early summer effect of heat-related morbidity.
Humidex was positively associated with hospital admissions in all study areas. Increases in hospital admissions with 1oC humidex increase in Spokane (0.9%) and Yakima (0.9%) were higher than in Greater Seattle (0.3%). This was consistent with the regional variation of socioeconomic indicators, such as unemployment rate and median household income. Conversely, humidex was not positively associated with mean length of hospital stay and mean charge per day in any of the study areas, since total charges were positively associated with humidex in some study areas and negatively associated in other study areas.
Although the study period is only one year and these results may not be generalized, this study elucidated the regional variation of threshold and strength of association to ambient heat.