Abstract
BACKGROUND: While temperature extremes have been shown to be associated with an increased risk of hospital admissions, evidence of their impact on the length of hospital stay, which may capture the lingering effects of temperature extremes, is scarce. OBJECTIVES: We aimed to evaluate the association between daily variation in ambient temperature and daily variation in daily total length of stay (daily TLOS), a composite measure encompassing the daily count of hospital admissions and their corresponding length of hospital stay among cardiopulmonary patients. Additionally, we quantified the burden of TLOS attributable to non-optimal temperatures among Hong Kong's older adult population. METHODS: We used a generalized linear regression with a distributed lag non-linear model to estimate the association between ambient temperature and daily TLOS. The analysis used 13 years of time-series data (1998-2010) on daily temperature and hospital admissions for cardiopulmonary diseases through accident and emergency departments among Hong Kong's older adult population. We quantified the attributable risk of TLOS by calculating the temperature-related days of hospital stay and the attributable fraction (AF). RESULTS: We recorded a total of 4,095,722 hospital stay days for cardiovascular patients and 4,492,697 days for respiratory patients. We found that both cold and heat were associated with increased TLOS for cardiopulmonary disease. The temperature-related AF was 11.5% (95% empirical CI: 5.3-17.2%) for cardiovascular disease, corresponding to an annual increase of 36,174 days (95% empirical CI: 15,286-57,018). For respiratory disease, the AF was 10.7% (95% empirical CI: 7.1-13.9%), equating to an annual increase of 36,897 days (95% empirical CI: 24,949-49,024) days annually. CONCLUSION: Extreme temperatures were associated with increased TLOS for cardiopulmonary patients in Hong Kong's older adult population. These findings highlight the need for hospitals to prepare in advance for extreme temperature events by implementing specific measures in terms of human resources and medical resources. In addition, the results provide valuable scientific evidence to support public health policies and inform hospital planning and management.