Abstract
Few studies have examined individual socioeconomic status (SES) as a potential modifier of ambient temperature-health associations, especially for temperature-related hospitalizations. We fit penalized distributed lag non-linear models within generalized additive models to study the short-term associations (0-3 days) between temperature and hospital admissions stratified by common causes, age, and individual SES, as determined by whether patients received public assistance (PA) to cover their medical fee at the time of hospitalizations, during the hot season (May 15 to October 15) in Hong Kong for the years 2010-2019. We calculated the ratio of relative risk (RRR) and corresponding 95% confidence interval (CI) to statistically test the difference of the associations between PA groups. For 75 + patients, the PA group had significantly increased risks of hospitalizations at higher temperature for most causes, with relative risks (RR, 99th %ile vs. 25%ile) and 95% CIs of 1.138 (1.099, 1.179), 1.057 (1.008, 1.109), and 1.163 (1.094, 1.236) estimated for all non-cancer non-external, circulatory, and respiratory admissions, respectively. There were slight decreases of RRs with higher temperature for 75 + patients without PA. The strengths of temperature-hospitalization associations were strongly and significantly different between PA groups for all examined causes for 75 + patients, with the most considerable discrepancy found for ischemic heart disease (RRR = 1.266; 95% CI, 1.137, 1.410). Hospitalizations for patients aged 15-74 were less affected by heat, and the difference of the associations between groups was small. Individual SES is a significant modifier of high temperature-hospitalization associations in Hong Kong among the elderly. Public health interventions are needed to better protect this subpopulation from adverse health impacts of high temperature.