[Effect of City-Specific Characteristics on Association between Heat and Ambulance Dispatches]

In this multicity study, we aimed to elucidate the city-specific factors affecting the association of high ambient temperature with ambulance dispatches due to acute illnesses.

We used the data of ambulance dispatches in 27 cities in Japan with more than 500,000 population excluding Tokyo, from May to September from 2012 to 2015. We included patients 20 years and older (≥20 years) and stratified them into three age groups (20-59, 60-79, and ≥80 years). We explored the city-specific pattern of the daily relative temperature (in temperature percentiles) and the risk of ambulance dispatches for each age group using a distributed lag nonlinear model and estimated the city-specific relative risks of ambulance dispatches at the 95/99 percentile temperature compared with the 77.6 percentile temperature defined as the reference temperature (T_ref). Then, the estimates were combined by performing meta-analyses for each age group. We also applied meta-regression models to explore whether the city-specific characteristics modified the association of temperature with ambulance dispatches.

The relative risks of the 95th percentile with respect to T_ref were 1.14 (95% confidence interval (CI): 1.12, 1.16), 1.16 (95% CI: 1.13, 1.20), 1.13 (95% CI: 1.10, 1.16), and 1.13 (95% CI: 1.00, 1.16), for all-age (≥20) and age-stratified groups (20-59, 60-79, and ≥80 years), respectively. We observed a higher relative risk for the ≥20 years age group in the cities with higher proportions of single-elderly, single-mother, and single-father households. We also found that the relative risk for the 20-59 years age group was higher in the cities with a higher proportion of blue-collar workers.

The present study provides insights into city-specific characteristics modifying heat-related health effects.