Student Research: Aubrey DeVine

, Environmental Health (EH), 2017
Faculty Advisor: Tania M Busch Isaksen

Spatial Analysis of Emergency Medical Service Calls and Extreme Heat in King County, WA


Background: Climate change is increasing allergens and environmental degradation, changing vector ecology, and intensifying severe weather and extreme heat (Center for Disease Control, 2016). Average global temperatures are steadily rising, along with the number and intensity of extreme heat events. Extreme heat has significant human health effects such as heat exhaustion, heat stress, and heatstroke. It is well known that morbidity, mortality, and hospitalizations increase during extreme heat events (Busch Isaksen et al., 2016; Medina-Ramon et al., 2006), but this research analyzed the relationship between extreme heat and Emergency Medical Service (EMS) calls. Objective: Building off the previous research of Calkins et al. (2016) who found increased relative risks for both Basic Life Support (BLS) calls and Advanced Life Support (ALS) calls on a King County-wide extreme heat day compared to a non-heat day, this study redefined the spatial scale of analysis to 4km x 7.5km grid cells. Additionally, it looked at county-wide temporal variations in EMS calls as well as examined the relationship of local extreme heat relative risk with community-level grid cell variables: percent tree canopy, percent impervious surfaces, total population, percent of population by age groups, percent development, percent water, average median income, and percent poverty. Methods: Using previously collected EMS data from Emergency Medical Technicians (EMT) of Seattle and King County Division of the Department of Public Health and meteorological data from the Climate Impacts Group, this study analyzed the spatial relationship between extreme heat and EMS calls in King County, WA. EMS data from the summer months of May 1st to September 30th, 2007 to 2012, was evaluated by calculating local crude relative risks for both BLS and ALS calls in each grid cell using a poisson regression and an a priori extreme heat threshold from Calkins et al. (2016). The more precise spatial scale in this analysis portrayed the spatial variation of heat risk in King County. Results: For both BLS and ALS calls, there were increased estimated relative risks of an EMS call on an extreme heat day compared to a non-heat day during peak humidex hours (3-6pm) and after peak humidex hours (6-10pm). Crude relative risks varied in grid cells across the county; significant increases in relative risk for BLS call volumes ranged from 1.05 (95% CI: 1.01, 1.09) to 4.13 (95% CI: 1.65, 10.32), while significant increases in relative risk for ALS call volumes ranged from 1.21 (95% CIs: 1.11, 1.32) to 6.50 (95% CIs: 1.06, 40.03). The results of regression models between relative risk and predictor variables showed negative correlations with grid cell percent impervious surface and positive correlations with grid cell percent canopy. Conclusion: The spatial distribution of EMS calls showed crude relative risks significantly vary across King County. Due to the strong negative relationship with impervious surfaces, we hypothesized BLS calls may be caused by recreational activities on extreme heat days. On the other hand, ALS calls were more severe ambulance calls and may be more driven by the presence of chronic health conditions. Further research and more patient specific variables need to be collected in order to confirm these hypotheses.