Advancing extreme heat risk assessments to better capture individually-experienced temperatures: A new approach to describe individual and subgroup vulnerabilities

Extreme heat risk assessments often rely on epidemiologic studies that used the nearest available outdoor airport temperatures (OATs) rather than individually-experienced temperatures (IETs) and frequently lack key individual-level determinants of exposure, including occupation, housing, and air conditioning. This hampers efforts to characterize heat burden inequities and guide interventions for vulnerable populations.

We developed an approach to estimate individual and subgroup-specific health impacts from modeled IETs before and during extreme heat events for three U.S. cities: Atlanta, Georgia (hot-humid), Detroit, Michigan (temperate), and Phoenix, Arizona (hot-dry).

IET profiles were estimated using modeled parcel-linked population microdata, housing-specific indoor temperatures from building energy models, ambient temperatures from urban-scale climate models, and time activity patterns from surveys. We linked each IET profile to daily OATs, then fit mixed-effects regressions to predict "equivalent" OATs (eOATs), based on IET, housing, and demographics. We assigned risk ratios (RRs) from existing literature on all-cause mortality, all-cause emergency department (ED) visits, and preterm births to each person-day's eOAT and estimated 5-day-extreme-heat absolute risks (ARs) by age-race-income-occupation subgroup.

The eOATs, RRs, and ARs differed between people due to variability in IETs and baseline health outcome incidence rates. All-cause mortality RRs ranges were 1.00-1.16 (Atlanta), 1.01-7.08 (Detroit), and 1.00-6.38 (Phoenix). All-cause-mortality ARs ranged 0.01-32 (Atlanta), 0.01-1,100 (Detroit), and 0.01-950 (Phoenix) per 100,000 persons. ED visit ARs ranges were 0.2-270 (Atlanta) and 0.04-6,200 (Phoenix) per 100,000 persons. Heat mortality ARs were higher among older adults and, only in Detroit, in young, Black, outdoor workers (median = 6.6 per 100,000) compared to young, non-Black, higher-income, indoor workers (median = 0.3 per 100,000).

When IETs can be estimated or directly measured, person-specific eOATs can be used to estimate the subgroup-specific heat-health burdens that would be experienced without adaptive behaviors. This approach could be adapted for other contexts to inform climate preparedness and justice policies. https://doi.org/10.1289/EHP15223.