Table 1.

Heat-related excess mortality (%, 95% eCI) by region, period, and climate change scenario, assuming no adaptation or population changes.

Region	RCP 4.5				RCP8.5
	2010s	2030s	2050s	2090s	2010s	2030s	2050s	2090s
Northern	2.8 (−0.9 to 6.1)	3.2 (−0.7 to 6.6)	3.6 (−0.4 to 7.2)	4.0 (−0.2 to 7.8)	2.9 (−0.9 to 6.2)	3.3 (−0.6 to 6.8)	4.2 (−0.0 to 7.9)	6.4 (1.1–12.0)
Northeast	0.9 (0.1–1.6)	1.2 (0.1–2.3)	1.6 (0.1–3.2)	1.9 (0.1–4.1)	0.9 (0.1–1.8)	1.3 (0.1–2.5)	2.0 (0.2–4.2)	4.3 (0.4–8.9)
Northwest	0.5 (−0.6 to 1.5)	0.7 (−0.8 to 2.0)	0.9 (−1.1 to 2.6)	1.1 (−1.3 to 3.3)	0.5 (−0.6 to 1.5)	0.8 (−0.9 to 2.2)	1.1 (−1.5 to 3.5)	2.3 (−3.5 to 7.0)
Eastern	1.8 (0.6–2.9)	2.3 (0.9–3.6)	2.8 (1.1–4.6)	3.3 (1.3–5.7)	1.8 (0.6–3.0)	2.4 (1.0–3.9)	3.5 (1.5–5.6)	6.2 (2.5–10.2)
Central	2.6 (−0.4 to 5.5)	3.1 (−0.1 to 6.0)	3.5 (0.2–6.6)	4.0 (0.5–7.2)	2.7 (−0.4 to 5.5)	3.2 (0.0–6.2)	4.0 (0.6–7.3)	6.3 (1.8–10.6)
Southern	1.9 (−0.5 to 3.7)	2.3 (−0.7 to 4.6)	2.9 (−1.2 to 6.2)	3.4 (−1.6 to 7.7)	1.9 (−0.5 to 3.7)	2.5 (−0.8 to 5.1)	3.6 (−1.8 to 7.8)	6.5 (−5.8 to 14.8)
Southwest	1.2 (−0.4 to 2.5)	1.4 (−0.7 to 3.2)	1.7 (−1.1 to 4.1)	2.1 (−1.4 to 5.1)	1.1 (−0.4 to 2.5)	1.5 (−0.8 to 3.4)	2.1 (−1.6 to 5.2)	3.9 (−4.2 to 10.1)
National	1.8 (0.2–3.3)	2.2 (0.4–3.9)	2.6 (0.5–4.6)	3.1 (0.6–5.5)	1.9 (0.2–3.3)	2.4 (0.4–4.1)	3.2 (0.6–5.6)	5.5 (0.5–9.9)

A distributed lag nonlinear model was used to estimate the district/county-specific temperature–mortality association with 14 days of lag adjusted for time trends and day of the week, which were pooled in a multivariate meta-analysis. Then, estimates of the attributable fraction of deaths due to high temperature, defined as temperatures above the optimum temperature, were calculated by the regional and national levels. Monte Carlo simulations generating 1000 samples were computed to produce empirical confidence intervals.