Epic Research is not viewable using Internet Explorer. Please try accessing it with an alternate browser.
Cosmos Study

Summer Heat-Related Illness Rates Have Risen About 50% Since 2017

September 22, 2026
Dual-Team Study
Team A:Kersten Bartelt, RNEmily Higgs
Team B:Blaine Franklin, PT, DPTGrant Keane

Key Findings

  • The intensity of heat-related illnesses peaks in the summer and has varied from year to year with a slight upward trend: in 2017, the summer rate was 24.4 per 100,000 patients and in 2025 reached a high of 36.5 per 100,000 patients.
  • Heat-related illness was concentrated in the South, where average July rates were about 76.0 per 100,000 patients seen, more than three times the average rate in the West (about 22.5 per 100,000).
  • As expected, heat-related illness rates and locally issued heat-alert frequency rise together across most climate regions: the South ranks highest on both, but the West is a striking exception, recording nearly as many July heat-alert days as the South while having the lowest illness rate of any region.

Extreme heat is the leading cause of weather-related death in the United States, and both the frequency and intensity of extreme-heat events have been rising.1 Heat makes people sick when the body can no longer cool itself, often because high humidity slows the evaporation of sweat, and the resulting illness can progress from mild heat cramps to heat exhaustion and, in serious cases, life-threatening heat stroke.1 To warn the public of increased risk, the National Weather Service (NWS) issues heat advisories and warnings based largely on the heat index, a combined measure of temperature and humidity rather than temperature alone.2 There is no single, fixed temperature at which an alert is issued: the thresholds are set locally, higher where hot weather is common, and they also shift through the season as people acclimate to heat as summer wears on.3 Knowing when heat-related illness occurs across the year, how its intensity has changed over time, and how it compares with where heat alerts are actually issued can help health systems and public health agencies anticipate demand and target prevention.

We examined outpatient and emergency encounters among more than 246 million U.S. patients between January 2017 and May 2026, identifying heat-related illnesses (heat cramps, heat exhaustion, heat syncope, heat stroke, heat fatigue, and heat edema). We examined these rates overall and across the nine National Oceanic and Atmospheric Administration (NOAA) U.S. climate regions based on where care was delivered. To place the regional patterns alongside heat exposure, we paired the illness rates with counts of how many days each month a NWS heat advisory, warning, or watch was in effect anywhere in each region, drawn from a publicly available archive of NWS alerts.4

Heat-related illness in the U.S. follows a sharp, repeating seasonal cycle, rising from roughly 1 to 2 cases per 100,000 patients seen in the winter months to a pronounced peak each summer, almost always in July. The intensity of the summer peaks varied year to year with a slight upward trend: the summer rate (June, July, and August) was 24.4 per 100,000 patients in 2017 and reached a high of 36.5 per 100,000 patients in 2025.

Figure 1
Rate of Heat-Related Illness per 100,000 Patients Seen Over Time
Rate of Heat-Related Illness per 100,000 Patients Seen Over Time
Figure 1. The rate of heat-related illness per 100,000 patients over time.

Across most climate regions, heat-related illness rates and heat-alert frequency rise together, as seen in Figure 2. The South stands out on both measures, with the highest illness rates (about 76.0 per 100,000 patients seen in July) and the most July days under a heat alert (averaging about 23 of the month’s 31 days), while lower illness regions such as the Northeast and Upper Midwest see correspondingly fewer alert days. The clear exception is the West, which was under a heat alert on about 20 days in July on average, yet had the lowest heat-illness rates of any region, about 22.5 per 100,000. That exception might reflect how the heat alert measure is built rather than an absence of heat risk: alert days are counted for each region as a whole, regardless of where within the region the alert applied or how many patients are seen there, so a large, geographically varied region can register many alert days on areas that contribute few patients.

Figure 2
Average Heat-Alert Days vs. Heat-Related Illness Rates in July by Climate Region
Average Heat-Alert Days vs. Heat-Related Illness Rates in July by Climate Region
Figure 2. July heat-related illness rates per 100,000 patients seen in 2017 to 2025, shown alongside the average number of days in July under a National Weather Service heat alert, by NOAA climate region.

These data come from Cosmos, a dataset created in collaboration with a community of Epic health systems representing more than 310 million patient records from 2,200 hospitals and more than 50,000 clinics from all 50 U.S. states, Canada, Lebanon, and Saudi Arabia. This study was completed by two teams that worked independently, each composed of a clinician and research scientist. The two teams came to similar conclusions. Graphics by Brian Olson.

References

  1. Extreme Heat and Climate Change. Center for Climate and Energy Solutions. Updated May 2026. https://www.c2es.org/content/heat-waves-and-climate-change/. Accessed July 7, 2026.
  2. Heat Information. National Weather Service. https://www.weather.gov/lwx/heat. Accessed July 7, 2026.
  3. Heat. National Weather Service Phoenix. https://www.weather.gov/psr/heat. Accessed August 26, 2026.
  4. IEM Web Services and APIs. (n.d.). Iowa Environmental Mesonet. Retrieved August 26, 2026, from https://mesonet.agron.iastate.edu/api

Data Definitions

Study period
Study population: inclusion
Study population: exclusion
Outcomes
Heat-related illness
Heat alerts
Climate region
Model specifications
Limitations