Heat-related deaths are vastly undercounted in the United States, with new estimates suggesting thousands more people die each year than official records show.
As summer temperatures rise with human-driven climate change, the nation’s count of heat-related deaths fails to capture the true toll, according to researchers. Inconsistent reporting and limited documentation on death certificates leave a large gap in the data, and experts warn that not grasping the full scope will cost lives.
To address that gap, a two-year investigation built a more comprehensive national estimate.
How big is the undercount?
The Centers for Disease Control and Prevention compiles annual heat-related deaths from state reports when heat is listed on death records. From 2004 to 2018, the official national average was about 700 deaths a year. In recent years the average has risen to around 1,700 and has sometimes exceeded 2,000.
Even those higher numbers fall short. An analysis conducted with Boston University estimates that on average between 2018 and 2025, roughly 9,000 deaths each year were influenced by heat. In some years the total likely surpassed 10,000. Researchers say the true figure could be higher due to data limitations and the study’s conservative methods.
Climate change is amplifying the danger. The heat wave season has lengthened by about a month and a half since the 1960s, and the number of heat waves in many major U.S. cities has doubled since the 1980s. Large heat domes now regularly expose tens of millions of people to dangerous conditions. Researchers say more frequent, longer and more intense events will reduce opportunities for the body to recover between bouts of extreme heat.
One key reason for the undercount is how heat interacts with existing conditions. Physicians note that extreme temperatures can trigger fatal events such as heart attacks in people with cardiovascular disease by forcing the heart to work harder to cool the body. Because heat leaves few clear signs, it is often omitted from death certificates, which then excludes those deaths from official tallies.
National risks — even in unexpected places
The investigation found heat is taking lives across nearly every region. Florida and Texas record the highest numbers overall, driven partly by population size, yet they still rank among the most affected even after adjusting for population.
States not typically seen as hot are also bearing significant losses. The model estimates more than 100 heat-related deaths each year in Michigan, about ten times the official count. In Massachusetts, the estimate exceeds 60 a year compared with roughly 5 in official records. In Florida, the gap is stark, with estimates above 1,900 annually, around 40 times the reported total.
People in cooler climates can face elevated risk because they are less acclimated and often lack cooling infrastructure. Air conditioning is nearly universal in Alabama, but only about 53% of homes in Washington state and 70% in Maine have it, even as heat risks grow there.
Counts are getting more accurate — but not everywhere
Official national numbers have climbed in part because extreme heat is worsening and because some jurisdictions have improved detection. Maricopa County, Arizona, has developed detailed protocols to better identify heat’s role, including asking about air conditioning functionality, ability to pay utility bills and medications that increase heat sensitivity.
Many places still undercount. Counties that include Los Angeles and Houston report relatively few heat-related deaths, yet modeling suggests hundreds likely die each year in each county due to heat. Local health officials acknowledge that official numbers can be substantial undercounts.
Improving accuracy is challenging and takes time. Hospitals, emergency departments and certifying physicians often lack standardized guidance to confidently identify heat’s contribution to a death. Until those gaps are closed, researchers say the undercount will persist, obscuring the urgency for policymakers, planners and families to take protective action.
Heat-related deaths and the research methods
Researchers analyzed weekly all-cause mortality from 2018 through 2025 in 530 counties with at least 50,000 residents, representing about 73% of the U.S. population in 2024. Data from 2020 was excluded because the COVID-19 pandemic disrupted typical mortality patterns. For comparison, they reviewed federal data listing heat exposure as a primary or contributing cause on death records.
Temperature estimates came from ERA5-Land, a global weather dataset. For each county and week, the team calculated a population-weighted average of daily maximum temperature, and ranked temperatures relative to each county’s usual climate to reflect local acclimatization.
Using space-time stratified case-crossover methods, the team estimated how unusually hot weeks affected deaths during the same and following week, allowing for nonlinear effects at very high temperatures. County-level results were combined within states to find the minimum mortality temperature, then extrapolated to the contiguous U.S. population. Heat-attributable deaths were defined as those associated with temperatures above the state-specific minimum mortality temperature. Analyses were conducted with the open-source cityClimateHealth R package.
Researchers plan to publish the full analysis in a scientific journal and will share additional reporting on whose deaths are being missed and how lives might be saved. For more on extreme weather risks, see how a Grand Canyon flash flood turned deadly.
2 Responses