NYC's heat warnings start at 95 degrees - but older patients flood ERs at 90

Two emergency departments sit ten miles apart in New York City. Both belong to the same health system. Both see patients over 65. But when summer heat arrives, they tell very different stories.

At the first hospital - a community-based academic center serving a racially and ethnically diverse, lower-income population - older adults started showing up with heat-associated illnesses when the heat index hit just 66 degrees Fahrenheit. By 90 degrees, the surge was unmistakable. At the second hospital, a larger academic medical center serving wealthier, predominantly white, privately insured patients, the researchers found no significant link between heat and emergency visits at all.

The gap between those two realities sits at the center of a new study published March 20 in JAMA Network Open, and it raises an uncomfortable question: are municipal heat warnings calibrated to protect the people most likely to suffer?

The 95-degree threshold and who it misses

New York City issues heat advisories when the heat index is forecast to reach 95 degrees Fahrenheit for at least two consecutive days, or 100 degrees for any duration. These thresholds trigger cooling center openings and public messaging campaigns. They are based on population-level mortality data - a broad view of when heat starts killing people across the city.

But the research team at NYU Grossman School of Medicine, led by Alexander Azan, assistant professor in the Department of Population Health, wanted to know whether those population-level thresholds matched what they were seeing in the clinic. So they examined electronic health records for patients aged 65 and older at their two emergency departments.

The dataset was substantial: 55,200 ED visits representing 15,092 unique patients at the first hospital and 19,559 at the second, spanning summer months from May through September across 2022 to 2024. Temperature data came from the LaGuardia Airport monitoring station, and the team calculated heat index values to allow direct comparison with the city's advisory thresholds.

Risk starts climbing at 66 degrees

At the community hospital, the relationship between heat and emergency visits began at a daily heat index maximum of just 66 degrees Fahrenheit. That is nearly 30 degrees below the city's advisory trigger. The association was modest at lower temperatures but amplified sharply between 90 and 101 degrees.

The 90-degree inflection point is the critical finding. It falls five degrees below the 95-degree threshold that currently activates the city's heat response infrastructure. The research team estimated that a healthcare system-based warning triggered on days reaching 90 degrees - on days without a city-issued advisory - could have prevented roughly 116 emergency visits at the community hospital during the three-year study period.

At the second hospital, serving a more affluent, privately insured population, the researchers found no statistically significant association between heat index and ED use among older adults. The contrast between the two sites is stark and telling.

Why income and race shape heat vulnerability

The divergence between the two hospitals tracks with established research on heat vulnerability. Older adults in lower-income communities face compounding risks. They are more likely to live in housing without reliable air conditioning, in neighborhoods with less tree cover and more heat-absorbing pavement, and in buildings that trap heat. They are more likely to have chronic conditions - heart disease, kidney disease, diabetes - that impair the body's ability to thermoregulate. They are more likely to take medications such as diuretics, beta-blockers, or anticholinergics that interfere with sweating or blood flow regulation.

The proportion of patients enrolled in Medicaid at the community hospital was twice that of the academic medical center. Medicaid enrollment is a rough proxy for poverty, and poverty correlates with nearly every factor that makes heat more dangerous.

Nationally, heat-related mortality has risen by roughly 17% per year since 2016. That trend is not distributed equally. It falls hardest on people who cannot afford to escape the heat.

Electronic health records as an early warning system

The study's practical contribution is methodological as much as clinical. Rather than relying solely on population-wide weather data to set thresholds, the researchers used hospital-specific electronic health records to identify where heat risks actually manifest.

Leora I. Horwitz, director of the Center for Healthcare Innovation and Delivery Science at NYU Langone and a co-investigator on the study, emphasized the broader applicability of this approach. Each health system serves a different patient population with different vulnerabilities. A universal threshold will inevitably be too high for some communities and too low for others.

The researchers are now planning to pair these locally relevant exposure thresholds with a comprehensive evaluation of social and structural risk factors - examining which specific health conditions in older adults are most sensitive to heat and how structural inequities compound the risk.

What the study cannot tell us

This is an observational analysis of emergency department visits, not a randomized trial. It can identify associations between heat exposure and ED use but cannot prove causation. Patients may come to the emergency department for reasons not directly captured in heat-related diagnosis codes, and some heat-associated illness may go uncounted if patients do not seek care at all.

The study covers only two hospitals within a single health system in one city. New York's urban heat island dynamics, building stock, and demographics are specific. The exact temperature thresholds identified here - 66 degrees, 90 degrees - may not apply to Houston, Phoenix, or rural Georgia. That is precisely the authors' point: each system needs to analyze its own data.

The three-year study window also limits the ability to account for year-to-year climate variability. And the use of a single weather monitoring station at LaGuardia Airport means the temperature data reflects airport conditions rather than microclimates in specific neighborhoods, where heat exposure can vary block by block.

The study also does not track what happens to patients after they leave the emergency department. Heat exposure that produces an ED visit may also contribute to longer-term health deterioration that this dataset cannot capture.

Closing the five-degree gap

The immediate implication is specific: for older patients in underserved communities in New York City, the danger threshold for heat is lower than the city's current advisory system recognizes. A five-degree difference in the trigger temperature could translate to days of additional warning time per summer - days when cooling centers could open earlier, when outreach workers could check on isolated seniors, when healthcare systems could pre-position resources.

Whether municipalities adjust their advisory thresholds is a policy question. But the study makes a clear case that healthcare systems do not need to wait for city-level action. Hospitals already have the data. Electronic health records can reveal the temperature thresholds at which their specific patient populations begin to suffer. The question is whether they choose to look.