And Then You're Dead

The Autopsy / Case File · Disparities

The County Line

In 2014, a person born in Summit County, Colorado could expect to live to 86.8. A person born the same year in Oglala Lakota County, South Dakota could expect 66.8 — a full generation of difference, inside the same country, the same federal government, the same national life-expectancy average.

Researchers at the Institute for Health Metrics and Evaluation built county-by-county life tables covering every year from 1980 to 2014, then measured not just how long Americans were living, but how unevenly that improvement had spread. The headline number — a 20.1-year gap between the best- and worst-off counties — is large enough to sound like a data error. It isn't. And the gap has been growing, not shrinking, for the entire 35-year period the study covers.

And Then You're Dead · September 2026 · 8 min read

Join the discussion →
A view of rolling forested Appalachian mountains and hollows from Pine Mountain in Letcher County, Kentucky
The view from Pine Mountain, Letcher County, Kentucky — one of the eastern Kentucky counties the underlying study identified among the lowest life-expectancy places in the country.

National life expectancy at birth in the United States reached 79.1 years in 2014 — 76.7 for men, 81.5 for women — a figure that, taken on its own, describes a country that has made real, sustained progress since 1980, when the national average sat at 73.8. But a national average is, definitionally, an average, and the study that produced it — published in JAMA Internal Medicine by researchers at the University of Washington's Institute for Health Metrics and Evaluation — was built specifically to look underneath that number, county by county, for the entire 1980-2014 period.

Twenty Years, County to County

What the researchers found was a gap between the country's best- and worst-performing counties of 20.1 years in 2014 alone — Summit County, Colorado, at the top with a life expectancy of 86.8, and Oglala Lakota County, South Dakota, home to the Pine Ridge Indian Reservation, at the bottom with 66.8. Nearby Colorado counties clustered near the top of the national distribution as well: Pitkin County at 86.5, Eagle County at 85.9. The lowest-performing counties clustered geographically too — parts of North and South Dakota with large Native American reservation populations, counties along the lower Mississippi River, and a band running through eastern Kentucky and southwestern West Virginia, the heart of central Appalachia.

0
years separating the highest- and lowest-life-expectancy U.S. counties in 2014 — 86.8 years in Summit County, Colorado, versus 66.8 in Oglala Lakota County, South Dakota
0
additional years the gap between the 1st and 99th percentile counties widened by between 1980 and 2014, even as the national average rose 5.3 years

The narrower 10th-to-90th-percentile gap — a measure less driven by extreme outlier counties — still ran 6.2 years in 2014, and the 1st-to-99th-percentile gap ran 10.7 years. Every one of those measures describes the same underlying pattern at a different level of extremity: geography, inside a single country with a single set of federal health and safety-net programs, is producing life-expectancy differences on the scale usually associated with comparing entirely different nations.

Growing, Not Shrinking

The most consequential finding in the study isn't the size of the 2014 gap on its own — it's the trajectory. Absolute geographic inequality in life expectancy, measured as the spread between the 1st and 99th percentile counties, increased by 2.4 years between 1980 and 2014, even as every county in the country was, on average, gaining life expectancy over that period. The gains were not remotely evenly distributed: researchers found that western and coastal counties saw substantially larger improvements over the 35-year window, while a number of counties, concentrated in the South, saw only minimal gains — enough that some individual counties' life expectancy either stagnated or, in a minority of cases, actually declined over a period when the nation as a whole was living meaningfully longer.

Everyone got a longer life on average. Not everyone got the same longer life. The distance between the two ends of that distribution has been growing for at least three and a half decades.

What Explains It — and What Doesn't, Alone

The study didn't stop at documenting the gap; it modeled what explained it, using 2009 data as its explanatory cross-section. Socioeconomic and race/ethnicity factors, considered alone, statistically accounted for about 60% of the variation in county-level life expectancy. Behavioral and metabolic risk factors — smoking, obesity, physical inactivity, high blood pressure, diabetes — considered alone, accounted for a larger 74% of the variation. Healthcare factors, such as insurance-coverage rates and access to care, considered alone, explained a comparatively modest 27%. Combined, all three factor groups together explained the same 74% that behavioral and metabolic factors explained on their own — a result the researchers interpreted as evidence that socioeconomic conditions drive much of their statistical association with life expectancy indirectly, by shaping the behavioral and metabolic risk factors that sit closer to the actual biological cause of death, rather than through some separate, independent pathway.

That finding matters for how the gap gets talked about politically. It isn't simply a healthcare-access story, where the fix is more clinics or more insurance coverage — access factors explained the smallest share of the three. It isn't simply a genetics-and-bad-luck story either, since the same counties' life expectancy has moved so much relative to each other over just 35 years, a timeframe far too short for population genetics to explain. The strongest single statistical driver is the accumulated weight of behavioral and metabolic conditions that are themselves substantially shaped by decades of local economic conditions, education levels, and the built environment — the density of grocery stores, the prevalence of physically demanding versus sedentary work, the local prevalence of chronic disease risk factors passed down across generations of the same community.

Two Counties, Two Very Different Systems

The two anchor counties in the 2014 data aren't interchangeable examples pulled at random — they sit inside two structurally different relationships to the American healthcare and economic system. Oglala Lakota County is coextensive with the Pine Ridge Indian Reservation, where healthcare for most residents runs through the Indian Health Service, an agency that has been documented for decades, across multiple federal oversight reports, as chronically underfunded relative to the per-capita healthcare spending the federal government provides through Medicare and the VA. Summit County, by contrast, is a Colorado ski-resort county with a comparatively young, affluent, insured population and none of the reservation-specific jurisdictional and funding constraints that shape healthcare delivery on tribal land. The 20-year gap between them isn't simply "one place has worse luck than the other" — it's the measurable output of two residents living under two different, unequal versions of the same country's health and economic infrastructure.

The Gap Persisting Into the Present

The 1980-2014 study is now more than a decade old at its endpoint, but the underlying pattern hasn't reversed. A separate, more recent analysis by the same Seattle-based research institute — using a different methodology that combines race and geography into distinct population groups rather than tracking counties alone — found the life-expectancy gap between the highest- and lowest-performing group had widened to 20.4 years by 2021, using National Vital Statistics System data through the COVID-19 pandemic. That later study, published in The Lancet in November 2024, isn't directly comparable figure-for-figure to the county-level study above, since it measures a different unit of analysis. But the two studies, run by overlapping research teams using the same underlying vital-records infrastructure roughly a decade apart, arrive at strikingly similar-sized gaps — around 20 years — which is itself a kind of finding: whatever produced this scale of geographic and demographic inequality in life expectancy hasn't meaningfully narrowed in the years since the original county-level analysis was published.

And Then You're Dead

Nobody's death certificate says "county of residence" as a cause. It says heart disease, or diabetes, or an overdose, the same diagnosis codes used everywhere in the country. But the county line itself was already doing quiet, compounding work on that outcome years before any specific diagnosis got written down — twenty years of it, on average, between the two ends of the same map.

That's it. That's the whole thing.

Comments

Loading…

Comments are anonymous and moderated.

Sources

Photo: "View from Pine Mountain (Kentucky)," by J654567, via Wikimedia Commons (CC BY-SA / GFDL).