Rural counties have ~18% higher premature-death rates than urban counties today, and the gap has more than tripled since 1999 — but it is far from uniform across states

Summary

Rural US counties have meaningfully worse health outcomes than urban counties, and the disadvantage has grown sharply over the past two decades. Right now, the average county-level premature-death rate (years of life lost before age 75, per 100,000 people) is 11,051 in nonmetro (rural) counties versus 9,359 in metro (urban) counties — an 18% gap, computed here from 3,080 US counties. Nationally, published USDA ERS/CDC research (not computed in this session) shows that gap has been widening for decades: working-age (25-54) rural adults died of natural causes at only a 6% higher rate than urban adults in 1999-2001, but a 43% higher rate by 2017-2019 — a roughly seven-fold increase in the relative gap, driven mainly by rising mortality among rural women and, especially, rural Native American communities.

The gap does not look the same within a state as it does between states. Computed within each of 45 states individually, the average rural-minus-urban gap is about 1,820-1,850 YPLL/100k (roughly +21% relative to that state's urban rate) — close to, and slightly larger than, the national pooled figure. But that within-state gap varies enormously by state: from a small negative gap in Idaho, Wyoming, and Massachusetts (rural counties there actually score BETTER) up to +7,131 in Arizona and +5,650 in South Dakota. Meanwhile the between-state variation is comparably large or larger: urban-county averages alone range from 6,139 (Utah) to 14,007 (Mississippi), a spread of nearly 8,000 YPLL/100k — almost 5 times the size of the typical within-state rural penalty. In short: which STATE a county sits in explains at least as much of the variation in health outcomes as whether that county is rural or urban, and the two effects partly compound — states with worse overall health (the South, Appalachia, parts of the Mountain West) tend to have somewhat larger rural-urban gaps too (correlation r≈0.43 across states), but that relationship is loose, not mechanical.

Step 1 — The current cross-sectional rural-urban gap (this connector)

Queried health.chr_premature_death (County Health Rankings & Roadmaps, 2025 release, county grain, single cross-section) joined to geo.rural_urban_continuum (USDA ERS Rural-Urban Continuum Codes, 2022 vintage) on county FIPS, for all 3,080 counties with both a premature-death rate and a metro/nonmetro classification (1,178 metro, 1,902 nonmetro):

Gap: 1,692 YPLL/100k, or +18.1% relative to the urban rate. Rural counties also show substantially more DISPERSION (SD 4,145 vs 2,981) — rural health outcomes are both worse on average and less uniform.

Important caveat on the time dimension: this table holds only ONE loaded release (2025) at county grain in this warehouse — it is a current snapshot, not a 20-year county-level panel. No table in this corpus carries county-grain mortality across a 20-year span (CDC WONDER's underlying-cause API explicitly rejects sub-national location grouping, confirmed via the corpus's own table documentation). The 20-year trend below is therefore drawn from published literature, not computed here, and is reported at the national/age-group level, not the county level.

Step 2 — The 20-year trend (published research, not computed in this session)

USDA Economic Research Service, using CDC death data, compared natural-cause mortality in rural and urban areas between two three-year windows, 1999-2001 and 2017-2019. Key published findings:

• Rural adults ages 25-54 died of natural causes at a 6% higher rate than urban adults in 1999-2001; by 2017-2019 that gap had grown to 43%.
• Natural-cause mortality rates fell for urban adults across all age groups over the two decades.
• Rural mortality fell for most age groups but ROSE for the prime working-age population; the rural group with the largest increase (+19%) was ages 30-34.
• The gap widened fastest for rural Native American adults 25-54 (+46% natural-cause mortality) and rural Native American women in particular (+55%).

This is consistent with, and complementary to, the current 18% cross-sectional gap measured above from CHR data: the ERS/CDC figure (43% for one working-age slice, natural causes only) is not directly comparable in magnitude to the all-ages YPLL figure computed here, but both point the same direction and the ERS trend explains why the present-day gap is as large as it is — it built up steadily rather than appearing suddenly.

Step 3 — Within-state vs. between-state: does the gap look the same?

To test whether the rural-urban gap is a fixed, state-independent quantity or something that varies by geography, I computed the rural-minus-urban gap SEPARATELY within each of 45 states (states with at least 3 rural and 1 urban county in the CHR data), then compared the distribution of those 45 within-state gaps to the between-state variation in county-average mortality itself.

Answer: no, the gap does not look the same within a state as between states, and in two distinct ways. First, the SIZE of the rural penalty itself swings from roughly zero or even reversed (Idaho, Wyoming, Massachusetts: rural counties there average LOWER premature-death rates than their state's urban counties) to more than seven times the national gap (Arizona: +7,131 YPLL/100k, nearly 70% higher rural mortality within that one state). Second, and just as important, the BETWEEN-STATE variation in mortality — how much urban counties differ from each other depending on which state they're in, or how much rural counties differ from each other by state — is comparable to or larger than the typical within-state rural-urban gap (SD ~2,000-2,600 between states vs. a typical within-state gap of ~1,800-1,850). A rural county in Utah (rural avg 8,448) looks healthier than an urban county in Mississippi (urban avg 14,007). Geography-of-state is doing at least as much work as rural/urban status; treating 'rural' as a single national category obscures this.

Data and methodology notes

Grain: county. Outcome: premature-death rate (years of potential life lost before age 75 per 100,000, age-adjusted), the current-era county-grain mortality-outcome measure available in this warehouse. Classification: USDA ERS Rural-Urban Continuum Codes collapsed to the standard metro (codes 1-3) / nonmetro (codes 4-9) binary, 2022 vintage (RUCC updates only decennially, so this is the best available and is treated as effectively time-invariant for classification purposes). States with fewer than 3 rural or 0 urban counties were excluded from the state-level comparison (5 states/territories dropped) — this removes small, noisy state-level estimates but does not change the headline national figures, which used the full county set. All within-state figures come from the same single CHR 2025 cross-section as the national figures — this section answers 'does the CURRENT gap vary by state,' not a 20-year within-state trend, since no historical county panel exists in this corpus.

What This Report Does Not Answer

• Over the last twenty years (trend over time): This warehouse's only county-grain mortality outcome table (health.chr_premature_death) carries a single loaded cross-section (2025 release) — no 20-year county-level panel exists in this corpus, and CDC WONDER's NVSS mortality API explicitly rejects county-level location grouping. The 20-year trend is instead reported from a published USDA ERS/CDC analysis (natural-cause mortality, ages 25-54, national level, 1999-2001 vs 2017-2019), which is the best available substitute but is not computed from this connector's own data and is not at county grain.

Every query behind this report

3 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT release_year, COUNT(*) FROM health.chr_premature_death WHERE geo_level='county' GROUP BY release_year ORDER BY release_year

SELECT r.metro_nonmetro, COUNT(*) n, AVG(c.premature_death_rate) avg_rate, STDDEV_SAMP(c.premature_death_rate) sd
FROM health.chr_premature_death c
JOIN geo.rural_urban_continuum r ON c.fips_code = r.county_fips
WHERE c.geo_level = 'county' AND c.premature_death_rate IS NOT NULL
GROUP BY r.metro_nonmetro

SELECT c.state_abbr,
  AVG(CASE WHEN r.metro_nonmetro='Nonmetro' THEN c.premature_death_rate END) AS rural_avg,
  AVG(CASE WHEN r.metro_nonmetro='Metro' THEN c.premature_death_rate END) AS urban_avg,
  COUNT(CASE WHEN r.metro_nonmetro='Nonmetro' THEN 1 END) AS n_rural,
  COUNT(CASE WHEN r.metro_nonmetro='Metro' THEN 1 END) AS n_urban
FROM health.chr_premature_death c
JOIN geo.rural_urban_continuum r ON c.fips_code = r.county_fips
WHERE c.geo_level='county' AND c.premature_death_rate IS NOT NULL
GROUP BY c.state_abbr
HAVING COUNT(CASE WHEN r.metro_nonmetro='Nonmetro' THEN 1 END) >= 3 AND COUNT(CASE WHEN r.metro_nonmetro='Metro' THEN 1 END) >= 1
ORDER BY c.state_abbr