Exercise Access Was Linked to 11% Lower Heart, Kidney and Metabolic Deaths in 3,028 Counties

A new nationwide analysis reports that county-level exercise access tracked with deaths from heart, kidney and metabolic disease across almost the entire United States. Counties with the thinnest exercise access and the highest inactivity carried the heaviest burden.

Exercise access — the share of residents living close enough to a park or a recreation facility to use one — was associated with lower cardiovascular-kidney-metabolic mortality across 3,028 US counties, according to an ecological study published in Frontiers in Public Health on October 1, 2026.

The analysis was carried out by Chen Zhao, Zhisheng Zhu, Jiaxian Zhang, Ziyue Ren and Hongwei Liu of the First Affiliated Hospital of Jinan University in Guangzhou, China. The team joined county-level death records drawn from CDC WONDER to physical activity measures from County Health Rankings, covering the years 2014 through 2019.

In total, 3,028 of the 3,235 US counties qualified, roughly 93.6% of the country. Counties were ranked into quartiles for exercise access and, separately, for physical inactivity. Age-adjusted mortality rates were then compared using quasi-Poisson generalized linear models, adjusted first for year alone and then for a fuller set of demographic and socioeconomic covariates.

Counties in the top exercise access quartile showed a rate ratio of 0.89 (95% CI 0.87 to 0.90) for total cardiovascular-kidney-metabolic mortality against counties in the bottom quartile. Looking at cardiovascular deaths on their own, the rate ratio was also 0.89 (95% CI 0.88 to 0.91).

How the Study Measured Exercise Access

Exercise access here is not a survey question about gym membership. The authors defined a reasonable distance as living within half a mile of a park, within one mile of a recreational facility in urban areas, or within three miles of such a facility in rural areas.

To build that measure, they combined park boundaries from the Esri USA Parks dataset with recreational facilities identified through Standard Industrial Classification codes, then overlaid 2020 census block population counts in ArcGIS Pro. The output is a single county-level percentage: how much of the population has exercise access by that definition.

Where the Gradient Was Steepest

The geography was not subtle. Counties in the South had the least exercise access, the highest levels of physical inactivity, and the greatest cardiovascular-kidney-metabolic death burden of any region in the analysis.

That clustering is the part of the paper most likely to matter to planners. It means the counties that would gain most from new trails, pools and recreation centers are, on average, the counties that have the fewest of them now.

The authors also stratified their estimates by sex, race and level of urbanization. The general direction of the association held across most of those subgroups, though the size of the effect varied between them. That consistency is one reason the combined estimate is reported with such narrow confidence intervals, and it is also why the regional concentration cannot be dismissed as an artifact of one demographic group.

Why the Measure Favors Walkable Counties

The distance thresholds deserve a second look, because they are not symmetrical. In urban areas a recreational facility counts if it sits within one mile. In rural areas the allowance widens to three miles, which quietly accepts that a car is part of the trip.

That design choice is defensible and it is also consequential. A rural county can register solid exercise access on paper while every visit still requires a vehicle, fuel and a block of free time. An urban county with the same score may put the same facility within a ten-minute walk.

So the exercise access variable is better read as a measure of theoretical reach than of lived convenience. It captures whether the infrastructure exists nearby. It does not capture sidewalks, lighting, safety after dark, opening hours, entry fees or whether a shift worker can get there at all.

What Physical Inactivity Added to the Picture

Inactivity was modeled separately, and the association ran the other way. Counties in the highest inactivity quartile carried a rate ratio of 1.23 (95% CI 1.22 to 1.25) for total cardiovascular-kidney-metabolic mortality compared with the least inactive counties.

Cardiovascular disease drove most of that difference, at a rate ratio of 1.24 (95% CI 1.23 to 1.26). The pattern is consistent with earlier device-based work, including the accelerometer analysis in which 12 minutes of vigorous activity a week matched 150 moderate minutes for lower mortality.

Why Exercise Access and Inactivity Are Not the Same Variable

It would be easy to treat the two measures as one story told twice. They are not. Exercise access describes what a place offers; inactivity describes what its residents actually do. A county can score well on one and badly on the other.

That gap is where most public health leverage sits. Building a trail does not by itself change behavior, which is part of why work on exercise participation and self-rated health in 76,827 adults has focused on who takes up opportunities rather than on who simply lives near them.

The occupational literature points the same way. Our earlier report on cardiometabolic risk across desk-based and physically demanding jobs found that where movement happens, not only how much of it there is, shapes the health signal.

Where the Component Results Got Complicated

The headline association held for the combined outcome and for cardiovascular deaths. It did not hold uniformly across all three components. The authors report that associations with diabetes and renal failure mortality were less consistent, including a non-monotonic gradient across inactivity quartiles for diabetes.

They also flag an unexpected direction in the highest inactivity quartile for renal failure deaths. Readers should treat the component-level numbers as unsettled and the combined estimate as the paper’s firmer result.

What an Ecological Design Can and Cannot Show

An ecological study compares aggregates. It can reveal that counties with more exercise access bury fewer residents from these causes. It cannot tell any individual reader what their own risk is, and it cannot establish that access caused the difference.

Fitness research has repeatedly shown how different the picture looks when measurement moves from the population to the person, as in the pooled work on estimated cardiorespiratory fitness and disease risk across three large cohorts.

What This Means for Your Own Week

Nothing in this paper tells you how much to train. What it does suggest is that the friction between you and a place to move is worth taking seriously, because at the population scale that friction shows up in mortality statistics.

A few practical readings that stay inside what the evidence supports:

• Proximity is a real constraint, not an excuse. If the nearest facility is far, the honest plan is one that does not depend on reaching it.
• Exercise access is a community variable. It responds to local budgets and zoning more than to personal willpower.
• The cardiovascular signal was the clearest. General activity guidance, not a specialized program, is what this evidence base supports.
• Anyone with existing heart, kidney or metabolic disease should set training plans with their own clinician rather than from population data.

Readers who want the behavioral side of the question will find more in our coverage of how 33 million gym visits pointed to coaching and group classes as habit supports, which addresses the step after access: actually showing up.