For a Copy of the full report, contact me here
The Geography of Diagnosed Diabetes in the United States
Rural demographic disparities, health-system capacity, and strategic implications
A Stark Geography of Burden
The 2024 Rural Health Information Hub county map reveals concentrated clusters of diagnosed diabetes across central and southern Appalachia, the Mississippi Delta, and large portions of the Southeast. The central analytical conclusion is not that rurality independently causes diabetes. Rather, rural residence organizes exposure to a distinctive combination of older age structure, lower income, lower educational attainment, food insecurity, higher obesity prevalence, transportation burden, limited clinical workforce, and weaker access to diabetes self-management education and support.
Five Leadership Conclusions
Is the organization’s diabetes prevention and care capacity distributed according to county need, or according to historical facility location and reimbursement convenience?
Data Sources and Analytical Approach
The primary visual is the RHIhub map “Diagnosed Diabetes Prevalence, 2024,” displaying county-level estimates in five categories. The analysis treats the map as a descriptive representation of place-based burden, while causal interpretation is derived from peer-reviewed survey and county studies.
County Prevalence Bands
Source: Rural Health Information Hub (2024), using CDC Diabetes County Data Indicators. Nonmetro counties follow OMB delineations (micropolitan plus noncore). The national view includes all counties and should not be interpreted as a rural-only map.
Place-Based Causal Framework
The framework separates four levels that are often conflated, avoiding treatment of “rural” as a biological characteristic. Rurality is instead a spatial organization of structural conditions. Outcomes feed back into household income, disability, workforce participation, and local institutional capacity.
Place & Policy
Determine resource distribution and exposure
Population Composition
Shapes baseline risk
Care Capacity
Affects diagnosis, control, and complication prevention
Outcomes
Feed back into income, disability, workforce, and institutions
Author synthesis based on Hill-Briggs et al. (2021), Dugani et al. (2024), and recent CDC rural diabetes studies.
Evidence Triangulation
| Source | Design | Analytical contribution | Interpretive constraint |
|---|---|---|---|
| RHIhub county map | County choropleth | Spatial distribution and prevalence bands | Descriptive; no uncertainty intervals or age adjustment |
| Yell et al. (2026) | NHIS, 2021–2024 | National rural-urban subgroup prevalence | Current, nationally representative; crude estimates and self-report |
| Onufrak et al. (2024) | NHIS, 2019–2022 | Region by metropolitan status | Tests whether regional rural patterns persist after adjustment |
| Khavjou et al. (2025) | BRFSS, 2021, 41 states | State-specific rural-urban differences | Shows how adjustment for demographics and obesity changes interpretation |
| Dugani et al. (2024) | County trends, 2004–2019 | Rurality and contextual determinants | Demonstrates explanatory importance of modifiable county conditions |
| Probst et al. (2024) | County service availability | DSMES, dialysis, hospitals, broadband | Assesses whether prevention capacity is located where need is greatest |
Analytical Conventions
Diagnosed only
Prevalence refers to self-reported diagnosed diabetes. It does not include undiagnosed diabetes or prediabetes.
Absolute vs relative
Percentage-point differences are absolute. Relative differences are ratios and should not be confused with absolute burden.
Crude vs adjusted
Crude estimates measure population burden; adjusted estimates address whether differences persist after accounting for measured covariates.
Clusters as hypotheses
Visual clusters are hypotheses about place-based burden, not proof of causation or individual-level risk.
Geospatial Findings
The 2024 visualization reveals a broad high-prevalence corridor across the Southeast, the Mississippi Delta, and central and southern Appalachia, with additional high-prevalence counties in parts of Oklahoma, Texas, New Mexico, and Arizona. Lower-prevalence clusters are more common in the Mountain West, Pacific Northwest, Upper Midwest, and Northeast. The pattern is consistent with the established “diabetes belt,” yet neighboring counties can occupy different prevalence categories, suggesting that state policy alone cannot explain local burden.
Because the national screenshot includes metro and nonmetro counties, dark shading in the South cannot be attributed exclusively to rural residence. The correct conclusion is that diabetes is geographically concentrated, and that rurality should be examined as one interacting dimension of that geography.
National Rural-Urban Difference, 2021–2024
Source: Yell et al. (2026), pooled National Health Interview Survey.
State Variation: The Limits of a Single National Estimate
State-level BRFSS analysis across 41 states found crude prevalence of 14.3% rural and 11.2% urban. Rural estimates ranged from 8.4% in Colorado to 21.3% in North Carolina; urban estimates ranged from 6.9% in Colorado to 15.5% in West Virginia. Nineteen states had a statistically significant unadjusted rural excess. After full adjustment, a significantly higher rural likelihood remained only in North Carolina and Oregon, and the national difference was no longer significant (Khavjou et al., 2025).
| Selected state | Rural | Urban | Absolute gap | Interpretation |
|---|---|---|---|---|
| North Carolina | 21.3% | 12.1% | +9.2 pp | Wide rural confidence interval; fully adjusted excess remained significant |
| Oregon | 20.4% | 9.2% | +11.2 pp | Very small rural sample and wide confidence interval; fully adjusted excess remained significant |
| Tennessee | 18.5% | 13.4% | +5.1 pp | High rural burden; adjusted difference attenuated |
| South Carolina | 17.8% | 13.5% | +4.3 pp | High prevalence in both rural and urban populations |
| West Virginia | 17.0% | 15.5% | +1.5 pp | Very high statewide burden, with a comparatively small rural-urban gap |
| Virginia | 16.4% | 10.9% | +5.5 pp | Rural excess attenuated after full adjustment |
A state can have a large rural-urban gap, a high burden in both settings, or both. Resource allocation should therefore use a two-axis matrix: absolute county burden and rural-urban disparity. Focusing on disparity alone can overlook high-need urban counties; focusing on prevalence alone can overlook structural inequity.
Demographic and Socioeconomic Gradients
Prevalence rises sharply with age in both rural and urban populations, but the rural excess appears in every adult age group below 65, converging only at 65 and older. This pattern challenges the assumption that rural diabetes is simply a by-product of an older population: the persistent excess among working-age adults indicates earlier onset, greater cumulative exposure to metabolic and social risk, or both.
Age: A Working-Age Rural Signal
Source: Yell et al. (2026). The 65+ rural-urban difference (20.8% vs 20.0%) is not statistically significant.
Source: Farrigan et al. (2024). The working-age rural population contracted over the same period.
Dual mandate
The rural population aged 65 and older grew from 7.4 million in 2010 to 9.7 million in 2023 while the working-age population contracted, placing more chronic-care demand on a smaller labor force. A rural diabetes strategy must therefore accomplish two objectives simultaneously: prevent or delay disease among younger adults and maintain function among a rapidly expanding older population.
Gradient Explorer
Use the pills to explore how education, income, food security, and body weight stratify the rural and urban burden. Economic resources are protective but do not fully neutralize place-based disadvantage, and these associations should not be reduced to individual choice: food prices, travel distance, retail structure, work schedules, housing, safe spaces for activity, and chronic stress shape diet and activity patterns (Hill-Briggs et al., 2021).
Source: Yell et al. (2026). At every level from high school completion upward, rural prevalence remained significantly higher. The similar burden in the lowest education category reflects very high prevalence in both settings, not equity.
Race, Ethnicity, and Compounded Place-Based Risk
Rural diabetes disparities occur within racial and ethnic groups, not only between them. The highest rural prevalence in the grouped 2021–2024 data occurred among non-Hispanic Black adults, consistent with a compounding-place framework: race and ethnicity index differential exposure to historical and current structures, while rurality shapes service availability, travel burden, employment, and local institutional capacity.
Source: Yell et al. (2026). The rural-urban difference among Hispanic adults is not statistically significant. The “other” category aggregates heterogeneous populations, including Asian and American Indian adults, and should be interpreted cautiously.
Attenuation after adjustment does not mean that rural geography is irrelevant. Age, income, education, and obesity are not random nuisances that happen to differ by geography. They are part of the mechanism through which decades of migration, labor-market restructuring, educational opportunity, food systems, and healthcare investment produce place-based health outcomes.
The Structural Profile Behind Crude Prevalence
Older
Rural adults were older on average in 2021 BRFSS data
Lower income
More likely to have household income below $35,000; less likely to reach $75,000 or more
Higher obesity
More likely to have obesity than urban counterparts
Lower attainment
More likely to have lower educational attainment
Source: Khavjou et al. (2025). These compositional differences help explain why statistical adjustment substantially attenuates the rural-urban association.
Insurance, Diagnosis, and the Detection Paradox
Where differences appear
Rural-urban differences were present among adults with public and private insurance but not among the uninsured. The absence of a difference among uninsured adults may partly reflect underdiagnosis, because the outcome requires prior contact with a healthcare professional. Counties with weaker access may appear less burdened than they truly are.
Executive dashboards should combine diagnosed prevalence with screening rates, preventable admissions, emergency utilization, pharmacy fills, hemoglobin A1c testing, and complication indicators.
Regional Heterogeneity and Health-System Capacity
Region modifies the relationship between metropolitan status and diagnosed diabetes. In 2019–2022 NHIS data, nonmetro prevalence ranged from 9.0% in the West to 14.8% in the South. The persistence of the Southern nonmetro effect after adjustment suggests that measured demographics and obesity do not fully capture regional conditions, rejecting a one-size-fits-all rural strategy.
The Nonmetro South as a Distinct Epidemiologic Context
Source: Onufrak et al. (2024), National Health Interview Survey.
Adjusted odds ratio: 1.62 after age, sex, race and ethnicity; 1.22 after adding income, education, and body weight. Both remain significantly above 1.0.
Candidate mechanisms for the residual Southern excess include physical inactivity, dietary environment, healthcare access, insurance policy, historical disinvestment, occupational patterns, chronic stress, discrimination, and institutional racism. The evidence does not permit attribution to any single mechanism.
The Prevention-Treatment Mismatch
Probst et al. (2024) classified counties in the highest diabetes-prevalence quartile at 14.4% or above. In those counties, the system was more than twice as likely to provide dialysis than structured diabetes education, exemplifying the inverse care law: communities with the greatest preventive need often have the least preventive capacity.
Source: Probst et al. (2024). High prevalence = top county quartile, 14.4% or higher. Even among counties with both a hospital and an FQHC, 42.9% lacked DSMES.
This pattern also reflects financing: dialysis has a stable reimbursement pathway and high per-patient expenditure, whereas DSMES generates comparatively limited revenue and cannot always be billed alongside other services.
Clinical and Operational Consequences
| System domain | Expected pressure in high-prevalence rural counties |
|---|---|
| Primary care | Higher panel complexity, medication titration, laboratory surveillance, and care-management demand |
| Emergency and inpatient care | Greater risk of hyperglycemic crises, infection, cardiovascular events, and avoidable admissions |
| Renal services | Increased CKD surveillance, nephrology referral, dialysis demand, and transportation burden |
| Imaging and diagnostics | Greater need for vascular imaging, cardiac evaluation, renal assessment, and complication workup |
| Ophthalmology | Retinal screening pathways, including tele-retinal models where specialist access is limited |
| Wound and vascular care | Foot surveillance, podiatry, limb-preservation protocols, and rapid referral for ischemia or infection |
| Pharmacy and finance | Medication affordability, insulin access, GLP-1 coverage, prior authorization, adherence, and benefit design |
| Workforce | Greater reliance on nurses, pharmacists, dietitians, community health workers, and remote specialist support |
Composition, Context, and Causality
Crude prevalence answers an operational question: how many people in a defined population are living with diagnosed diabetes? Adjustment answers a different etiologic question: would a rural-urban difference remain if populations were statistically comparable? Both are legitimate, but they should not be substituted for one another. Executives must plan for crude burden because that burden determines staffing, utilization, and cost.
Crude estimates
Measure real population burden. In 41-state BRFSS data, 19 states showed a significant unadjusted rural excess.
Adjusted estimates
Full adjustment eliminated the national rural-urban difference and most state differences (Khavjou et al., 2025). Dugani et al. (2024) found adjustment for county conditions reversed or attenuated the association. These findings identify modifiable pathways, not absolution.
Rurality as a Container of Cumulative Disadvantage
Rurality is analytically useful when treated as a container for linked exposures: labor-market decline, population loss, aging, educational opportunity, transportation distance, food retail structure, broadband access, hospital fragility, and workforce scarcity. It is analytically misleading when treated as a single binary cause. The practical unit of intervention is the county or multi-county system of risk and resources, not “the rural patient” in the abstract.
The Spatial Scale Problem
County maps can both reveal and conceal. Counties are meaningful administrative units, but they are internally heterogeneous: a county seat may hold a hospital, pharmacy, and grocery store while residents at the periphery travel an hour or more, and patients cross county and state borders for care. Refined analysis should include travel time, referral flows, telehealth reach, road networks, and cross-border utilization.
Diagnosed Prevalence Is Not Total Disease Burden
Source: CDC (2026). More than one-quarter of adults with diabetes are undiagnosed. Because diagnosis requires access, county estimates may reflect both disease occurrence and detection capacity.
Economic Materiality
Source: Parker et al. (2024).
Disproportionate indirect burden
Diagnosed diabetes cost the United States an estimated $412.9 billion in 2022: $306.6 billion in direct medical costs and $106.3 billion in reduced productivity and other indirect costs. Rural communities may experience a disproportionate indirect burden because chronic disease affects a smaller labor pool, informal caregiving capacity, transportation, and the financial stability of local employers and hospitals.
The map is best understood as a demand signal. Demographics explain much of that signal, regional context modifies it, and prevention infrastructure often fails to match it. The appropriate response is a place-based operating model that combines clinical care, public health, financing, and community infrastructure.
Strategic Implications for Healthcare Leadership
Six strategic moves translate the evidence into an operating response, moving from intelligence to network design, prevention capability, prevention timing, therapeutic equity, and financing reform.
Executive Dashboard: Recommended Measures
| Domain | Core measures |
|---|---|
| Population burden | Diagnosed prevalence; estimated undiagnosed burden; prediabetes; incidence; age at diagnosis |
| Control | A1c testing; A1c control; blood pressure; statin use; kidney-protective therapy; medication adherence |
| Complications | Emergency visits; preventable admissions; CKD progression; dialysis starts; amputations; retinal disease |
| Access | Primary care, DSMES, pharmacy, retinal screening, podiatry, nephrology, endocrinology, travel time |
| Equity | Outcomes by rurality, county, age, payer, income, race and ethnicity, and food security |
| Financial | Total cost of care; avoidable utilization; payer mix; uncompensated care; program contribution margin |
| Experience | Patient-reported access, treatment burden, affordability, transportation, digital usability, trust |
Rural Diabetes System Readiness Index
A proposed 100-point index creating a disciplined, repeatable assessment of whether local capacity is proportional to documented need. The index is conceptual and has not been psychometrically validated; weights are strategic starting points requiring local validation.
Readiness = 0.20(D) + 0.20(C) + 0.20(E) + 0.15(M) + 0.15(S) + 0.10(G)
where each domain is scored 0–100: D = detection, C = clinical capacity, E = education and nutrition, M = medication and monitoring, S = structural supports, G = governance.
| Domain | Weight | Illustrative evidence |
|---|---|---|
| Early detection and risk identification | 20% | Screening reach, prediabetes identification, risk stratification, community access points, closed-loop follow-up |
| Primary care and clinical capacity | 20% | Panel capacity, same-week access, care management, specialist escalation, kidney and cardiovascular protection |
| DSMES and nutrition access | 20% | Accredited program reach, referral completion, dietitian access, culturally and linguistically appropriate delivery |
| Medication and monitoring access | 15% | Affordability, pharmacy access, prior authorization, glucose monitoring, remote monitoring, adherence |
| Transportation, food, and broadband supports | 15% | Travel time, nonemergency transport, food resources, device access, broadband reliability, digital literacy |
| Data integration and governance | 10% | County-level analytics, disparity measures, accountability, community partnerships, continuous improvement |
Interactive County Readiness Calculator
Score each domain from 0 to 100 for a county or service area. The weighted index, classification band, and domain profile update in real time.
Material gaps in prevention, clinical access, or structural supports require an executive remediation plan.
Interpretation Bands
| Score | Classification | Executive interpretation |
|---|---|---|
| 80–100 | High readiness | Capacity is broadly aligned with need; focus on outcomes, equity, and continuous optimization |
| 60–79 | Moderate readiness | Core services exist, but reach, integration, or subgroup performance remains incomplete |
| 40–59 | Low readiness | Material gaps in prevention, clinical access, or structural supports require an executive remediation plan |
| 0–39 | Critical readiness deficit | The delivery system cannot reasonably meet current burden; regional partnership and external investment are urgent |
Need-Readiness Matrix
The index should never be interpreted without a need measure. Set the county’s need level (prevalence and projected demand) below; the marker combines it with your calculated readiness score to place the county in its priority quadrant.
High need / Low readiness
Highest priority: prevalence, projected demand, and service deficits converge
High need / High readiness
Focus on outcomes and efficiency
Low need / Low readiness
Regional access safeguards rather than full local duplication
Low need / High readiness
Maintain and share capacity regionally
Phased Action Agenda
A 36-month sequence moving from baseline intelligence to a regional operating model to scaled, financed, and evaluated programs.
Establish the Baseline
0–6 months- Download county-level prevalence data and produce separate all-county, metro, and nonmetro maps for the service region
- Create a local diabetes burden profile using age, income, education, race and ethnicity, food security, obesity, insurance, and travel-time data
- Inventory primary care, DSMES, pharmacy, retinal, podiatry, nephrology, endocrinology, dialysis, transportation, and broadband capacity
- Calculate a preliminary Readiness Index and identify high-need, low-readiness counties
- Adopt an executive dashboard with baseline measures, disparity stratification, and accountable owners
Build the Regional Operating Model
6–18 months- Create a hub-and-spoke DSMES and care-management network across hospitals, rural health clinics, FQHCs, pharmacies, and community organizations
- Implement closed-loop electronic referrals for DSMES, retinal screening, kidney evaluation, podiatry, and vascular care
- Deploy community health workers and mobile screening to reach working-age adults and communities distant from clinical sites
- Expand pharmacist-led medication management, remote glucose monitoring, tele-endocrinology, and tele-nephrology where broadband and digital support are adequate
- Negotiate payer support for prevention, remote monitoring, transportation, and navigation, with explicit rural risk adjustment
Scale, Finance, and Evaluate
18–36 months- Tie county-level outcomes to value-based contracts, community-benefit strategy, and capital planning
- Evaluate reach, clinical outcomes, avoidable utilization, patient experience, equity, and return on investment
- Use stepped-wedge or matched-county designs where feasible to distinguish program effects from secular trends
- Scale successful models while preserving local adaptation and trusted community relationships
- Reassess need and readiness annually, with public reporting where appropriate
90-Day Executive Decision List
Track leadership commitments interactively. Progress is session-only and resets on reload.
Limitations, Research Priorities, and References
The evidence supports three simultaneous conclusions: rural adults experience a materially higher crude prevalence, much of the difference is explained by demographic and modifiable contextual conditions, and the nonmetro South retains an excess after adjustment. The most actionable finding is the prevention-capacity mismatch, which finances the consequence more reliably than the prevention.
Limitations of the Current Synthesis
The supplied map is a screenshot with categorical bands; the underlying county CSV, denominators, uncertainty intervals, and modeling metadata were not included.
The screenshot uses the “All” filter, so visual interpretation cannot distinguish metro from nonmetro counties without additional views or data extraction.
Most national studies use self-reported diagnosed diabetes, omitting undiagnosed disease; differential detection can bias geographic comparisons.
Crude and adjusted results answer different questions, and the cited studies differ in years, rurality definitions, samples, and statistical models.
County residence is an imperfect measure of actual access: patients cross county borders and within-county travel times vary greatly.
The proposed readiness index is a strategic framework, not a validated measurement instrument; domain weights and thresholds require empirical testing.
The report synthesizes population-level evidence and should not be used to infer individual risk or to stigmatize rural communities.
Research Priorities
| Priority | Research objective |
|---|---|
| Longitudinal geography | Track county incidence, diagnosis, control, migration, and mortality rather than prevalence alone |
| Detection bias | Estimate undiagnosed disease using laboratory or modeled approaches and compare screening intensity by rurality |
| Travel-time access | Replace binary county availability with network travel time and cross-border referral analysis |
| Causal mechanisms | Examine food environment, physical activity, stress, discrimination, state policy, and healthcare market structure |
| Intervention evaluation | Test regional DSMES, community health worker, pharmacy, remote monitoring, and mobile-care models |
| Economic evaluation | Estimate county and system-level costs, avoided complications, productivity effects, and return on prevention |
| Readiness validation | Assess reliability, construct validity, predictive validity, and responsiveness of the proposed index |
References
Centers for Disease Control and Prevention. (2026). National diabetes statistics report.
cdc.govDugani, S. B., Lahr, B. D., Xie, H., Mielke, M. M., Bailey, K. R., & Vella, A. (2024). County rurality and incidence and prevalence of diagnosed diabetes in the United States. Mayo Clinic Proceedings, 99(7), 1078–1090.
DOIFarrigan, T., Genetin, B., Sanders, A., Pender, J., Thomas, K. L., Winkler, R. L., & Cromartie, J. (2024). Rural America at a glance: 2024 edition (EIB No. 282). USDA Economic Research Service.
DOIHill-Briggs, F., Adler, N. E., Berkowitz, S. A., Chin, M. H., Gary-Webb, T. L., Navas-Acien, A., Thornton, P. L., & Haire-Joshu, D. (2021). Social determinants of health and diabetes: A scientific review. Diabetes Care, 44(1), 258–279.
DOIKhavjou, O., Tayebali, Z., Cho, P., Myers, K., & Zhang, P. (2025). Rural-urban disparities in state-level diabetes prevalence among U.S. adults, 2021. Preventing Chronic Disease, 22, 240199.
DOIOnufrak, S., Saelee, R., Zaganjor, I., Miyamoto, Y., Koyama, A. K., Xu, F., Pavkov, M. E., Bullard, K. M., & Imperatore, G. (2024). Prevalence of self-reported diagnosed diabetes among adults, by county metropolitan status and region, United States, 2019–2022. Preventing Chronic Disease, 21, 240221.
DOIParker, E. D., Lin, J., Mahoney, T., Ume, N., Yang, G., Gabbay, R. A., ElSayed, N. A., & Bannuru, R. R. (2024). Economic costs of diabetes in the U.S. in 2022. Diabetes Care, 47(1), 26–43.
DOIProbst, J. C., Yell, N., Benavidez, G. A., McNatt, M. K., Browne, T., Herbert, L., Zahnd, W. E., & Crouch, E. (2024). Dialysis more available than patient education in counties with high diabetes prevalence. Preventing Chronic Disease, 21, 240052.
DOIRural Health Information Hub. (2024). Map of diagnosed diabetes prevalence, 2024.
ruralhealthinfo.orgRural Health Information Hub. (2026). Rural data visualization help.
ruralhealthinfo.orgYell, N., Asiedu-Danso, M., Odahowski, C. L., Crouch, E., & Benavidez, G. A. (2026). Geographic and sociodemographic patterns in prevalence of diagnosed diabetes, U.S., 2021–2024. Preventing Chronic Disease, 23, 250288.
DOIData note: all quantitative values in figures and tables were transcribed from the cited peer-reviewed publications and federal sources. Figures are original visual syntheses prepared for this report; no independent reanalysis of individual-level microdata or the underlying county CSV is claimed.