Investigate America’s Declining Healthspan

Overview

U.S. life expectancy recovered to 79.0 years in 2024, but it remained about two years below the OECD average despite the United States devoting approximately 18% of GDP to healthcare. Avoidable mortality also remained unusually high compared with peer nations.

Life expectancy alone does not measure whether those additional years are lived in good health. The more useful concept is healthspan: the portion of life spent free from serious disease, disability, or functional limitation.

The central research question is:

Why are Americans spending so many years with chronic disease, disability, and reduced functional capacity, and which interventions produce the greatest gains in healthy life expectancy?

The investigation should distinguish:

• Lifespan: expected years alive;
• Healthspan: expected years lived in good health;
• Disability-free life expectancy: years without activity-limiting disability;
• Healthy life expectancy: life expectancy adjusted for disease and disability;
• Morbidity compression: illness is delayed into a shorter period near the end of life;
• Morbidity expansion: people survive longer but spend more years with illness;
• Working-age health loss: disability or chronic disease that limits employment before retirement.

The recommended primary units of analysis are:

• State-year;
• County-year;
• Age-sex-population group;
• Disease-state-year;
• Individual-year where longitudinal data are available.

---

Action Items

Initial Evidence

Several findings justify the investigation:

• U.S. life expectancy reached 79.0 years in 2024, but remained among the lowest in the OECD.
• The United States spends more on healthcare per person and as a share of GDP than comparable countries while experiencing weaker mortality and access outcomes.
• Chronic illness, obesity, mental-health disorders, and preventable mortality remain major contributors to poor U.S. health outcomes. In 2022, approximately 42% of U.S. adults were affected by obesity.
• Recent commentary cautions that rising chronic-disease prevalence may reflect several mechanisms simultaneously: real health deterioration, population aging, improved diagnosis, and changing disease definitions.
• Disability has substantial labor-market consequences. In 2025, 22.8% of people with a disability were employed, compared with 65.2% of people without a disability.
• State and county health outcomes differ sharply, indicating that national outcomes are not inevitable and that local systems, environments, and policies may matter.
• Recent research describes a reversal in the historical relationship between rising income and longevity: during the 2010s, many U.S. places became richer without achieving corresponding life-expectancy gains.

These findings establish poor health performance but do not determine whether the main causes are healthcare, behavior, environmental exposure, economic insecurity, aging, diagnosis, or social conditions.

Hypotheses

H1 — Chronic disease is beginning earlier in life

Americans may be developing obesity, diabetes, cardiovascular risk, depression, chronic pain, and other conditions at younger ages.

Evidence to seek:

• Age of first diagnosis declines;
• Disease prevalence rises within the same age groups;
• Multiple chronic conditions become more common among working-age adults;
• Functional limitation begins earlier;
• Younger cohorts have worse health at the same age than previous cohorts.

H2 — Medical care extends survival without proportionally restoring function

Treatment may prevent death while leaving patients with disability or chronic impairment.

Evidence to seek:

• Mortality falls while disability prevalence remains stable or rises;
• Survival after major diseases improves without equivalent functional recovery;
• High-spending diseases produce limited gains in healthy life-years;
• Rehabilitation access predicts better functional outcomes;
• Regions with similar survival differ substantially in disability.

H3 — The United States underinvests in prevention

Resources may be concentrated on treating advanced disease rather than preventing risk factors and detecting disease earlier.

Evidence to seek:

• Preventive-service access predicts later healthspan;
• Primary-care shortages predict earlier disability;
• Counties improving blood pressure, smoking, activity, or diabetes control later gain healthy years;
• Public-health investment predicts lower preventable morbidity;
• Preventive spending produces delayed but measurable savings.

H4 — Obesity and metabolic disease are central drivers

Obesity, insulin resistance, diabetes, and related cardiovascular conditions may account for a large proportion of unhealthy years.

Evidence to seek:

• Metabolic conditions explain substantial geographic healthspan variation;
• Earlier obesity predicts later disability;
• Changes in food, activity, sleep, and medication use predict cohort differences;
• Effects remain after income and age adjustment;
• Interventions improve function rather than only biomarkers.

H5 — Mental illness increasingly reduces healthy and economically active years

Depression, anxiety, substance-use disorders, and severe mental illness may reduce work, physical activity, social connection, and long-term health.

Evidence to seek:

• Mental-health conditions precede labor-force exit;
• Treatment access predicts functional recovery;
• Mental and physical conditions interact;
• Effects are especially strong among younger adults;
• Mental-health burden explains part of the rise in disability claims or activity limitations.

H6 — Environmental and occupational exposures create concentrated health loss

Air pollution, heat, toxic exposure, unsafe work, noise, and poor housing may generate preventable disease and disability.

Evidence to seek:

• Exposure changes precede disease changes;
• Burden is geographically concentrated;
• Regulatory or remediation events improve outcomes;
• Exposure effects remain after demographic controls;
• Working-age disability is elevated in high-risk occupations and locations.

H7 — Economic insecurity damages healthspan

Housing instability, food insecurity, debt, unpredictable work, and lack of paid leave may cause illness or prevent effective disease management.

Evidence to seek:

• Economic shocks precede health deterioration;
• Stable housing or income interventions improve health;
• Cost-related medication nonadherence predicts complications;
• Paid-leave access predicts recovery;
• Effects are not explained solely by baseline health.

H8 — Weak primary care allows manageable conditions to become disabling

Patients without regular primary care may receive diagnosis and treatment later.

Evidence to seek:

• Primary-care access predicts fewer disability-producing complications;
• Continuity of care improves disease control;
• Rural closures precede worsening outcomes;
• Emergency-department dependence predicts lower healthspan;
• Benefits are larger in medically underserved areas.

The United States has relatively few primary-care physicians per capita compared with peer countries.

H9 — Long COVID contributed to working-age disability

Persistent post-COVID symptoms may have increased activity limitations and labor-force difficulties among selected groups.

Evidence to seek:

• Long-COVID prevalence predicts changes in disability and employment;
• Effects differ by occupation and remote-work feasibility;
• Vaccination and treatment history affect outcomes;
• Disability changes exceed pre-pandemic trends;
• Results remain after accounting for other health conditions.

Recent analysis found unequal long-COVID prevalence and activity limitation across demographic groups and less access to remote work among many workers with disabilities.

H10 — Increasing diagnosis exaggerates part of the apparent decline

Some measured growth in chronic disease may reflect better screening, broader definitions, or improved survival.

Evidence to seek:

• Diagnosis rises without equivalent change in symptoms or function;
• Biomarker trends differ from diagnosis trends;
• Definition changes create measurable discontinuities;
• Mortality and hospitalization do not move with recorded prevalence;
• Survey-based limitations provide different trends from medical records.

H11 — Healthspan varies more than lifespan across U.S. communities

Places with similar life expectancy may differ greatly in disability, chronic disease, independence, and quality of life.

Evidence to seek:

• Healthspan rankings differ from lifespan rankings;
• Some counties achieve long life with high disability;
• Positive-deviant regions achieve both longevity and functional health;
• Medical spending predicts lifespan and healthspan differently;
• Local environments explain part of the difference.

---

Research Plan

1. Define measurable healthspan outcomes

Measure:

• Healthy life expectancy;
• Disability-free life expectancy;
• Years lived without activity limitation;
• Years lived without major chronic disease;
• Self-rated healthy years;
• Quality-adjusted life expectancy;
• Working-life expectancy;
• Years lost to disability.

A basic measure using Sullivan’s method is:

Where:

• (L_x) is person-years lived in age interval (x);
• (\pi_x) is the proportion reporting unhealthy status or disability.

Report both the healthy years and the expected unhealthy years:

2. Establish national and geographic trends

Construct state- and county-level series from approximately 2000 onward.

Analyze separately:

• 2000–2009;
• 2010–2019;
• 2020–2022;
• 2023 onward.

Measure whether:

• Life expectancy increased;
• Disability-free expectancy increased;
• Unhealthy years expanded or compressed;
• Trends differed by age, sex, income, race, and geography.

3. Build a disease-burden dataset

Include:

• Cardiovascular disease;
• diabetes;
• obesity;
• cancer;
• chronic respiratory disease;
• musculoskeletal disorders;
• chronic pain;
• depression and anxiety;
• substance-use disorders;
• dementia;
• long COVID.

For each disease, estimate:

• Prevalence;
• incidence;
• age of onset;
• mortality;
• disability weight;
• medical spending;
• employment effects;
• healthy years lost.

4. Separate prevalence from functional impairment

A diagnosis does not necessarily imply disability.

Compare:

• Diagnosed condition;
• reported symptoms;
• activity limitations;
• work limitations;
• hospitalization;
• medication use;
• self-rated health.

This will help distinguish improved detection from genuine health decline.

5. Construct a county-level healthspan panel

Core variables:

• Mortality;
• disability;
• chronic disease;
• primary-care access;
• insurance;
• medical spending;
• income;
• education;
• housing;
• food insecurity;
• pollution;
• heat exposure;
• transportation;
• physical activity;
• smoking;
• social isolation;
• rurality.

Estimate:

Use lagged predictors because many exposures affect health over years.

6. Identify positive-deviant communities

Find counties or states with:

• Longer healthspan than expected;
• Lower disability at comparable age;
• Better outcomes at lower spending;
• Strong outcomes despite economic disadvantage;
• Rapid improvement over time.

Match them with comparable locations and investigate:

• Primary care;
• public health;
• built environment;
• food access;
• social connection;
• preventive care;
• economic stability.

7. Analyze working-age health loss

For adults ages 18–64, examine:

• Disability;
• labor-force participation;
• sickness absence;
• reduced hours;
• occupation changes;
• early retirement;
• Social Security Disability Insurance participation.

Estimate:

Adjust for age, education, occupation, and prior employment.

8. Compare spending with healthy years gained

For each disease or region:

This should not be interpreted mechanically. Some spending improves comfort, dignity, or rare-disease care without creating large population-level gains.

9. Evaluate policy and environmental interventions

Possible quasi-experiments include:

• Medicaid expansion;
• primary-care expansion;
• smoking restrictions;
• clean-air regulations;
• lead remediation;
• walkability improvements;
• nutrition benefits;
• paid sick leave;
• heat protections;
• vaccination programs;
• diabetes-prevention programs.

Use:

• Difference-in-differences;
• synthetic controls;
• interrupted time series;
• event studies;
• regression discontinuity where eligibility thresholds permit.

10. Analyze cohort deterioration

Compare generations at the same age.

For example:

This distinguishes population aging from younger generations becoming less healthy.

11. Build a Healthspan Dashboard

Display:

• Life expectancy;
• healthy life expectancy;
• unhealthy years;
• disability-free years;
• chronic conditions;
• working-age disability;
• preventable mortality;
• spending;
• primary-care access;
• environmental and social risk.

Always include uncertainty intervals.

12. Conduct robustness checks

Repeat analysis using:

• Different disability definitions;
• Self-rated health;
• activity limitations;
• disease-free life;
• quality-adjusted years;
• State and county units;
• Survey and administrative data;
• With and without pandemic years;
• Alternative age standardization;
• Different disability weights.

13. Produce the Final PowerPoint

Recommended structure:

1. Lifespan is not healthspan;
2. U.S. longevity and health spending;
3. Expected healthy and unhealthy years;
4. Chronic disease by age and cohort;
5. Working-age disability;
6. Geographic healthspan inequality;
7. Medical treatment versus prevention;
8. Environmental and economic conditions;
9. Primary-care access;
10. Positive-deviant communities;
11. Which hypotheses survived testing;
12. Highest-value interventions.

The presentation must distinguish:

• Diagnosis from disability;
• Longer survival from better health;
• Population aging from cohort deterioration;
• Medical spending from prevention;
• Association from causation;
• Life expectancy from healthy life expectancy.

---

Resources

Data, APIs, and Where to Get Data

CDC WONDER

Use for:

• Mortality;
• cause-specific death rates;
• age-adjusted mortality;
• population denominators;
• county and state comparisons.

CDC’s final 2024 mortality release provides current U.S. life-expectancy and cause-of-death benchmarks.

CDC PLACES API

Use for local estimates of:

• Chronic disease;
• disability;
• self-rated health;
• mental health;
• physical inactivity;
• obesity;
• smoking;
• preventive care.

PLACES provides county-, place-, tract-, and ZIP Code Tabulation Area estimates.

Behavioral Risk Factor Surveillance System

Use for:

• Self-rated health;
• activity limitation;
• disability;
• chronic disease;
• obesity;
• mental health;
• health behavior;
• access to care.

Survey weights and changing questionnaire modules must be handled carefully.

National Health Interview Survey

Use for:

• Functional limitation;
• disability;
• chronic conditions;
• pain;
• mental health;
• insurance;
• employment;
• healthcare access.

NHIS is one of the strongest sources for national healthspan estimation.

Medical Expenditure Panel Survey

Use for:

• Medical spending;
• chronic disease;
• disability;
• service use;
• prescriptions;
• insurance;
• health status.

MEPS can connect health conditions with annual costs and functional outcomes.

Health and Retirement Study

Use for longitudinal analysis of:

• Aging;
• disease onset;
• disability;
• cognition;
• employment;
• retirement;
• wealth;
• mortality.

This is especially useful for estimating transitions between healthy, disabled, and deceased states.

Current Population Survey Disability Supplement

Use for:

• Disability;
• employment;
• unemployment;
• occupation;
• labor-force participation.

BLS publishes annual disability labor-force statistics.

American Community Survey API and PUMS

Use for:

• Disability;
• employment;
• age;
• income;
• insurance;
• housing;
• transportation;
• county and state population controls.

CMS Data APIs

Use for:

• Medicare spending;
• chronic-condition prevalence;
• hospital use;
• geographic variation;
• preventable admissions;
• provider access.

National Health Expenditure Accounts

Use for national healthcare expenditure by:

• Service;
• payer;
• population;
• time.

Area Health Resources Files

Use for:

• Primary-care supply;
• hospitals;
• health centers;
• workforce;
• demographics;
• county-level resources.

Health Professional Shortage Area Data

Use for:

• Primary-care shortage;
• mental-health shortage;
• dental shortage;
• underserved areas.

EPA Air Quality System API

Use for:

• Fine particulate matter;
• ozone;
• pollutants;
• monitoring stations;
• exposure changes.

EPA EJScreen and Related Environmental Data

Use for environmental exposures and community vulnerability. Confirm current availability and methodology before implementation.

NOAA APIs

Use for:

• Heat;
• wildfire smoke conditions;
• weather extremes;
• climate exposure.

USDA Food Access Research Atlas

Use for:

• Food access;
• distance to supermarkets;
• vehicle access;
• neighborhood socioeconomic conditions.

Census Household Pulse Survey

Use for:

• Long COVID;
• mental health;
• food insecurity;
• healthcare access;
• activity limitations;
• economic stress.

Social Security Administration Data

Use for:

• Disability program applications;
• awards;
• beneficiaries;
• geographic patterns.

Administrative definitions do not capture all disability and are affected by program rules.

WHO Global Health Observatory API

Use for international comparison of:

• Healthy life expectancy;
• life expectancy;
• disease burden;
• mortality;
• risk factors.

OECD Health Statistics

Use for peer-country comparisons involving:

• Life expectancy;
• healthy life measures where available;
• spending;
• chronic disease;
• workforce;
• utilization.

Foundational Resources

• CDC, Mortality in the United States, 2024.
• Commonwealth Fund, U.S. Health Care from a Global Perspective, 2026.
• Washington Post, Chronic Disease Rates Are Growing. Here’s the Unexpected Story.
• BLS, People with a Disability: Labor Force Characteristics—2025.
• The U.S. Mortality Crisis as a Preston Curve Reversal.

Central Scientific Caution

A rise in chronic-disease prevalence does not necessarily mean that Americans are becoming less healthy at the same rate. Prevalence can rise because:

1. More people develop a disease;
2. Disease begins earlier;
3. Diagnosis improves;
4. Diagnostic definitions broaden;
5. People survive longer after diagnosis;
6. The population ages.

The strongest investigation will therefore compare:

• Disease prevalence;
• Incidence;
• age of onset;
• symptoms;
• functional limitations;
• mortality;
• survival;
• employment;
• medical spending.

The most important policy finding may be that the United States is succeeding at keeping people alive after disease develops while failing to delay disease onset or restore functional health.

Resources/Instructions

REPLACE THIS TEXT -If there is a website which has documentation that helps with this issue provide the link(s) here.

• If this issue requires access to 311 data, please answer the following questions:
  • Do you need a one-time or ongoing dump of the data?
  • Do you need subset of data (i.e. certain years) or the entire data set (approx. 4 million rows or 11 GB)?
    • If a subset is needed, please define subset characteristics (i.e. date range, etc.)
  • Do you need online access via an API or a download of data?