Mapping the US health insurance gap.
Where should a healthcare team investigate coverage barriers first? We used Census estimates to compare the scale of the uninsured population across states, then added income context to identify where deeper access research could be useful.
People in the defined under-65 population
Estimated people without health insurance
Uninsured share, derived from state totals
State-level geographies compared
Counts above are sums of state estimates. The population is the SAHIE under-65 poverty universe, not the full US population.
Choose where to investigate coverage barriers first.
Imagine a healthcare organization planning a coverage outreach pilot with resources to work in only a few locations. Choosing the state with the highest uninsured rate answers one question: where is the coverage gap most widespread? Choosing the state with the largest uninsured population answers another: where are the most people affected? This analysis compares both measures so the team can build a geographic shortlist before researching local barriers and delivery options.
Who is included?
People under age 65 across all sexes, races, ethnicities, and income levels, in 50 states and Washington, DC. The denominator follows the Census SAHIE poverty universe: the population for whom poverty status is determined. It is not every US resident under 65. The income comparison narrows this same group to people at or below 200% of the poverty threshold.
What you can explore
Two maps distinguish the prevalence of being uninsured from the number of people affected. A state explorer shows the population denominator, uninsured count, rate, and uncertainty. An income comparison adds context. Together, these outputs support a geographic research shortlist; they do not identify individual people or determine benefit eligibility.
A high rate and a large population tell different stories.
How widespread is the gap?
Uninsured people as a share of the defined population.
How many people are affected?
Estimated number of uninsured people in each state.
Texas has the highest state uninsured rate, at 19.1%, and the largest uninsured population, about 5.02 million. California ranks third by uninsured count, despite a lower rate of 6.9%. Population scale changes the planning question.
For example, Oklahoma has an uninsured rate of 14.4%, compared with 6.9% in California. California nevertheless has about 2.25 million uninsured residents, compared with about 476,000 in Oklahoma. A team focused on the prevalence of coverage gaps may prioritize differently from a team planning the capacity needed to serve a large population.
Darker-to-lighter colors indicate larger values. DC is included in the data explorer; its small footprint is difficult to see at this map scale. These are descriptive comparisons, not tests of statistical significance.
Start with scale, then check local feasibility.
of the uninsured population is in three states.
Texas, California, and Florida together account for roughly 9.7 million uninsured people in this dataset. A national outreach plan could start by investigating these states, then use county estimates and local evidence to narrow its focus.
The next step would be to compare counties within those states and speak with local organizations and residents. Can the team deliver services there? Which languages and communication channels matter? What prevents people from getting or keeping coverage? The state totals help choose where to investigate; local research determines whether a pilot is practical.
Add income context before designing outreach.
All income levels
At or below 200% of the poverty threshold
The lower-income group includes 74.7 million people, of whom 12.3 million are uninsured. It represents 27.4% of the defined population but 45.9% of its uninsured residents.
For outreach planning, this is a reason to examine affordability and coverage navigation in the next research phase. It is not evidence that those barriers explain every uninsured person’s situation. Interviews and local program information would be needed to understand the reasons behind the pattern.
Both percentages are calculated from summed counts. Income differences describe an association; they do not establish why someone is uninsured or whether they qualify for a particular program.
A population definition behind every number.
Rate margins of error use the Census 90% confidence level.
A shortlist, not an automatic answer.
The analysis gives a team three starting points: states with high uninsured rates, states with large uninsured populations, and places where the income comparison warrants a closer look. The right shortlist depends on the pilot’s purpose and capacity.
If the goal is broad population reach
Investigate Texas, Florida, and California first because they account for 36.4% of the uninsured people in this defined population. Compare counties and existing service capacity before committing resources.
If the goal is to understand concentrated coverage gaps
Use high-rate states such as Texas, Oklahoma, and Georgia as candidates for deeper research. Examine county variation and recruit residents for interviews about their coverage experiences.
These are illustrative planning paths. The analysis does not show how many people can be contacted, enrolled, or converted into customers.
Turn the findings into a focused research plan.
Plan coverage outreach
Use rate and count together to shortlist places for coverage education or enrollment support research. Validate program eligibility, local partners, and delivery capacity before launching.
Design access research
Select contrasting geographies for interviews about affordability and coverage barriers. Recruit participants directly; state estimates cannot identify individual patients or their experiences.
Prioritize deeper geographic analysis
Move from state patterns to county SAHIE estimates. Add service locations, transportation, and local demographic context to understand where a coverage gap may coincide with an access gap.
Build a repeatable monitoring view
Keep the population, filters, and definitions consistent across annual releases. Evaluate uncertainty before labeling a change as meaningful.
What can you conclude from these data?
Does the uninsured count tell us who qualifies for benefits?
No. Eligibility depends on the program and individual circumstances. SAHIE estimates describe insurance coverage for population groups; they do not identify eligible individuals.
Why focus on people under age 65?
This study uses the under-65 group supplied in SAHIE. It keeps the population definition consistent across states. People age 65 and above are outside the scope, so the findings should not be presented as estimates for all residents.
Can this become a county-level map?
Yes. SAHIE also provides county estimates. A county analysis would retain the same demographic and income definitions, then check missing values and margins of error before making local comparisons.
Does a high uninsured rate prove poor access to care?
No. Insurance coverage is one part of access. Provider availability, transportation, affordability, appointment capacity, and residents’ experiences need additional data or research.
How we built the analysis.
- Source: Census Bureau 2024 Small Area Health Insurance Estimates (SAHIE), released August 2026. SAHIE produces model-based estimates using ACS and additional inputs; these are not a direct enumeration of every uninsured person.
- Selection: State geography (geocat 40), under age 65 (agecat 0), all races (racecat 0), both sexes (sexcat 0). The main view uses all incomes (iprcat 0); the income view uses at or below 200% of the poverty threshold (iprcat 1).
- Measures: NIPR is the group population; NUI is the estimated uninsured count; PCTUI is the group's uninsured percentage. National summary percentages are derived from summed counts across 50 states and DC.
- Uncertainty: State rate and count margins of error are preserved in the downloadable data. Margins of error are not summed. No statistical significance is claimed for state rankings or differences.
- Scope: This analysis does not measure disease prevalence, willingness to buy, reachable audiences, benefit eligibility, or causal effects. People age 65 and above are outside this analysis.
This is an independent analysis of public data, created to demonstrate our approach to population analytics. The outreach scenario illustrates a possible use of the findings. No outreach program was launched or evaluated as part of this study.
View all state estimates
| State | Defined population | Uninsured | Uninsured rate | 90% rate MOE | ≤200% poverty rate |
|---|---|---|---|---|---|
| Alabama | 4,045,561 | 403,170 | 10.0% | ±0.3 pp | 15.8% |
| Alaska | 608,380 | 76,031 | 12.5% | ±0.7 pp | 21.9% |
| Arizona | 5,929,505 | 763,269 | 12.9% | ±0.4 pp | 20.5% |
| Arkansas | 2,450,638 | 278,142 | 11.3% | ±0.4 pp | 16.9% |
| California | 32,354,492 | 2,247,123 | 6.9% | ±0.1 pp | 10.9% |
| Colorado | 4,914,052 | 456,679 | 9.3% | ±0.3 pp | 16.6% |
| Connecticut | 2,881,150 | 200,524 | 7.0% | ±0.3 pp | 13.3% |
| Delaware | 799,746 | 65,990 | 8.3% | ±0.6 pp | 14.2% |
| District of Columbia | 576,620 | 27,877 | 4.8% | ±0.6 pp | 7.8% |
| Florida | 17,852,452 | 2,442,995 | 13.7% | ±0.2 pp | 20.8% |
| Georgia | 9,088,938 | 1,270,400 | 14.0% | ±0.3 pp | 22.0% |
| Hawaii | 1,190,344 | 48,504 | 4.1% | ±0.3 pp | 6.9% |
| Idaho | 1,617,673 | 178,927 | 11.1% | ±0.5 pp | 17.7% |
| Illinois | 10,196,344 | 828,292 | 8.1% | ±0.2 pp | 14.1% |
| Indiana | 5,577,037 | 502,456 | 9.0% | ±0.3 pp | 14.3% |
| Iowa | 2,568,781 | 165,300 | 6.4% | ±0.3 pp | 11.4% |
| Kansas | 2,392,792 | 241,618 | 10.1% | ±0.4 pp | 17.8% |
| Kentucky | 3,678,229 | 296,932 | 8.1% | ±0.3 pp | 12.4% |
| Louisiana | 3,631,605 | 333,790 | 9.2% | ±0.4 pp | 13.5% |
| Maine | 1,050,637 | 75,311 | 7.2% | ±0.5 pp | 11.6% |
| Maryland | 5,021,506 | 362,232 | 7.2% | ±0.3 pp | 13.7% |
| Massachusetts | 5,617,274 | 184,401 | 3.3% | ±0.2 pp | 6.4% |
| Michigan | 8,003,812 | 492,951 | 6.2% | ±0.2 pp | 9.6% |
| Minnesota | 4,660,683 | 278,902 | 6.0% | ±0.2 pp | 11.6% |
| Mississippi | 2,298,836 | 278,104 | 12.1% | ±0.5 pp | 17.4% |
| Missouri | 4,963,334 | 457,105 | 9.2% | ±0.3 pp | 15.3% |
| Montana | 877,995 | 97,609 | 11.1% | ±0.6 pp | 17.9% |
| Nebraska | 1,626,305 | 132,594 | 8.2% | ±0.4 pp | 14.5% |
| Nevada | 2,673,405 | 366,864 | 13.7% | ±0.5 pp | 21.2% |
| New Hampshire | 1,083,585 | 62,725 | 5.8% | ±0.5 pp | 11.8% |
| New Jersey | 7,638,856 | 686,755 | 9.0% | ±0.3 pp | 18.4% |
| New Mexico | 1,629,253 | 218,391 | 13.4% | ±0.6 pp | 18.0% |
| New York | 15,658,365 | 918,375 | 5.9% | ±0.2 pp | 9.2% |
| North Carolina | 8,775,652 | 899,415 | 10.2% | ±0.3 pp | 17.1% |
| North Dakota | 633,565 | 41,539 | 6.6% | ±0.5 pp | 12.0% |
| Ohio | 9,417,063 | 760,419 | 8.1% | ±0.2 pp | 13.2% |
| Oklahoma | 3,314,645 | 475,968 | 14.4% | ±0.4 pp | 21.0% |
| Oregon | 3,393,289 | 223,111 | 6.6% | ±0.3 pp | 10.9% |
| Pennsylvania | 10,089,772 | 729,152 | 7.2% | ±0.2 pp | 12.7% |
| Rhode Island | 865,232 | 46,214 | 5.3% | ±0.6 pp | 10.3% |
| South Carolina | 4,255,249 | 475,998 | 11.2% | ±0.4 pp | 17.8% |
| South Dakota | 722,259 | 71,147 | 9.9% | ±0.6 pp | 17.4% |
| Tennessee | 5,812,776 | 658,811 | 11.3% | ±0.3 pp | 18.7% |
| Texas | 26,257,677 | 5,024,269 | 19.1% | ±0.3 pp | 30.8% |
| Utah | 3,042,028 | 284,716 | 9.4% | ±0.4 pp | 16.7% |
| Vermont | 486,291 | 24,977 | 5.1% | ±0.6 pp | 8.6% |
| Virginia | 7,063,460 | 572,321 | 8.1% | ±0.3 pp | 14.6% |
| Washington | 6,537,336 | 512,682 | 7.8% | ±0.2 pp | 14.5% |
| West Virginia | 1,349,265 | 98,721 | 7.3% | ±0.5 pp | 11.0% |
| Wisconsin | 4,693,812 | 300,348 | 6.4% | ±0.2 pp | 11.7% |
| Wyoming | 463,384 | 57,898 | 12.5% | ±0.9 pp | 20.8% |
Sources
Census Bureau · SAHIE dataset and documentation
Original 2024 estimates · Census CSV ZIP
Map geometry · PublicaMundi US states GeoJSON
Analysis prepared October 4, 2026. Continental map uses an Albers equal-area projection. Alaska and Hawaii use separate insets. Map outlines are illustrative, not an official Census boundary product.
What population or access question should your data answer?
Share your question, available data, and timeline. We will propose a focused analytical scope.
Get a quote →