The Gap Index

The Gap Index scores every state on five weighted dimensions of emergency response capacity: destination risk (25%), EMS staffing (20%), fire-service volunteer dependency (20%), demand pressure from population aging (20%), and police staffing (15%). Higher scores mean wider gaps. Scores are relative, not absolute — running from 32.5 to 64.3. The index measures structure, not outcomes.

Download the brief — 16 pages, all fifty states ranked and profiled. The full methodology follows below.

How the Gap Index was built

The Gap Index measures structural capacity for emergency response across all fifty states. It combines five dimensions, each capturing a different way that the distance between a person in trouble and the help they need can widen. A higher score means a wider gap.

The index measures structure, not outcomes. It does not observe dispatch times, mutual aid agreements, road geometry, or the performance of any individual agency. Two states with identical scores may have very different actual response times. What the index describes is the standing capacity that exists before any call is placed.

Every input is public. Every weight is published below. Where a data source has a known limitation, that limitation is stated rather than smoothed over.

Design principle

Wherever it was possible to do so honestly, each dimension pairs a measure of resource with a measure of exposure — how much capacity exists, and how much territory or population that capacity has to cover. Volunteer Dependency combines firefighter staffing composition with rural population share. Destination Risk combines hospital closures and closure risk with rural population as the denominator. The remaining three dimensions are single-measure, because no defensible exposure pairing was available in public data at the state level.

The five dimensions

1. Volunteer Dependency — 20%

Eighty percent county-population-weighted volunteer and paid-per-call firefighter share, drawn from the USFA National Fire Department Registry; twenty percent rural population share from the 2020 Census. Actual firefighter headcounts were used rather than department-type classification, and paid-per-call staff were grouped with volunteers. The five New York City counties were merged into a single unit. Forty-five counties with no registered department were excluded rather than imputed, representing roughly 0.1 percent of national population. State population coverage ranges from 98.66 percent to 100 percent, with thirty-three states at full coverage. Range: 3.6 percent to 89.5 percent volunteer share.

The registry is self-reported by fire departments and updated continuously, with no fixed publication vintage. It is the weakest provenance of any source in this index, and it carries twenty percent of the weight. The coverage figures above are published for that reason.

2. EMS Staffing — 20%

Employed EMTs and paramedics per 100,000 residents, inverted so that fewer providers produces a higher gap score. Source is BLS Occupational Employment and Wage Statistics, May 2025, SOC codes 29-2042 and 29-2043, against 2025 Census population estimates. Range: 48.9 per 100,000 (Hawaii) to 152.0 (Vermont). A paramedics-only alternative was tested and rejected; it correlated with the combined measure at only 0.537, indicating a different construct rather than a cleaner one.

3. Police Staffing — 15%

Employed police and sheriff’s patrol officers per 1,000 residents, inverted. BLS OEWS May 2025, SOC 33-3051. Range: 1.14 per 1,000 (Washington) to 2.73 (Mississippi). This occupational code covers patrol officers and sheriff’s deputies; it excludes detectives, supervisors, and federal officers, and therefore produces lower counts than the FBI’s broader “sworn officer” definition. FBI Law Enforcement Employees data was evaluated as an alternative and rejected — reporting is voluntary, nine states fall below 90 percent population coverage, and two below 70 percent.

4. Destination Risk — 25%

Two components, each expressed per 100,000 rural residents. Historical closures (40 percent) come from the UNC Sheps Center rural hospital closure database, 2005 to present: 197 closures, tier-weighted so that the 165 facilities losing all emergency capability count fully and the 32 retaining emergency or urgent care count as half. Forward risk (60 percent) uses the Center for Healthcare Quality and Payment Reform’s count of hospitals at immediate risk of closing, May 2026. The rural population denominator is 2020 Census rural share applied to 2020 Census state population. The two components correlate at 0.54.

5. Demand Pressure — 20%

Forty percent share of population aged 65 and over; sixty percent share aged 85 and over. American Community Survey five-year estimates, 2020–2024, table S0101, with an effective midpoint near 2022. The heavier weight on the 85-plus share reflects that emergency call volume rises steeply in the oldest cohort. The two components are related but distinct, correlating at 0.736 when measured as shares — and the rank order between them flips for several states.

Combining the dimensions

Each dimension was normalized to a 0–100 scale using min-max normalization against the observed range across the fifty states. A score of 100 marks the state with the greatest gap on that dimension; a score of 0 marks the least. Normalized dimension scores were then combined using fixed weights:

Gap Index =

• 0.20 × Volunteer Dependency
• 0.20 × EMS Staffing
• 0.15 × Police Staffing
• 0.25 × Destination Risk
• 0.20 × Demand Pressure

Z-score normalization was tested and rejected as unnecessary: all five dimension standard deviations fall within a narrow band, from 16.9 to 23.1, so min-max does not distort relative spread. The composite was deliberately not re-normalized. Rescaling the final score to a full 0–100 range would erase the compression that is one of the index’s central findings.

Reading the scores

Gap Index scores run from 32.5 to 64.3. They are relative, not absolute. A score of 64 does not mean 64 percent of anything; it means Kansas shows the highest combined structural gap among the fifty states.

The scores compress toward the middle because the dimensions are largely independent. Composite standard deviation is 6.8, against dimension standard deviations of 16.9 to 23.1. A state that scores badly on one measure will usually score near the middle on others.

Ranks are shown for display only. No calculation in this index uses rank position — all arithmetic operates on scores. Where states tie on a dimension, they share a rank.

Independence and sensitivity

The five dimensions are largely independent. Of ten pairwise correlations among dimension scores, nine fall below |0.32|. The exception is volunteer fire dependency and measured EMS staffing, at −0.63.

That correlation reflects the single largest limitation in this index: occupational employment data does not capture volunteer EMS providers. States that rely heavily on volunteer fire services also tend to rely on volunteer or hybrid EMS, and those responders do not appear in BLS employment counts. Vermont illustrates the effect clearly — it ranks first on volunteer fire dependency and last on measured EMS staffing at the same time. The measured EMS figure is accurate as far as it goes; it simply does not go as far as the reality.

Rankings were tested under alternative weightings, including equal weights across all five dimensions. No state moved more than twelve positions. Kansas ranks first under the published weighting and second under equal weights, reflecting its concentration of rural hospital risk against median performance elsewhere.

No state ranks in the top ten on more than three of the five dimensions. Only Maine, Iowa, and Vermont reach three.

What this index does not show

The Gap Index describes structural conditions. It does not establish causation, and it does not predict individual outcomes. A high score does not mean a given emergency call will go unanswered, and a low score does not mean a community is well served. Local factors the index cannot see — mutual aid networks, dispatch quality, geography, traffic, hospital specialization — shape actual response as much as the structural measures here.

The index is a starting point for asking questions about a place, not a verdict on it.

Notes on individual states

Delaware scores zero on Destination Risk. This is a substantive zero, not missing data: Delaware has recorded no rural hospital closures since 2005, and the CHQPR methodology identifies no eligible isolated rural community hospitals in the state.

Massachusetts scores zero on both Destination Risk components. Verified as genuine — no rural hospital closures since 2005 and none currently at immediate risk.

Vermont reports 152.0 EMTs and paramedics per 100,000, the highest measured rate in the country and well clear of second place. The absolute count is 790 people. The figure was verified and retained without adjustment. See the note above on volunteer EMS.

Kansas shows 28 hospitals at immediate risk of closing, the highest count of any state, against a rural population of roughly 800,000. This figure anchors the top of the Destination Risk scale and was neither capped nor winsorized.

Sources and data vintages

Data compiled August 2026.