# The Mental Health Access Gap is Two Problems, Not One
**A Trellison Institute Working Paper — Executive Brief**
*May 2026 · 2-page summary of `mh_gap_article_v1_peer_review.md`*
---
## The finding
We joined the CDC's PLACES tract-level mental health prevalence (78,815 census tracts, 244.5 million American adults) to the CMS National Provider Identifier registry (102,036 licensed mental-health providers in 8,609 ZIP codes), then computed a drive-time-to-nearest-provider for every tract using Census Gazetteer 2024 ZIP centroids.
**The result splits the access conversation into two distinct problems with different policy answers.**
| Population | Drive time | Share of adults | Count | The problem |
|---|---|---:|---:|---|
| In-range | ≤ 30 minutes | **91.3%** | ~223M | Capacity, not distance |
| Edge | 30–60 minutes | 6.1% | ~15M | Mixed |
| Desert | > 60 minutes | **2.6%** | ~6.3M | Geographic access |
For **6.3 million Americans**, the gap is physical: no provider is reachable in under an hour. This population clusters in three regions — the Texas-Mexico border, interior Alaska, and the rural Mountain West. Brewster County, Texas is 5 hours 21 minutes from the nearest provider. The policy lever is geographic: telehealth, mobile crisis teams, the federal Mental Health Professional Shortage Area program, and HRSA loan-repayment incentives.
For **238 million Americans**, the gap is not distance — the gap is the appointment. The provider is close, the insurance card works on paper, but availability doesn't exist. The most striking signal in this dataset is the **college-town pattern**. In Cabell County, West Virginia (Marshall University), the nearest psychiatrist is 0.7 minutes away by car. Adult frequent mental distress: 36.9%. Eight college towns across eight states show the same pattern — high need, immediate supply, structural unavailability. Lafayette County, Mississippi (Ole Miss). Forrest County (Southern Miss). Strafford, New Hampshire (UNH). Black Hawk, Iowa (UNI). Burlington, Vermont (UVM). Lincoln, Nebraska. Burlington and Strafford share the pattern with the West Virginia and Mississippi towns — geography is not the variable; capacity is.
## What works
The residual analysis — within-state OLS of log-gap-ratio against insurance coverage, with z-score classification — surfaces the inverse: places where the gap is *smaller* than the demographic model predicts. The top of that list is concentrated in two policy regimes.
**Medicaid expansion + CCBHC payment model.** Wayne County, Michigan (Detroit). Kent County, Michigan (Grand Rapids). Jackson County, Missouri (Kansas City). Fairfax County, Virginia. Each of these counties has structural access deficits on paper that the residual analysis shows are *not* translating into the worst outcomes. Michigan accepted federal Medicaid expansion in 2014; Missouri voters approved it in 2020. Detroit and Grand Rapids host federally-designated Certified Community Behavioral Health Clinics — a payment model that compensates providers for taking Medicaid patients on real terms.
The pattern across negative outliers is consistent: **state Medicaid expansion plus a sustainable payment model for community mental health closes the gap**.
## Why this matters
The public-health conversation around mental health access treats "the gap" as a unified problem. The data does not. Conflating physical distance and provider availability leads to policy mismatches — telehealth investments that don't reach the 6.3M who need them, network adequacy enforcement that doesn't address the 238M who can't get an appointment.
**Separating the two clarifies the levers**:
- For the **6.3M geographic gap**: HRSA HPSA program expansion, telehealth parity laws, mobile crisis funding, NHSC loan repayment.
- For the **238M capacity gap**: BHWET workforce expansion (HRSA Behavioral Health Workforce Education and Training), CCBHC payment model expansion to all states, Medicaid network adequacy enforcement, parity law enforcement under MHPAEA.
## Method, in one paragraph
Tract-level prevalence from CDC PLACES (BRFSS small-area estimates, peer-reviewed). Provider supply from CMS NPPES (May 2026 pull) across five mental-health taxonomies. ZIP-centroid drive-time proxy: haversine distance × 1.4 road-multiplier, urban (35 mph) and rural (55 mph) speed splits by tract population, neighbor-state catchment. Gap ratio = (need × 1000) / state-level provider density per 100K adults. Residual regression within-state with z-score outlier classification at ±1.5σ. Population-weighted dartboard sampling stratified by residual class for case-study tracts. Full reproducibility package, including the parameterized `atlas.need_vs_access_framework_v1` tool, in supplementary materials.
## What's next
The framework is reusable. The same pipeline applied to other "need-vs-access" domains is in scope: poverty safety-net programs, English-language acquisition, jobs-to-job-seekers, postsecondary education, library access, police per capita, maternal care, dental care, broadband, oncology, crisis response. **The audit is the product.** The need-vs-access framework + Mental Health Gap study is Trellison demonstrating the methodology audit standard applied to its own original research.
## Reference
Full working paper: `docs/mh_gap_article_v1_peer_review.md` (~6,500 words, 10 sections).
Dataset: `analysis_outputs.mh_gap_tract_v1` (78,815 rows); supplementary CSV at `data-stories/mental_health_gap/dataset_v1/mh_gap_tract_v1.csv`.
Visualizations: `data-stories/media-assets/mental_health_gap/viz_v0/national_{drive_time,residual,dartboard}_*.mp4`.
**Methodology rating**: Trellison Institute, pending review.
**Evidence chain certificate**: LedgerWell, pending.
**Working paper version**: v1.0 draft.