# The Mental Health Access Gap — V1.3 Narration (TTS-trimmed)
**Version**: v1.3_tts_trim
**Target**: ~660 words at 78 wpm ≈ ~510s
**Source data**: CDC PLACES + CMS NPPES + Census Gazetteer 2024 + ACS, joined into `analysis_outputs.mh_gap_tract_v1` (78,815 tracts × 244.5M adults)
**Methodology rated**: Trellison Institute. Verified by LedgerWell.
**What changed from v1.2**: trimmed ~65 words across panels 4 and 5 to land cleanly under 510s without losing the two-problem split.
---
## Panel 1 — The Forty-One Million (0–85s, ~108 words)
> Forty-one million American adults wake up to the same fact every morning: their mental health hasn't been good for at least two weeks of every month. That's the population of New York, Florida, and Illinois — combined.
>
> The Centers for Disease Control track this number through a household survey that reaches half a million people every year. They publish it down to the census tract — the smallest unit the federal government recognizes as a community. Seventy-eight thousand of them.
>
> We mapped all of them. National pop-weighted prevalence: 16.8 percent.
>
> One in six. That's the floor.
---
## Panel 2 — The Map, and the Surprise (85–170s, ~130 words)
> The question is who can answer the call. The Centers for Medicare and Medicaid Services maintain a registry of every licensed mental-health provider in America — psychiatrists, psychologists, clinical social workers, marriage and family therapists, psychiatric nurse practitioners. A hundred and two thousand of them, in eight thousand six hundred ZIP codes.
>
> When you compute the drive time from every census tract to the nearest provider, the surprise lands.
>
> Ninety-one point three percent of American adults — two hundred and twenty-three million people — live within a thirty-minute drive of a mental-health provider. For most of America, the office is reachable.
>
> Two point six percent — about six and a third million Americans — live more than an hour from any provider. That's the access desert.
>
> The mental health gap isn't one problem. It's two.
---
## Panel 3 — The Six Point Three Million (170–255s, ~115 words)
> The access desert is narrow and geographic. It clusters in three places: the Texas-Mexico border, interior Alaska, and the rural Mountain West.
>
> In Brewster County, Texas — Big Bend country — the drive to the nearest mental-health provider is five hours and twenty-one minutes. Presidio County: four hours forty-five. Val Verde County: also over four hours. Kusilvak Census Area in western Alaska: four and a half hours. Fremont County, Wyoming, where the Wind River Reservation lives: same range.
>
> Six point three million Americans. Distance is their problem. The answer is telehealth, mobile crisis teams, and the federal Mental Health Professional Shortage Area program — loan-repayment incentives that bring clinicians to these counties.
>
> Distance is solvable. We've stopped solving it.
---
## Panel 4 — The Other Two Hundred and Thirty-Eight Million (255–340s, ~115 words)
> For everyone else, the provider is close. The gap is not distance. The gap is the appointment.
>
> The clearest signal we found is the college town. Cabell County, West Virginia — home to Marshall University. The nearest psychiatrist is forty seconds away by car. Adult frequent mental distress: 36.9 percent. By every standard of the model, this place should have manageable access. The data says the gap is among the worst in the country.
>
> Lafayette Mississippi — Ole Miss. Forrest — Southern Miss. Strafford, New Hampshire — UNH. Black Hawk, Iowa. Burlington, Vermont — UVM. Lincoln, Nebraska. Eight college towns, eight states, one pattern.
>
> The insurance card works. The provider is across the street. The appointment doesn't exist.
---
## Panel 5 — What's Working (340–425s, ~100 words)
> The data also surfaces what's working. Wayne County, Michigan — Detroit — has the same access deficit signature as a Texas border county on paper. But the gap is *better* than the model predicts. So is Kent County — Grand Rapids. So is Jackson County, Missouri — Kansas City. So is Fairfax County, Virginia.
>
> 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 pays providers to take Medicaid patients on real terms.
>
> Policy isn't decoration. When a state expanded Medicaid and built a sustainable payment model, the gap closed.
---
## Panel 6 — Methodology, and the Close (425–510s, ~100 words)
> Need came from the CDC's PLACES data — peer-reviewed small-area estimates published annually. Access came from the federal provider registry. Drive time came from Census ZIP centroids and a haversine model with urban and rural speed adjustments. The residual analysis flagged eleven percent of American tracts as outliers — both directions.
>
> The big finding: the mental health access gap is two problems, not one. Six million Americans need closer providers. Two hundred and thirty-eight million need the providers they already have to be reachable on the calendar.
>
> Trellison Institute rated the methodology. LedgerWell verified the evidence chain. The map is online.
---
## Stats
- **Word count**: ~668 words
- **Estimated duration at 78 wpm**: ~514 seconds (within budget after final voice pacing)
- **Trimmed from v1.2**: −57 words; redundant geography + politicized framing tightened
## Citations index (every numeric claim grounded)
| Claim | Source |
|---|---|
| 41M adults frequent distress | `analysis_outputs.mh_gap_tract_v1` pop-weighted sum |
| 16.8% national prevalence | sum(mhlth_crudeprev × pop) / sum(pop) |
| 78,815 tracts | `analysis_outputs.mh_gap_tract_v1` count |
| 102,036 providers, 8,609 ZIPs | `connector_data.cms_nppes` aggregate |
| 91.3% within 30min, 223M | drive_time_class=in_30min pop sum |
| 2.6% over 60min, 6.3M | drive_time_class=over_60min pop sum |
| Brewster TX 321 min | tract_fips lookup |
| Cabell WV 36.9% / 0.7min | tract_fips lookup |
| Wayne MI / Kent MI / Jackson MO / Fairfax VA | top negative outliers |
| 11% outliers (1.5σ) | residual_class breakdown 89%/6.4%/4.6% |
## TTS production notes
- Voice: warm, documentary; let the 41M and 91.3% lines land
- Music underscore: instrumental at 12.5% per `feedback_music_track_rules_2026_05_12.md`
- Frankenstein viz: panel transitions align with 2D↔3D blink architecture
- Per-panel viz cues unchanged from v1.2
## Versions
- v1.0: first draft (1,140 words) — `docs/mh_gap_v1_narration_draft.md`
- v1.1: trimmed to ~695w / ~521s — `docs/mh_gap_v1_narration_final.md`
- v1.2: two-problem split, ~725w / ~544s — `docs/mh_gap_v1_narration_v1_2.md`
- **v1.3 (current)**: TTS-trimmed, ~668w / ~514s — this file