Protocol & scope
United States mental-health access — prevalence definitions, usable outpatient supply, IMD/acute capacity, crisis continuum, Medicaid rates, parity/ERISA, precedents, and the disability-awards seam with GBMT-12. Imports the gubment method (M1–M10): two-source rule on headline claims, kill conditions declared up front, one red team, one structurally blinded re-score, honesty box.
Eight workstreams executed 2026-08-11; scorecard pass 2 (blind-reconciled) the same day. Full markdown: research protocol · Phase 0 findings.
Starred anchors — priors stated before evidence
| # | Prior | Phase 0 / workstream result |
|---|---|---|
| 1 ★ | AMI ≈20–23%; SMI ≈5–6% | Holds (2024): AMI 23.4% (61.5M); SMI 5.6% (14.6M). DSM-IV model + functional impairment locked. |
| 2 ★ | No federal by-payer wait/acceptance series | Confirmed absence. KC1 fires. |
| 3 ★ | Psychiatrist Medicaid acceptance <50% | Supported: Bishop NAMCS 43.1%; Brahmbhatt secret-shopper 17.8% appointment success (4 cities). |
| 4 ★ | "95% beds gone since 1955" citogenic | Confirmed single-root TAC/state-hospital arithmetic. KC2 fires. |
| 5 ★ | Contemporary beds ≈10–15/100k | State-hospital staffed ≈11.7/100k (TAC 2016) — construct-limited. |
| 6 ★ | IMD still blocks most adult MH FFS stays | Holds. §1115 SMI/SED patchwork, not national repeal. |
| 7 ★ | 988 volume roughly doubled | Direction holds; prefer Jan ~2.5× answered or GAO +87% routed. Diversion unproven. |
| 8 ★ | Mental disorders ≈⅓ of working-age awards | Corrected: 12.7% of 2023 disabled-worker awards. Stock → GBMT-12. |
Workstream findings
§2 Licensed ≠ available
H2 Supported on two methods; KC1 carried. Flagship usable-supply wedge.
ws02-licensed-available.md§3 IMD & acute
Construct A/B locked; H3 Supported on statute/CMS; no 1955 absolutes.
ws03-imd-acute.md§4 Crisis / 988
H4 Supported — volume up; multi-state diversion causal series not located.
ws04-crisis-988.md§5 Medicaid rates / CCBHC
H5 Supported; CCBHC access package verified with DY4 softening disclosed.
ws05-medicaid-rates.md§6 Parity / ERISA
H6 Supported; KC3 fires — commercial O4 band-only for self-funded majority.
ws06-parity.md§8 Political economy / seam
H8 Not supported as written — awards 12.7%, not ~⅓.
ws08-political-economy-seam.mdScorecard — pass 2 leadership
#5 crisis continuum leads W1; #2 CCBHC leads W2 and W3; #10 do-nothing last under every weighting. Asylum-scale #4 never leads. Full matrix and honesty box: ws09-scorecard.md. Short publish-order note: ws10-sequencing.md.
Red team & blind re-score
Red team (pass 1): #2 O1 5→4, O2 4→3; #6 O1 4→3; measurement row named missing. ws09-red-team-log.md.
Structurally blinded re-score: 7 cells moved, leadership held. ws09-rescore-log.md · raw blind scores.
All research documents
Generated pages for every markdown file in the filing: All mental health research documents →