Date: 2026-08-11. Adjudicates H7 with equal effort against the bed-rebuild narrative (M3 steelman). Cross-cites §3 (IMD / Construct A–B beds; boarding), §4 (crisis continuum, not 988 diversion), §5 (CCBHC access), §6 (parity exams). Does not re-litigate KC2 “beds since 1955” citogenesis.
PRISMA-lite search log
| Field | Logged |
|---|---|
| Strings | IMD 1115 SMI waiver outcomes beds boarding; Kendra's Law AOT Swartz Swanson hospitalization selection; Laura's Law California DHCS legislative report AOT; CCBHC Mathematica access wait evaluation; Massachusetts DOI behavioral parity market conduct 8K; New York DFS 11 NYCRR 38 behavioral network adequacy; VHA Directive mental health same day access wait standard; OECD psychiatric beds per 1000 Japan Italy United States community care; Australia psychiatric bed occupancy readmission OECD; mobile crisis arrest diversion Swanson Michigan; BHCC walk-in crisis ED Kalb |
| Date range | Prefer 2009–2025 evaluations with methods; OECD/WHO bed ratios contemporary; AOT NY Duke/OMH as primary domestic natural experiment |
| Inclusion | Measured access/capacity/outcome evaluations; official parity exam statutes/reports; VA published access standards; peer-country bed + community mixes with explicit transferability notes. Exclude advocacy bed-shortage slogans without inventory construct; exclude 1955% claims. |
| Equal-effort counter-search | Peer systems with low beds + high acute pressure; boarding literature; TAC/NRI state-hospital shortage reports; arguments that community substitution failed |
| Searched | 2026-08-11 |
1. What H7 claims (steelman)
H7 — STEELMAN, unfashionable direction: Deinstitutionalization’s bed decline is real, but the binding deficit today is outpatient and crisis continuum, not a return to asylum-scale inpatient stock; jurisdictions that invested in community systems show better access without proportional bed restoration.
Equal effort means this workstream also steels the opposite: acute boarding, forensic state-hospital occupancy, and low-bed peer systems with high readmission / ED pressure still argue for some inpatient rebuild. The scorecard will weigh both; this file decides whether H7’s formal criteria fire.
2. Domestic instruments with measured effects
2.1 MH IMD §1115 waivers (financing, not proven bed rebuild)
Carry §3 / Phase 0: IMD exclusion remains the default Medicaid purchase constraint for adult IMD stays; SMI/SED §1115 patchwork (CRS IF10222 as of 2025-01-14) is not national repeal.
Measured capacity effect: McBain et al. (PLOS Medicine 2025, HCRIS) — MH IMD waiver states did not show statistically significantly higher Construct B hospital psych bed rates. Waivers address a financing rule; they are not yet evidence that the waiver geography rebuilt beds.
Transferability: High for other Medicaid financing instruments (repeal / broaden waiver). Low as a bed-stock lever until comparative boarding/bed evaluations appear.
2.2 CCBHC expansion states (community access without bed rebuild)
Carry §5 / anchor 9 (ASPE/Mathematica): adult time-to-eval 9.0→5.4 days (DY1→DY2); clients +~9%; 94% open-access/same-day reported; ED/hospital DID heterogeneous.
For H7: this is a community-capacity instrument with measured outpatient access gains and no requirement of asylum-scale bed restoration. It is one of the two evaluations H7 needs for criterion (b).
Transferability: High within Medicaid safety-net PPS states; payment-design perishable (GAO-21-104466 on PPS alignment).
2.3 Crisis continuum (mobile crisis / stabilization — not 988)
Carry §4 independence audit: 988 lacks multi-state causal ED/arrest diversion. Adjacent instruments do show community-tier gains:
| Instrument | Finding | Bed rebuild? |
|---|---|---|
| Swanson et al. Michigan mobile crisis (IPTW) | ~45% lower 11-mo arrest incidence vs LE-only | No |
| Kalb et al. BHCC walk-in (Health Serv Res) | Zip BHCC availability ↔︎ ~2.8% lower mean MBD ED utilization | No |
For H7: these are the second family of community-crisis evaluations showing access/acute-adjacent gains without large bed rebuilds. They steel architecture #5’s middle tier; they do not refute H4 on 988 itself.
2.4 AOT / Kendra’s / Laura’s — selection is the whole game
New York (Kendra’s Law) — best domestic identification stack:
Swartz et al., Psychiatric Services 2010 (NYS OMH + Medicaid; 3,576 AOT consumers, 1999–2007): vs pre-order period, initial 6-month order → hospital admission odds OR 0.77 (≈25% lower); renewal → OR 0.59; hospital-days odds also down (OR ~0.80–0.84). Companion Duke/OMH evaluation reports large pre/post drops in hospitalization, homelessness, and arrest — pre/post on selected eligibles.
Selection discipline (mandatory): AOT statutes select people with prior hospitalization, violence risk, or non-adherence. Pre/post improvements can reflect regression to the mean, service intensification that accompanies the order (ICM/ACT), and court leverage — not a population-average “involuntary outpatient cures boarding.” Score architecture #7 as targeted engagement for a selected SMI band, not as a substitute for rate floors or system-wide bed policy.
California (Laura’s Law): DHCS legislative reports (county-reported aggregates; e.g. 2022–23 window cited in state AOT materials) show large % drops in homelessness / hospitalization / LE contact among enrolled participants — same selection/pre-post structure; county opt-out and reporting inconsistency limit causal transfer. Treat as corroborating direction for the selected band, weaker than NY’s linked Medicaid claims design.
2.5 MA / NY parity enforcement (realized commercial access pressure)
| Jurisdiction | Instrument | What it shows |
|---|---|---|
| Massachusetts | M.G.L. c.26 §8K (Acts of 2022): DOI parity market-conduct exams ≥ every 4 years, NQTL focus including network adequacy and reimbursement; CY2023 parity summary documents exams of 21 carriers underway | State exam machinery that can see fully insured plans; open-exam confidentiality limits published findings this pass |
| New York | DFS 11 NYCRR 38 behavioral network adequacy / wait standards; 11 NYCRR 230.3 parity compliance programs; Insurance Law §343 biennial parity reports; OAG industry investigation lineage (2018 report) | Numeric wait/network rules + comparative-analysis duties for state-regulated plans |
Cross-cite §6 / KC3: these precedents strengthen fully insured O4 enforcement; they do not reach the ERISA self-funded majority. Transfer to national commercial access is band-limited.
2.6 VA mental-health access standards (measurement precedent)
VHA publishes explicit access standards and facility-level wait metrics for mental health — the measurement object this filing’s KC1 says the civilian system lacks nationally by payer.
Carried standards (Uniform MH Services / successor directives + GAO-16-24 lineage): same-day / next-business-day initial MH screening for new non-urgent need; comprehensive evaluation within a stated window (historically 14- then 30-day preferred-date framing); established-patient scheduling under VHA Directive 1230; outpatient MH staffing guidance (e.g. Directive 1161 productivity/staffing ratios).
GAO caveat (load-bearing): preferred-date wait metrics can understate time from first request; clarifying guidance was repeatedly recommended. The steelman point for this filing is not that VA always meets the standard — it is that a national integrated payer can define and publish realized-access instruments. That is the precedent KC1 points at for civilian measurement architecture.
Transferability: High as a measurement template; low as a delivery transplant (VA eligibility, integrated budget, no Medicaid IMD geography).
3. International peers — bed ratios ≠ access, with transferability
OECD-class psychiatric bed rates differ by >80× across high-income countries (e.g. Japan ~2.58 beds per 1,000 vs Italy ~0.08 vs US ~0.35; OECD median ~0.64 per 1,000 in 2022 extracts cited in McBain-adjacent PLOS Medicine commentary on global bed variation). Burden of disease does not explain the spread; history, financing, and community substitution do.
| Peer | Pattern | Transferability to US scorecard |
|---|---|---|
| England / UK NHS | Long deinstitutionalization + GP/community pathway; psychiatric beds far below historic asylum peak; access fights are wait/community capacity, not “rebuild asylums first” | Financing (NHS single payer) and Mental Health Act civil commitment travel badly. Use as: community-first systems can run at moderate beds if primary/community capacity exists |
| Netherlands | Historically higher bed intensity than Anglosphere peers; active policies to rebalance toward community | Shows high beds are a policy choice, not a clinical necessity bound to prevalence |
| Australia | Low OECD bed rank; high occupancy / short LOS; elevated 30-day schizophrenia/bipolar readmissions vs peers with more beds (MJA / OECD readmission comparisons) | Counter-steelman: low beds + incomplete community offset → acute pressure. Supports some inpatient capacity in the mix — not asylum-scale restoration |
Locked transferability sentence: Peer bed ratios inform whether beds are the main correlate of access once community capacity is in frame; they do not import another country’s civil-commitment or financing law into US Medicaid architecture scoring.
4. Equal effort against H7 — what the bed-rebuild side gets
| Counter-evidence | Role |
|---|---|
| §3 Construct A: state-hospital staffed ~10.8/100k; 52% forensic (TAC 2023); NRI 2025: 90% of states report bed shortage | Acute/forensic inpatient pressure is real on the public hospital construct |
| §3 Construct B stable ~28.4/100k with STACH IPU share falling | Composition stress inside hospitals even without asylum rebuild |
| Pearlmutter MA boarding; national psych ED boarding literature | Hours-to-days boarding is measured; Medicaid/uninsured worse |
| Anchor 12 direction: jail MH prevalence ≫ state-hospital bed census | De-facto “beds” in carceral settings — acute unmet need, not proof community failed |
| Australia low-bed / high-readmission pattern | Community substitution is not automatic |
| AOT/IMD literature still uses hospitalization as an outcome | Inpatient capacity remains clinically load-bearing for a selected SMI band |
Honest synthesis: H7 does not claim acute beds are worthless. It claims the binding contemporary deficit for realized access is outpatient/crisis usable supply (and measurement), and that asylum-scale restoration is the wrong primary architecture. The counter-evidence supports scoring architecture #4 as a targeted acute/forensic capacity play, demoted as the lead “fix mental health” lever if H7 supports.
5. H7 adjudication
H7 — STEELMAN: bed story dated; community capacity was the right bet.
| Criterion | Result |
|---|---|
| (a) Per-capita bed ratios in high-access peer regions are not the main correlate of boarding/access once outpatient capacity is controlled | Met directionally. OECD bed rates vary enormously at similar disease burden; peers with community-heavy mixes (UK) and CCBHC/crisis evaluations show access movement without proportional bed restoration. Clean multi-country regression with outpatient controls not located — disclose as qualitative + case evidence, not a fitted international model. |
| (b) ≥2 evaluations of community-crisis systems show access gains without large bed rebuilds | Met. (1) CCBHC Mathematica/ASPE access package; (2) mobile crisis / BHCC peer-reviewed continuum results (§4). Independent roots; neither is a bed-rebuild program. |
| Refute path: boarding and unmet acute need track bed supply after outpatient controls | Does not fire as a clean refute. Boarding and Construct A shortages are real, but waiver-state bed null (McBain) and community-access gains without bed rebuild block “beds are the binding lever after outpatient controls.” |
| Indeterminate if bed and outpatient series cannot be jointly observed | Partially true nationally (KC1; no payer-resolved joint series) — not used to kill H7 because criterion (b) evaluations jointly observe access under community instruments |
Adjudication: Supported.
Steelman takeaway for scorecard / KC4: Lead with rebuild outpatient + crisis continuum (+ measurement), not asylum-scale bed restoration. Keep a narrow acute/forensic inpatient architecture on the board for O2 stress; demote “beds since 1955 → rebuild asylums” as citogenic and mis-targeted.
Implications for scorecard
- Architecture #4 (bed rebuild / state hospital): score on Construct A/B contemporary need + forensic share — not historical % decline; rank below #1/#2/#5 if weighting favors O1.
- Architecture #2 (CCBHC) and #5 (crisis continuum): strongest H7-consistent measured packages.
- Architecture #3 (IMD repeal/waiver): financing steelman; bed-supply proof still weak.
- Architecture #7 (AOT): selected-band instrument with selection footnotes mandatory.
- Architecture #6 (parity with ERISA teeth): MA/NY show what state exams can do for fully insured lives; KC3 still limits commercial O4.
Confidence: Moderate–strong
Strong on CCBHC and NY AOT administrative evaluations; strong that OECD bed ratios are not disease-burden-determined. Moderate on international outpatient-controlled boarding correlations (not a fitted model). Moderate–weak that any single continuum study travels nationally — disclose geography.