Date: 2026-08-11. Extends Phase 0 anchors 2 (KC1), 3 (Medicaid acceptance), and the H2 kill-condition test. Does not re-litigate KC1 — the absence of a national by-payer wait/acceptance series is carried as a finding.
PRISMA-lite search log
| Field | Logged |
|---|---|
| Strings | psychiatrist Medicaid acceptance secret shopper; Medicaid phantom network mental health claims; Bishop NAMCS psychiatrists insurance; Brahmbhatt Schpero JAMA 2024; Zhu Oregon Medicaid phantom networks; HRSA mental health HPSA population; GAO mental health network adequacy wait by payer |
| Date range | Prefer post-2015 for second acceptance method; NAMCS 2009–10 retained as Method 1 root |
| Inclusion | Peer-reviewed secret-shopper or claims studies with multi-state/national or multi-city frame; federal HPSA methodology; GAO/OIG on standards vs series. Exclude single-clinic anecdotes and association wait surveys as national series. |
| Searched | 2026-08-11 |
1. KC1 carried (not re-litigated)
Finding (Phase 0, confirmed): No federal or ≥80%-population series publishes outpatient MH appointment wait or new-patient acceptance by payer. Circulating waits are vendor/association (e.g., National Council “48-day” figure via HRSA briefs) or one-off studies. CMS has MA network wait standards and Medicaid managed-care secret-shopper rules — standards ≠ a measured national median series.
Under the protocol, that collapse is a headline shape, not a caveat: the country argues about a shortage it does not measure at the point of use. O1 scores on proxies with that disclosure.
2. Method 1 — NAMCS acceptance (Bishop 2014; Phase 0)
Bishop, Press, Keyhani & Pincus, JAMA Psychiatry 2014 (NAMCS 2009–2010, office-based physicians):
| Psychiatrists | Other physicians | |
|---|---|---|
| Medicaid acceptance | 43.1% (95% CI 34.9–51.7) | 73.0% |
| Private noncapitated | 55.3% | 88.7% |
Frame: national survey of acceptance, not secret-shopper realized appointments. Age of data (~15 years) disclosed. Commercial acceptance materially higher than Medicaid in the same frame.
Wen et al. (JAMA Psychiatry 2019) extend NAMCS through Medicaid expansion (2010–2015) — same instrument family, not a second independent method.
3. Method 2 (owed) — delivered: secret-shopper + claims
3a. Secret-shopper (multi-city) — Brahmbhatt & Schpero, JAMA 2024
Audit of psychiatric prescribing clinicians (psychiatrists, NPs, PAs) listed as accepting new Medicaid patients in the largest Medicaid managed-care directories in NYC, Los Angeles, Chicago, Phoenix (May–Jul 2023):
- Only 17.8% of sampled clinicians were reachable, accepted Medicaid, and offered a new-patient appointment
- 27.2% of offices had any appointment (sampled or alternate at same practice)
- City range for any appointment: NYC 36.3%, Phoenix 30.0%, Chicago 27.5%, LA 15.0%
- Median waits among available appointments: Phoenix 11 days → LA 64 days (IQR up to 126 in LA); waits up to ~6 months in outliers
- Among failures: 15.2% bad/out-of-service numbers; 35.0% no answer on two attempts
Independence from Bishop: different method (secret shopper vs NAMCS acceptance), different years, different outcome (realized appointment vs stated acceptance). Multi-city, not full national. No commercial arm in this study (authors note the limitation; cite prior work for the Medicaid–commercial gap).
3b. Claims / phantom networks — Zhu et al., Health Affairs 2022 (Oregon Medicaid)
Directory listings vs 2018 claims (provider counted “in-network” if ≥5 unique Medicaid beneficiaries with claims):
- 58.2% of directory listings were phantom (no Medicaid claims activity)
- Mental-health prescribers: 67.4% phantom
- Mental-health nonprescribers: 59.0% phantom
- Primary care: 54.0% phantom
Independence: claims reconstruction vs survey vs shopper. Single state (Oregon), so it strengthens the Medicaid realized-access story but does not alone satisfy “multi-state or national” for H2. Useful as Method 2b / corroboration that directory “participation” ≠ usable supply.
Ludomirsky et al., Health Affairs 2022 (cited by Brahmbhatt): Medicaid managed-care care highly concentrated among a small share of listed physicians — same claims-family insight, concentration rather than acceptance rate.
4. HPSA vs realized access (two-method shortage rule)
| Method | What it measures | 2024–2026 snapshot | Limit |
|---|---|---|---|
| HPSA (HRSA) | Population-to-provider (often psychiatrist-only) ratios + poverty/age/travel/SUD scores; geographic, population, facility designations | Aug 2024: ~122M (~⅓ of US) in a Mental Health HPSA (HRSA State of the Behavioral Health Workforce 2024). Dec 2, 2025: ~137M (~40%) (HRSA 2025 brief). Quarterly summary (as of Jun 30, 2026): 7,109 MH HPSA designations; designated population 157.1M; 26.53% of need met; 7,825 practitioners needed to remove designations | Counts licensed/assigned supply geography, not payer acceptance or new-patient wait. Most MH HPSAs still psychiatrist-ratio based (footnote in HRSA quarterly). |
| Realized access | Appointment offer under a stated payer | Brahmbhatt: 17.8% usable among directory-listed Medicaid clinicians in 4 cities; Bishop: 43.1% psychiatrist Medicaid acceptance nationally (2009–10) | No national by-payer series (KC1) |
Two-method verdict: HPSA designations show a large geographic “shortage” footprint. Realized-access methods show that even where clinicians appear in Medicaid directories, usable new-patient access is far below headcount. Shortage-as-HPSA and shortage-as-appointment are not the same object. Scorecard cells that cite only HPSA population shares without a realized-access proxy fail this workstream.
5. H2 adjudication
H2 (flagship / KC1 test) — Licensed ≠ available, and Medicaid is the wedge. A majority of the apparent outpatient shortage is non-participation in Medicaid (and narrow networks), not an absolute absence of licensees.
| Criterion | Result |
|---|---|
| ≥2 independent methods show Medicaid new-patient acceptance (or realized new-patient access) below 50% in multi-state/national frame | Met. Method 1 NAMCS: 43.1% national. Method 2 secret-shopper: 17.8% appointment success among listed Medicaid clinicians across 4 large cities (multi-city / multi-state plans). Claims phantom rates in Oregon push effective directory participation well below 50% for MH specialties. |
| and commercial acceptance materially higher in the same frame | Met in Method 1 (55.3% private vs 43.1% Medicaid for psychiatrists; gap vs other physicians larger still). Not tested in Method 2 (Brahmbhatt has no commercial arm). |
| Refute path (Medicaid acceptance >70%, or licensed supply below panel benchmark even at 100% participation) | Does not fire. |
Adjudication: Supported. Two independent methods place Medicaid usable access below 50%; the commercial wedge is documented in the national NAMCS frame. Disclose: Method 2 measures directory-realized access (often worse than stated acceptance) and lacks a same-study commercial comparator; KC1 still means O1 has no national instrument.
Second acceptance method: FOUND (Brahmbhatt 2024 secret-shopper; Zhu 2022 claims as claims-family corroboration).
Implications for scorecard
- Flagship whitepaper lead remains measurement failure (KC1) + Medicaid usable-supply wedge (H2).
- Architectures that only grow licensed headcount without rate/network/directory integrity score poorly on O1/O3.
- HPSA population shares are allowable as supporting geography, never as the O1 cell alone.
- Secret-shopper and claims methods are the right proxy class for O1 until a national series exists.
Confidence: Strong on the wedge direction; Moderate on national point estimates
Strong that Medicaid usable access ≪ licensed/directory counts (two methods, consistent direction). Moderate that any single percentage (43% vs 18%) generalizes to all therapists nationally — Bishop is psychiatrists/office-based/aged; Brahmbhatt is urban prescribing clinicians in four cities.