GUBMENTPlain talk · policy frontier
Filings / Elder care / Sources / GBMT-7 — Extraction ledger, research r
GBMT-7 · Research record · No. 7

GBMT-7 — Extraction ledger, research record

elder-care/research/ledger-record.md
This is a working research document from the elder care filing, published as written — including the parts later corrected. It is the underlying record for Whitepaper No. 7, not a summary of it.

Verification Protocol V3(2): every claim in elder-care/docs/, elder-care/research/ and elder-care/baseline/ that does not already appear on the public pages. Claims that do appear there are carried at their P/S numbers in ledger-public.md and are not re-numbered here; where a record file states the same fact differently from the public page, it is numbered here, because the divergence is itself the finding.

264 rows (R001–R264). Verdicts per V7; CONFIRMED-INTERNAL = reproducible from the committed record with no external referent.


docs/research-inquiry.md — protocol, hypotheses, anchor priors (R001–R052)

# Claim Verdict
R001 Status line: Phase 0 executed 2026-08-03, verdict GO, anchors 7 and 11 broken/reframed CONFIRMED-INTERNAL
R002 §2 seed H2.1: unpaid family caregiving accounts for the majority of LTSS hours delivered nationally UNVERIFIABLE — never tested; superseded by the valuation kill condition
R003 §2 seed H2.2: Medicaid, not Medicare, is the primary public LTSS payer CONFIRMED
R004 §2 seed H2.3: no unified national HCBS waiting-list count exists; incompatible waiver definitions, unknown duplicate rates CONFIRMED (KFF states the series is not comparable across states or over time)
R005 §2: the 85+ band is the load-bearing denominator for LTSS need CONFIRMED (ASPE: 40% of 85+ had severe LTSS needs vs 8% of 65–74)
R006 §3 framing: treat workforce as the primary binding constraint, per the childcare/housing rhyme CONFIRMED-INTERNAL — a prior, correctly labelled
R007 §3 seed H3.1 (turnover/vacancy binds more tightly than waiver funding) CONFIRMED-INTERNAL as stated; verdict INDETERMINATE
R008 §3 seed H3.2 (immigration restriction as first-order supply constraint) CONFIRMED-INTERNAL
R009 §3: direct-care turnover is among the highest of any major occupation group UNVERIFIABLE at primary tier
R010 §4: CLASS failed at the design stage, not the implementation stage CONFIRMED
R011 §4 seed H4.1: private LTC insurance is not viable at a price point most households will pay UNVERIFIABLE — never adjudicated in any findings file
R012 §4 seed H4.2: WA Cares is the load-bearing precedent for a payroll-tax design surviving a real legislature CONFIRMED-INTERNAL
R013 §5: nursing home care is a mandatory Medicaid entitlement; HCBS mostly optional capped waivers CONFIRMED at statutory tier (see P094/P095)
R014 §5: Money Follows the Person history of lapses and reauthorizations UNVERIFIABLE — never researched
R015 §5: Olmstead's ADA integration mandate as the legal lever forcing state rebalancing UNVERIFIABLE — named, never reached; no citation anywhere in the record
R016 §5 seed H5.1: largest waiting lists correlate with institutionally-biased spending mix, not low spending CONFIRMED-INTERNAL — verdict "consistent with (n=1)"
R017 §6: CMS Care Compare and the Special Focus Facility program as quality instruments UNVERIFIABLE — named, never researched
R018 §6 seed H6.1: PE-owned facilities show a measurable staffing/outcome delta CONFIRMED
R019 §7: assisted living has no federal Conditions of Participation, no Care Compare equivalent, no uniform staffing or incident-reporting requirement CONFIRMED at secondary tier (V5 negative; consistent with GAO-18-179)
R020 §7: state licensure-category proliferation is a first-order measurement problem CONFIRMED (GAO-26-107884: AL "aren't uniformly defined across federal claims systems")
R021 §7: NIC MAP is a private, subscription, investor-facing source and should be flagged as such CONFIRMED-INTERNAL
R022 §7 seed H7.1 (incidents systematically undercounted) CONFIRMED
R023 §7 seed H7.2 (Medicaid AL footprint growing) CONFIRMED-INTERNAL — verdict: data gap
R024 §8: OAA Title III-E and Title III-C predate every current caregiver proposal by decades and are the baseline CONFIRMED
R025 §8 seed H8.1 (imputed value exceeds total paid LTSS spending) CONFIRMED-INTERNAL — superseded by the kill condition
R026 §8 seed H8.2 (NFCSP orders of magnitude below caregiver value and HCBS spending) CONFIRMED in direction; see S083 on the exponent
R027 §9: Medicaid waiver architecture §1915(c)/(i)/(k), §1115 demonstrations CONFIRMED at statutory tier (42 U.S.C. §1396n)
R028 §9 seed H9.1: is the binding lever federal or state? CONFIRMED-INTERNAL
R029 §11 seed list of nine candidate architectures CONFIRMED-INTERNAL
R030 §11 seed H11.1: demand-side money without workforce capacity inflates wages/prices without adding care hours UNVERIFIABLE — never adjudicated; ws13 finds the opposite for BIP-style packages
R031 Anchor 1 definition: 56% = 2+ ADLs/90+ days or severe cognitive impairment CONFIRMED at agency-primary tier (ASPE 2022, HIPAA-style trigger)
R032 Anchor 1: ASPE 2019 also reports 48% receive paid LTSS lifetime CONFIRMED at agency-primary tier
R033 Anchor 1: Kemper/Komisar/Alecxih 2005 is a Brookings-ICF model of independent lineage CONFIRMED (ASPE 2019 cites it as an independent projection)
R034 Anchor 1: "the current ASPE figure is materially lower (56%)" and 70% is a "citation trap" OVERSTATED — see S018; the two use different models and different definitions
R035 Anchor 2: PHI's figure is BLS-derived, not an independent second source CONFIRMED-INTERNAL — an M1 two-source disclosure, correctly made
R036 Anchor 3: must specify "standalone individual" to avoid understating participation CONFIRMED-INTERNAL
R037 Anchor 4: AARP's 2026 update is replacement-cost, blended wage by task type CONFIRMED at secondary tier
R038 Anchor 4: the 350B600B→$1.01T move is "partly from methodology revision, not just growth" CONFIRMED-INTERNAL — and the qualifier the whitepaper drops (see P141)
R039 Anchor 4: CBO 2013 used HRS data; RAND 2014 used ATUS UNVERIFIABLE at primary tier (cbo.gov blocked; RAND PDF 404)
R040 Anchor 5: left blank, queued CONFIRMED-INTERNAL
R041 Anchor 6: premiums delayed by HB 1732/1733 UNVERIFIABLE at primary tier
R042 Anchor 6: private-insurance opt-out window Oct 2021–Dec 2022 UNVERIFIABLE at primary tier
R043 Anchor 6: SB 5291 (May 2025, eff. Jan 2026) loosened vesting (10yr, breaks allowed, or 3-of-last-6) and added opt-outs; no rate/cap change UNVERIFIABLE at primary tier — and see R044
R044 Anchor 6: rate "unchanged since enactment", cap "auto-inflation-adjusted" OVERSTATED — RCW 50B.04.080(1) moves rate-setting to the Pension Funding Council from 2026-01-01, capped at .58%
R045 Anchor 6: benefits payable July 1, 2026, "confirmed independently (AARP, DSHS, KOMO)" CONFIRMED at statutory tier, one tier deeper than the record reached
R046 Anchor 7: "CMS's own repeal cites the moratorium as running to 2035, one law-firm summary said 2034; not reconciled" CORRECTED — the repeal rule (90 FR 55687) says "September 30, 2034" / "before October 1, 2034" throughout, as does Pub. L. 119-21 §71111
R047 Anchor 7: rural hardship-waiver terms are moot CONFIRMED
R048 Anchor 10: "not RFS, not JPE — the seed didn't specify venue" (venue now on record) CONFIRMED
R049 Anchor 10: "No direct econometric rebuttal found" CONFIRMED-INTERNAL — the honest version of P121
R050 Anchor 10: two independent systematic reviews (BMJ 2023, 55 studies; ScienceDirect 2025, US-specific) reinforce the direction UNVERIFIABLE at primary tier
R051 Anchor 11: "since SNF census is flat-to-declining while AL is the growing setting, a same-year comparison could narrow or reverse the gap" CONFIRMED-INTERNAL
R052 Anchor 12: Medicaid HCBS ≈ $116B (FY2020, KFF) to $162B (CY2020 combined fed/state) UNVERIFIABLE at primary tier

docs/phase0-findings.md (R053–R084)

# Claim Verdict
R053 No kill condition fires as a stop CONFIRMED-INTERNAL
R054 §12's first named kill condition fires exactly as specified CONFIRMED-INTERNAL
R055 The second kill condition (§3) was not testable within Phase 0 scope; left open CONFIRMED-INTERNAL
R056 Three independent knobs explain the valuation spread: population scope, wage-rate assumption, hours source OVERSTATED — elapsed time is a fourth, unnamed driver (R038 concedes it)
R057 "None of the three studies is wrong; they are answering different questions" CONFIRMED-INTERNAL
R058 §8 must report a conditional range by scope and method CONFIRMED-INTERNAL
R059 Rule vacated on statutory-authority grounds in both courts OVERSTATED — 90 FR 55687 records a major-questions holding for N.D. Tex. only; the N.D. Iowa ground is not stated by CMS, and neither opinion has been read
R060 The N.D. Iowa case was a 20-state AG suit UNVERIFIABLE — the repeal rule names Kansas v. Kennedy without a party count
R061 "HHS withdrew its own appeals of both losses in fall 2025" UNVERIFIABLE — 90 FR 55687 records the N.D. Tex. case as "appealed June 2, 2025 to the Fifth Circuit" and says nothing about withdrawal
R062 Congress's moratorium is "a belt-and-suspenders legislative kill shot on top of the court losses" CONFIRMED-INTERNAL
R063 Independently confirmed via AHA, Federal Register, CANHR, Center for Medicare Advocacy, Duane Morris, Skilled Nursing News CONFIRMED-INTERNAL
R064 The rule was "modeled to save 13,000 lives/year" UNVERIFIABLE — appears nowhere in 89 FR 40876's fetched text; no citation in the record
R065 §11's federalism hypothesis gets a strong empirical answer: state-level laws are the more durable lever OVERSTATED — one reversal is not a durability comparison; the record's own patchwork finding cuts against it
R066 AL ~1.016M (NPALS 2022) vs SNF ~1.24M (CASPER July 2025); SNF ~20% larger UNVERIFIABLE at primary tier
R067 The claim is right in order of magnitude, wrong on direction CONFIRMED-INTERNAL
R068 NIC MAP resident count was paywalled and unobtainable CONFIRMED-INTERNAL
R069 CLASS was a pure adverse-selection failure, not a funding or political one OVERSTATED — repeal came in a fiscal-cliff deal; political causation is not excluded by the actuarial record
R070 HHS tried employer auto-enrollment, tiered/phased benefits, and an 80% benefit cutback after 5 years; all left "high uncertainty about long-run solvency" CONFIRMED at secondary tier
R071 Sebelius, Oct 2011: "I do not see a viable path forward for CLASS implementation" CONFIRMED at secondary tier
R072 ATRA 2012 simultaneously created a Commission on Long-Term Care CONFIRMED at statutory tier (Pub. L. 112-240 §643)
R073 CBO independently projected the same structural problem (growing deficits post-2030) UNVERIFIABLE (cbo.gov blocked)
R074 Texas is one of six states (with FL, IA, OK, OR, SC) that don't screen for eligibility before adding to an interest list OVERSTATED — KFF's current brief reports eight such states, of which Texas is the one with interest lists only
R075 Texas 2015: 198,538 raw entries → 101,948 unduplicated (~95% inflation) CONFIRMED at secondary tier
R076 55,009 people on the Texas list were already receiving other Medicaid LTSS CONFIRMED at secondary tier
R077 Texas STAR+PLUS HCBS headcount ~15,850 against ~24,000 slots, secondary-aggregator only CONFIRMED-INTERNAL (self-disclosed gap)
R078 Texas did not report FY2020 LTSS expenditures to CMS's Mathematica rebalancing study CONFIRMED at secondary tier
R079 Florida was rejected as a case (non-responsive to KFF's 2025 survey); Illinois disqualified (no waitlist) CORRECTED — KFF's brief records Illinois waiting lists that grew from 14,444 to 15,905 between 2023 and 2024
R080 ACS 2024 1-year table-based Summary File, B01001, gives population 65+ and 85+ by state CONFIRMED-INTERNAL (pipeline committed)
R081 Two pipeline bugs caught by inline identity checks, recorded in the incompatibility log CONFIRMED-INTERNAL
R082 LTSS spending by payer vs utilization by setting was not pulled; needs NHEA + CMS-64 CONFIRMED-INTERNAL
R083 Re-ranking for full execution: §8 rises, §6/§9 change shape, §5 gets a sharper opening, §11 gains a clean test case, §3 holds CONFIRMED-INTERNAL
R084 WA Cares payouts are weeks old; any performance analysis works with essentially no outcome data CONFIRMED

research/ws02-spending-findings.md (R085–R094)

# Claim Verdict
R085 NHEA "nursing care facilities & CCRCs" mixes short-stay Medicare rehab with long-stay Medicaid custodial care CONFIRMED (KFF describes the same methodology issue)
R086 Medicaid shows as only 30.4% of that NHEA line — an artifact of the mixing UNVERIFIABLE at primary tier
R087 "home health" mixes skilled Medicare home health with Medicaid personal-care HCBS CONFIRMED
R088 Medicaid's $253B splits ~47pts HCBS / 14pts institutional UNVERIFIABLE at primary tier
R089 Up from a 52% split of 29/23 in 2000 UNVERIFIABLE at primary tier
R090 Urban DYNASIM split: Medicare 9.9%, Medicaid 34.3%, OOP 52.3%, private insurance 2.7% UNVERIFIABLE at primary tier
R091 A Health Affairs critique found DYNASIM diverges substantially from NHEA accounting for the same period UNVERIFIABLE at primary tier
R092 DYNASIM is a projection model, not current-spending accounting — treat as a modeling exercise CONFIRMED-INTERNAL
R093 CBO's 2013 report used a third, broader Medicare-inclusive definition UNVERIFIABLE (cbo.gov blocked)
R094 Confidence: medium-high on the headline range because "multiple KFF compilations converge" OVERSTATED — red team #6 already found these share upstream NHEA assumptions; the tier line was not updated

research/ws03-findings.md (R095–R110)

# Claim Verdict
R095 ARPA §9817 was a temporary 10-point FMAP increase earmarked for HCBS CONFIRMED at agency-primary tier (MACPAC)
R096 Only 5% ($1.7B) targeted waiting-list reduction directly CONFIRMED (MACPAC Fig. 1: $1.7B) — the fact that most undercuts P064, and it is absent from the public pages
R097 Only three states (CA, NM, TX) specifically targeted that category CONFIRMED at secondary tier
R098 National median home care worker wage rose $13.07 (2014, infl-adj.) → $16.77 (2024) UNVERIFIABLE at primary tier
R099 Pennsylvania direct-care turnover held at 44–65%, projected 37,000-worker shortfall by 2026 UNVERIFIABLE at primary tier
R100 No source tracks hours of care delivered per enrollee comparably across states CONFIRMED (MACPAC's own measurement-gap finding)
R101 §9817 carried no evaluation requirement; states kept no mandated common baseline CONFIRMED at agency-primary tier (MACPAC: "Section 9817 did not include a requirement to evaluate")
R102 States running concurrent initiatives "encountered difficulty isolating the impact of each initiative" CONFIRMED — verbatim in MACPAC
R103 A GSA/American Institutes for Research national evaluation was still pending CONFIRMED (MACPAC: "A report is anticipated in 2026")
R104 Foreign-born share 28% of the direct-care workforce nationally (PHI, ACS-based, up from 21% in 2011) UNVERIFIABLE at primary tier
R105 36.5% for home health aides (MPI/AIC, 2019 ACS), vs 17% for the overall US workforce, 6.9% undocumented CONFIRMED at secondary tier; STALE as a present-tense claim (2024 base is 19.2%)
R106 ~69–74% of NYC's home health/personal care aide workforce is immigrant; CA health workforce ~35% UNVERIFIABLE at primary tier
R107 DHS terminated the healthcare "sensitive locations" enforcement exemption (Jan 21, 2025) UNVERIFIABLE at primary tier
R108 An executive order (Feb 19, 2025) threatens federal payments to sanctuary-facilitating entities UNVERIFIABLE at primary tier
R109 Justice in Aging's Feb 2026 fact sheet found no documented data on actual staffing disruptions CONFIRMED-INTERNAL (V5 negative, correctly labelled)
R110 EPI's "400,000 jobs threatened" is a modeled projection, not an observed loss CONFIRMED-INTERNAL

research/ws04-insurance-reform-findings.md (R111–R120)

# Claim Verdict
R111 The 2015 study was legislature-mandated, by Milliman via DSHS/ALTSA CONFIRMED at secondary tier
R112 "the original study has not yet been read in this record" CONFIRMED-INTERNAL — and this pass did not reach it either
R113 The public-benefit option modeled alongside it became WA Cares CONFIRMED at secondary tier
R114 The adverse-selection bridge is "the researcher's inference," not a Milliman finding CONFIRMED-INTERNAL
R115 Braun & Kopecky (Cleveland Fed) point to tying private LTC insurance to Medicaid asset-test offsets UNVERIFIABLE — abstract-level only, self-disclosed
R116 No named federal reinsurance-for-LTC bill or Academy-endorsed reinsurance mechanism was found UNVERIFIABLE (V5 negative)
R117 The Bipartisan Policy Center proposal leans on auto-enrollment with opt-out, not reinsurance CONFIRMED at secondary tier
R118 That is equally explained by upfront cost asymmetry as by mechanism efficacy CONFIRMED-INTERNAL (red team #10)
R119 Scorecard effect: CC 2, IQ/WS 1, CR 1, AA 2 for the private-LTC row CONFIRMED-INTERNAL — superseded by the pass-3 board
R120 "The result is real but scope-limited negative evidence, not 'tested and rejected'" CONFIRMED-INTERNAL — contradicts the §4 digest headline at S100

research/ws05-findings.md (R121–R128)

# Claim Verdict
R121 Header corrected on red-team review from SUPPORTED to CONSISTENT WITH (n=1, correlational, confounded) CONFIRMED-INTERNAL
R122 Indiana split its Aged & Disabled Waiver in July 2024 into PathWays (60+) and Health & Wellness (under 60) CONFIRMED at secondary tier
R123 Both new waivers hit approved capacity almost immediately CONFIRMED at secondary tier
R124 Combined waitlist ≈12,800 (July 2024) → over 17,000 (February 2026) CONFIRMED at secondary tier — 19 months, not the whitepaper's eighteen (P098)
R125 Confirmed independently by FSSA/DDARS, KFF's national survey, and state-level journalism CONFIRMED-INTERNAL
R126 Top HCBS-heavy states reach 83.2% UNVERIFIABLE at primary tier
R127 Indiana's 2024 waiver restructuring may itself be a confounder CONFIRMED-INTERNAL
R128 Meta-finding: states useful for testing H5.1 tend to be those with poor spending data — flagged as speculation CONFIRMED-INTERNAL

research/ws06-09-findings.md (R129–R145)

# Claim Verdict
R129 38 states + DC had standards exceeding "the old, pre-2024 federal floor (0.3 HPRD)" CORRECTED — 42 CFR 483.35 contains no numeric HPRD floor, before or after the repeal; the restored text requires "sufficient" staff plus an RN 8 consecutive hours/day, 7 days/week
R130 California 3.5, New York 3.5, DC 4.1 (highest), Florida 3.6, Arkansas 3.36 HPRD UNVERIFIABLE at primary tier (secondary aggregator; deviations #11 discloses this)
R131 New York's 2021 law requires 40% of revenue go to resident-facing staffing UNVERIFIABLE at primary tier
R132 29 states require under 3.5 HPRD, 15 under 2.5 UNVERIFIABLE at primary tier — and the counterweight the public pages omit (see P108/P114)
R133 "state law is a genuinely uneven patchwork, not a uniform substitute for the dead federal floor" CONFIRMED-INTERNAL — omitted from both public pages
R134 No state staffing mandate has been struck down or repealed UNVERIFIABLE (V5 negative)
R135 A fourth industry lawsuit targeting NY's funding/staffing provisions is ongoing; $2,000/day fines recently reaching ~20 facilities UNVERIFIABLE at primary tier
R136 No state legislation specifically reacting to the 2026 rescission is yet visible; a ~6-month window is too short to read as declining to respond CONFIRMED-INTERNAL (red team #12)
R137 Federal replacement bills S.3886 and H.R.8100 have not passed UNVERIFIABLE at primary tier
R138 Health Affairs panel: 2010–2023, mandates raised staffing ~5% (CNAs +5.7%, LPNs +7.5%, no RN change) CONFIRMED at author-institution tier; the publication year is 2026, not 2025 (see S114)
R139 Labor costs rose less than revenue; margins statistically unchanged; no closure effect CONFIRMED at author-institution tier (+273Klaborvs+546K net patient revenue)
R140 Scoped to mandates studied (up to ~4.1 HPRD), not to ratio mandates in general CONFIRMED-INTERNAL (red team #13)
R141 The IV estimate is "notably larger than the naive OLS estimate (+10%)" CORRECTED — the OLS effect is 0.3 pp, "about 2% of the mean"; the direction holds, the magnitude does not
R142 The authors' interpretation is that correlation understates the effect because PE homes selectively admit healthier patients CONFIRMED — verbatim in the paper
R143 Also +19% billed spending, ~50% higher antipsychotic use CORRECTED — IV spending effects are +8% (per stay) and +6% (stay + 90 days); the paper reports no antipsychotic result at all
R144 A JAMA COVID-era comparison prevents a claim that the evidence is uncontested CONFIRMED-INTERNAL
R145 PE-ownership disclosure/oversight is a candidate architecture independent of the staffing-ratio question CONFIRMED-INTERNAL — never added to the board (re-score log records the same gap)

research/ws07-findings.md (R146–R155)

# Claim Verdict
R146 GAO-18-179: only 22 of 48 studied states could report critical-incident counts CONFIRMED (GAO: "Twenty-six state Medicaid agencies could not report…")
R147 Incident types covered: abuse, neglect, medication errors, unexplained deaths CONFIRMED (GAO highlights name the same categories)
R148 Only 34 of 48 made any incident data public CONFIRMED — verbatim in GAO
R149 GAO-26-107884 (2026) "confirms the gap persists" CORRECTED — the 2026 report is Assisted Living Facilities: Information on Federal Spending and Medicaid Coverage (June 2, 2026) and makes no incident-reporting finding
R150 Medicare+Medicaid spent at least $12B on AL services in 2024, "likely an undercount" CONFIRMED — verbatim in GAO ($3.5B Medicaid + $8.5B traditional Medicare)
R151 CMS finalised a rule (May 2024) requiring annual state critical-incident reporting, effective 2027 at the earliest CONFIRMED at secondary tier
R152 Washington Post: 2,000+ elopement incidents since 2018, ~100 deaths UNVERIFIABLE — paywalled, secondhand, self-disclosed
R153 No rigorous state-by-state AL-vs-SNF per-resident incident-rate comparison exists CONFIRMED-INTERNAL (V5 negative, correctly labelled)
R154 Sources disagree on Medicaid AL-waiver state count (41–46) — a measurement-consistency issue, not a conflict CONFIRMED-INTERNAL
R155 NCAL/AHCA: ~18–20% of AL residents Medicaid-supported, services only, never room-and-board UNVERIFIABLE at primary tier

research/ws08-findings.md (R156–R170)

# Claim Verdict
R156 Cash & Counseling ran 1998–2003 in AR/FL/NJ, RWJF/ASPE-funded, Mathematica-evaluated CONFIRMED (ASPE hosts the evaluation reports)
R157 Satisfaction: AR 60.8% vs 42.7% "very satisfied" UNVERIFIABLE at primary tier this pass
R158 Treatment caregivers reported less physical strain and better self-rated health across all three states CONFIRMED at secondary tier
R159 Control-group members frequently didn't receive services they were authorized for CONFIRMED at secondary tier
R160 Costs higher in year 1 in most states (AR elderly +17%, FL nonelderly +14%, both significant) UNVERIFIABLE at primary tier
R161 Narrowing to non-significant in Arkansas by year 2 but persisting in FL/NJ UNVERIFIABLE at primary tier
R162 Self-direction is available in all 50 states + DC via 1915(c)/(i)/(j)/(k) and 1115 CONFIRMED at statutory tier for the authorities; state coverage UNVERIFIABLE
R163 ~1.5M people self-directed Medicaid HCBS in 2023 UNVERIFIABLE at primary tier
R164 Usage among 65+ dual-eligibles ranges 0.1% to 99.6% by state UNVERIFIABLE at primary tier
R165 The self-direction/agency-workforce interaction is explicitly untested (PMC12605748) CONFIRMED-INTERNAL
R166 Credit for Caring reintroduced as S.925 / H.R.2036 (March 2025, bipartisan) UNVERIFIABLE at primary tier
R167 Parameters: nonrefundable, 30% of expenses above a $2,000 floor, capped $5,000; earned-income floor $7,500; phase-out above 75K/150K MAGI UNVERIFIABLE at primary tier
R168 No CBO/JCT score exists; JCT typically scores only at markup CONFIRMED-INTERNAL (V5 negative, correctly labelled as a finding)
R169 MetLife 2011 lifetime cost: $324,044 (women), $283,716 (men) — dated, no successor found CONFIRMED-INTERNAL as dated; figures UNVERIFIABLE at primary tier
R170 BRFSS 2021: 19.6% vs 12.9% poor mental health, plus elevated depression, obesity, smoking UNVERIFIABLE at primary tier (cdc.gov 403 to every method tried)

research/ws10-international-findings.md (R171–R186)

# Claim Verdict
R171 Germany: payroll social insurance, 3.6% of wages (4.2% childless 23+), ~2.5% of GDP UNVERIFIABLE at primary tier
R172 By design covers only about half of actual LTC cost — a partial-insurance model CONFIRMED at peer-reviewed tier (NBER w31870: privately financed share ~60% including informal care)
R173 German contribution rate rose again for 2025; federal loans needed to hold 2026 rates steady UNVERIFIABLE — sourced to a German consumer-finance blog
R174 German direct-care vacancies rose 110% in a decade against 45% demand growth and 13% workforce growth, attributed to "NBER/Geyer" CORRECTED (citation) — NBER w31870 (Long-term Care in Germany, Nov 2023) contains no vacancy, demand-growth, or workforce-growth series; its only 13% is the 2021 foreign-national share of the LTC workforce
R175 Most transferable German piece is the partial-coverage design itself CONFIRMED-INTERNAL
R176 Japan: 50% enrollee premiums (age 40+) + 50% general revenue, ~1.8% of GDP, in-kind benefits only UNVERIFIABLE at primary tier
R177 The favorable "Japan succeeded" narrative clusters around Campbell, Ikegami and Gibson CONFIRMED-INTERNAL, with completeness disclaimed (red team #20)
R178 Geyer (PMC7054649) argues cross-country LTC comparisons are undermined by definitional heterogeneity and excluded informal-care costs CONFIRMED-INTERNAL — recorded as the critique's claim, not adopted
R179 Japan projects a 570,000-worker shortfall by 2040 at flat service levels UNVERIFIABLE at primary tier (MHLW not reached)
R180 Copay increases toward 30% "are already being proposed" CONFIRMED-INTERNAL — and the correct wording the whitepaper overstates (P161)
R181 Transferable Japanese fragment is the age-40 pre-funding logic CONFIRMED-INTERNAL
R182 Netherlands: flat 9.65% income-related contribution; Wlz funds the most severe/institutional cases after 2015 UNVERIFIABLE at primary tier
R183 Highest LTC spend in the OECD (4.4% of GDP, 2021, projected toward 6% by 2070) UNVERIFIABLE at primary tier
R184 Municipalities and insurers benefit from pushing people into the nationally-funded Wlz CONFIRMED at peer-reviewed tier (PMC11984009 cited)
R185 Dutch workforce shortage projected at 266,000 by 2035 (already 44,000 short) UNVERIFIABLE at primary tier
R186 No non-strained counterexample was sought in this pass CONFIRMED-INTERNAL (red team #21)

research/ws11-scorecard.md (R187–R204)

# Claim Verdict
R187 Three scoring passes, not one; pass 2 was instruction-blinded with filesystem access throughout CONFIRMED-INTERNAL
R188 25 of 40 ranked cells moved; the blind scorer differed on 26; one kept as a logged judgment split CONFIRMED-INTERNAL
R189 Anchored scale as authored (passes 1–2), preserved alongside the amendment CONFIRMED-INTERNAL
R190 Anchor 3 widened to cover "no evidence either way"; anchor 1 narrowed to require documented adverse effect CONFIRMED-INTERNAL
R191 The original scale is asymmetric: 4 and 5 require evidence, 1 and 2 accept its absence CONFIRMED-INTERNAL
R192 The same defect the housing re-score named on 2026-08-06 CONFIRMED-INTERNAL
R193 ~two thirds of pass 3's corrected cells are below-neutral unevidenced cells returning to neutral CONFIRMED-INTERNAL (20 of 25 = 80%; "two thirds" understates)
R194 Federal LTC insurance basis: German vacancies 110% against 13% workforce growth CORRECTED (citation) — see R174
R195 Federal AL standards basis: "GAO-18-179 and GAO-26-107884 document a persistent government-verified incident-reporting gap" CORRECTED — see R149
R196 Federal AL standards basis: ~10-year congressional moratorium CORRECTED — see R046
R197 OAA/NFCSP row is now entirely unevidenced; CR held at 3 by both scorers independently CONFIRMED-INTERNAL
R198 Rankings under three weightings (corrected scale) CONFIRMED-INTERNAL — all 24 sums independently recomputed and reproduce exactly
R199 Sensitivity under the original scale: private LTC reform still last; HCBS de-capping leads WS and CC; C&C leads CR; wage floor leads nothing CONFIRMED-INTERNAL — all 24 sums independently recomputed and reproduce exactly
R200 "Do not cite a ranking from this file without also citing which scale it came from" CONFIRMED-INTERNAL
R201 The private-LTC row scored 2/1/1/1/2 with exactly one evidenced cell CONFIRMED-INTERNAL
R202 "Cash & Counseling ties or trails nationalized WA Cares insurance and the wage floor" is withdrawn CONFIRMED-INTERNAL
R203 Federal LTC social insurance now ranks last under all three weightings, on evidence CONFIRMED-INTERNAL — but its two "evidenced" cells rest on R173/R174
R204 A three-way tie at 21 under every weighting between wage floor, OAA/NFCSP and unstructured cash CONFIRMED-INTERNAL

research/ws11-rescore-log.md and ws11-blind-scores-2026-08-10.md (R205–R226)

# Claim Verdict
R205 Structural blinding via scripts/batch-rescore.py: one tool-less request, evidence inlined CONFIRMED-INTERNAL
R206 Supersedes the 2026-08-03 pass-2 re-score as the filing's independence claim CONFIRMED-INTERNAL
R207 Elder care was excluded from the 2026-08-06 cross-filing pass at commit 40b7372 CONFIRMED-INTERNAL
R208 Section A: five corrections that hold regardless of the scale change CONFIRMED-INTERNAL
R209 A2 HCBS-WS 2→4 on ws13's BIP stacked DID (+13.24% HCBS workforce, no institutional offset) CONFIRMED at peer-reviewed tier (PMID 40845205 cited)
R210 A7 AL-WS 3→4 on the Health Affairs panel; pass 2 stopped one step short CONFIRMED-INTERNAL
R211 A7 AL-IQ 2→3 as an evidenced wash under the published anchor 3 CONFIRMED-INTERNAL
R212 A1 CR and AA 4→3; under the published scale they would fall to 2 CONFIRMED-INTERNAL
R213 "These two above-neutral cells are the only unevidenced cells the published board scored up" CONFIRMED-INTERNAL
R214 Section B: exactly twenty below-neutral unevidenced cells returned to neutral CONFIRMED-INTERNAL (counted: 20)
R215 "Not one was corrected because new evidence arrived" CONFIRMED-INTERNAL
R216 Section C: five cells where reasoning was replaced and the score held CONFIRMED-INTERNAL
R217 A3 WS=3's published basis wrongly used the Health Affairs staffing-mandate panel; moved to A7 CONFIRMED-INTERNAL
R218 A4 C&C-CC kept at 1; the blind scorer called it "my most aggressive score" CONFIRMED-INTERNAL
R219 Section D: A4 AA judgment split kept at 4 against a blind 5 CONFIRMED-INTERNAL
R220 Section E: fourteen cells matched outright, enumerated CONFIRMED-INTERNAL (counted: 14; 26 + 14 = 40)
R221 The PE-mortality finding attaches to no row; blind scorer: "if PE oversight were row A11, it would be the only row with a 5 on IQ" CONFIRMED-INTERNAL
R222 PACE blind row corroborated ws12 on 4 of 5 cells; IQ contested downward 4→3 CONFIRMED-INTERNAL
R223 Credit for Caring has a full blind row (3/3/3/2/3) that is not entered CONFIRMED-INTERNAL
R224 Three passages leaking prior cell values were redacted, each with a visible marker CONFIRMED-INTERNAL
R225 Residual un-redacted risks itemised (dangling links; ws13's directional phrasing) CONFIRMED-INTERNAL
R226 "This is a re-score, not a fact-check … Whether the underlying findings are true is Phase 1" CONFIRMED-INTERNAL — this pass is that Phase 1, and it found the underlying findings are not uniformly true

research/ws12-pace-sequencing.md (R227–R246)

# Claim Verdict
R227 PACE is a Medicare program and Medicaid state-plan option for adults meeting the state NH level-of-care standard CONFIRMED at agency tier (CMS)
R228 The organization accepts prospective Medicare and Medicaid capitation and must provide the full covered-service package CONFIRMED at agency tier
R229 CMS requires services at a PACE center, the home, and inpatient facilities; new organizations face federal application and state-readiness review CONFIRMED at agency tier
R230 CMS's PACE manual specifies eleven interdisciplinary-team roles CONFIRMED at agency tier (manual cited with page numbers)
R231 Expanding PACE creates demand for the same direct-care labor the filing has not shown can expand CONFIRMED-INTERNAL — labelled a design fact, not a scalability claim
R232 CMS/Abt evaluation: lower hospital admissions and fewer NH days at six months, more community days, higher initial life satisfaction CONFIRMED at agency tier (report cited with page numbers)
R233 Differences narrowed at later follow-up; the report flags selection bias and small late samples CONFIRMED at agency tier
R234 2024 JAMA Health Forum systematic review: PACE subset found reduced long-term NH stays in three of four studies CONFIRMED at peer-reviewed tier (PMID 39028653)
R235 ASPE/Mathematica matched evaluation (8 states, 2006–2008 enrollees, followed to 2011): Medicaid capitation exceeded predicted Medicaid spending in every reported interval CONFIRMED at agency tier
R236 ASPE's literature review reached the same guarded bottom line; the two are not independent CONFIRMED-INTERNAL
R237 CMS July 2026 directory: 206 active National PACE contracts CONFIRMED-INTERNAL (extraction method published and reproducible)
R238 201 report visible enrollment totaling 79,758; five blank because CMS masks values ≤10 CONFIRMED-INTERNAL (deviations #27 records the correction)
R239 Contract mailing addresses fall in 34 states; address state ≠ service-area state CONFIRMED-INTERNAL
R240 52 listed contracts became effective 2024–2026 (eight in 2026); this does not measure net growth CONFIRMED-INTERNAL
R241 "historically hard to scale" remains NOT MEASURED as a causal claim; retire the shorthand CONFIRMED-INTERNAL
R242 Provisional PACE row: CC 2, IQ 4, WS 3, CR 3, AA 4 — not entered CONFIRMED-INTERNAL
R243 Sequencing: build the workforce before treating PACE as capacity expansion CONFIRMED-INTERNAL
R244 Sequencing: do not sell PACE as a cost-containment first move CONFIRMED-INTERNAL
R245 Four prospective M3 criteria, all returning Indeterminate on the 2026-08-06 follow-up CONFIRMED-INTERNAL
R246 The protocol's prior order (workforce first, HCBS de-capping second, caregiver payment last) "remains directionally sound" OVERSTATED — the pass-3 board withdraws the scorecard's support for it under both scale readings; ws12 was not updated

research/ws13-rebalancing-evidence.md (R247–R253)

# Claim Verdict
R247 2026 Health Services Research stacked DID (ACS 2005–21): BIP adoption raised HCBS workforce 13.24%, no significant institutional-workforce effect CONFIRMED at peer-reviewed tier (PMID 40845205)
R248 2025 BMJ Open generalized DID: ~5 pp fewer long NH stays for Medicaid beneficiaries living alone in BIP completer states CONFIRMED at peer-reviewed tier (DOI cited)
R249 HRS does not directly observe HCBS use; concurrent MLTSS adoption remains a confound CONFIRMED-INTERNAL
R250 2025 state-year study found no significant wage association with HCBS spending CONFIRMED at peer-reviewed tier (PMID 39425476)
R251 2023 DID: each added HCBS dollar associated with $0.74 of added total LTSS spending; the self-financing claim is not supported CONFIRMED at peer-reviewed tier (PMID 37326313)
R252 2026 IV/border study: HCBS use reduced Medicare community-initiated home-health use by 1.02 pp — substitution, not total savings CONFIRMED at peer-reviewed tier (PMID 41919377)
R253 MACPAC's 2025 rate-setting review: states use inconsistent compensation inputs; no public wage series is required CONFIRMED at agency tier

research/deviations-log.md, red-team-log.md, baseline/ (R254–R264)

# Claim Verdict
R254 Deviations log freeze point: research-inquiry.md as committed at 3d1785b (2026-08-03) CONFIRMED-INTERNAL
R255 31 numbered deviation entries CONFIRMED-INTERNAL (counted)
R256 Deviation #2: the moratorium end-year is unreconciled between CMS (2035) and a law-firm summary (2034) CORRECTED — resolved at statutory and agency-primary tier as 2034
R257 Deviation #19: red-team corrections were not line-edited into six findings files CONFIRMED-INTERNAL
R258 Deviation #23: that line-edit pass was completed 2026-08-06 OVERSTATED — ws02's confidence tier (R094) and ws04's §4 headline (S100/R120) still read pre-red-team
R259 Deviation #30: the blind matrix differed on 26 of 40; 25 corrected, 1 judgment split; 14 matched CONFIRMED-INTERNAL
R260 Deviation #31: three redactions, each left a visible marker; PACE scored from an evidence-only extract CONFIRMED-INTERNAL
R261 Red-team log: 22 attacks, 4 "partially lands", 2 "answered, no change", 1 "answered, minor addition", rest absorbed CONFIRMED-INTERNAL
R262 Baseline incompatibility #1: ACS .dat estimate E_n sits at field 2n-1; the off-by-one pulled the MOE jam value -555555555 CONFIRMED-INTERNAL
R263 Baseline incompatibility #2: a SUMLEVEL-only join admitted ~550 extra geographies (603 vs the correct 53) CONFIRMED-INTERNAL
R264 Baseline population pipeline (ACS 2024 1-year SF, B01001) is committed and auditable CONFIRMED-INTERNAL

Record coverage: 264 extracted · 264 verdicts · 0 unaddressed.

← All Elder care research documents Sources digest Read the whitepaper