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

GBMT-7 — Extraction ledger, public pages

elder-care/research/ledger-public.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. Every factual assertion on site/elder-care/index.html (P001–P218) and site/elder-care/sources/index.html (S001–S207). 425 rows. Verdict vocabulary per V7: CONFIRMED · CORRECTED · OVERSTATED · UNVERIFIABLE · STALE. CONFIRMED-INTERNAL = reproducible from the committed record, no external referent (scorecard arithmetic, self-descriptions of process).

Adjudication and sources are in verification-log.md. Extraction was performed by the adjudicating model directly — nothing was delegated — so the V3 guardrail is reported as a self-consistency re-extraction (see the log's "Extraction guardrail" section) rather than as a diff against a cheaper ledger.


A. Whitepaper — site/elder-care/index.html

# Claim Line Verdict
P001 Title: "Elder Care — Gubment Policy Whitepaper No. 7" 15 CONFIRMED-INTERNAL
P002 meta description: workforce-vs-funding test, dead staffing rule, one architecture with a real trial 16 CONFIRMED-INTERNAL
P003 og:title "workforce or funding — which one actually caps home care?" 21 CONFIRMED-INTERNAL
P004 og:description "The workforce theory came back INDETERMINATE. A federal staffing rule got killed twice." 22 CONFIRMED-INTERNAL
P005 canonical www.gubment.com/elder-care/ 17 CONFIRMED-INTERNAL
P006 Whitepaper No. 7, Series GBMT-7, Filed 2026-08-03 244 CONFIRMED-INTERNAL
P007 H1: does workforce or funding actually cap home care? 245 CONFIRMED-INTERNAL
P008 "A feasibility assessment of LTSS for older Americans" 246 OVERSTATED — scope; majority of the HCBS evidence used is not older-adult (see log C1)
P009 "a federal staffing rule killed twice" 246 CONFIRMED
P010 "the one architecture in the whole candidate set with a real trial behind it" 246 CONFIRMED
P011 Dossier SUBJECT: LONG-TERM CARE 252 CONFIRMED-INTERNAL
P012 Dossier SCOPE: $415B A YEAR 253 CONFIRMED
P013 Dossier SCOPE: 61% MEDICAID 253 CONFIRMED
P014 Dossier FINDING: MEDICARE MOSTLY DOESN'T PAY FOR THIS 254 CONFIRMED
P015 Dossier STATUS: WHITEPAPER NO. 7 — LIVE 255 CONFIRMED-INTERNAL
P016 Rail verdict "Medicare mostly doesn't pay for this." 279/463 CONFIRMED
P017 Cite line: Policy Whitepaper No. 7, Aug 2026 299/483 CONFIRMED-INTERNAL
P018 Inquiry tested whether workforce capacity, not Medicaid funding, is the binding HCBS constraint 310 CONFIRMED-INTERNAL
P019 "the same shape of finding our childcare and housing reports reached" 310 CONFIRMED-INTERNAL
P020 "a 2021 federal funding surge earmarked for direct-care wages" 310 OVERSTATED — ARPA §9817 was a 10-pt HCBS FMAP bump, not wage-earmarked (log B2)
P021 Pattern equally explained by economy-wide post-pandemic wage inflation 310 CONFIRMED
P022 Hypothesis reported INDETERMINATE, not confirmed 310 CONFIRMED-INTERNAL
P023 2024 federal minimum-staffing rule vacated by two courts 310 CONFIRMED (90 FR 55687)
P024 "mooted by a ten-year congressional moratorium" 310 CORRECTED — enactment 2025-07-04 → 2034-09-30 (Pub. L. 119-21 §71111) ≈ 9 yr 3 mo
P025 "dead, not merely contested" 310 CONFIRMED
P026 Nine candidate architectures scored on cost/quality/workforce/caregiver relief/autonomy 310 CONFIRMED-INTERNAL
P027 Only Cash & Counseling anchored in a genuine RCT 310 CONFIRMED
P028 C&C is self-direction letting Medicaid beneficiaries pay family members 310 CONFIRMED
P029 C&C does not win under every weighting once its cost increase is scored 310 CONFIRMED-INTERNAL
P030 That correction was made on independent re-score 310 CONFIRMED-INTERNAL
P031 Second, structurally blinded re-score moved 25 of 40 cells 310 CONFIRMED-INTERNAL
P032 It forced withdrawal of the most confidently stated finding 310 CONFIRMED-INTERNAL
P033 Scoring scale marked instruments down for being under-researched 310 CONFIRMED-INTERNAL
P034 The declared clear loser had almost no evidence either way 310 CONFIRMED-INTERNAL
P035 Corrected, most of the board sits at neutral 310 CONFIRMED-INTERNAL
P036 "the domain's conventional wisdom is less settled than either political party assumes" 310 UNVERIFIABLE — no party-position analysis in the record
P037 "our own scorecard was more confident than its evidence" 310 CONFIRMED-INTERNAL
P038 56% will need severe LTSS 330 CONFIRMED (ASPE Aug 2022 rev.)
P039 $415B total US LTSS spending 331 CONFIRMED (KFF, 2022)
P040 61% Medicaid's share 332 CONFIRMED (KFF)
P041 "$209M OAA caregiver support, total" 333 OVERSTATED + UNVERIFIABLE — Title III-E only, not all OAA caregiver funding; FY2026 figure not reached at primary tier
P042 >half turning 65 develop a severe LTC need before death — 56% on the current model 335 CONFIRMED
P043 Not the 70% figure "still circulating from a superseded estimate" 335 OVERSTATED — 70% is ASPE 2019 (HRS), a different model and definition, not formally superseded
P044 Medicare covers acute and narrow post-acute care, not LTC 335 CONFIRMED
P045 Medicaid reached only after asset spend-down 335 CONFIRMED
P046 Medicaid covers 61% of the $415B 335 CONFIRMED
P047 "home- and community-based spending having overtaken institutional spending nationally in 2013" 335 CORRECTED — the 2013 crossover is a Medicaid LTSS fact; attached here to the all-payer $415B total
P048 No single canonical "LTSS spending" figure exists 335 CONFIRMED
P049 Federal categories mix Medicare rehab with Medicaid custodial care, understating Medicaid's share 335 CONFIRMED
P050 Src: ASPE 2022 (DYNASIM4) · KFF 10 Things · CMS NHEA 336 CONFIRMED
P051 Plain talk: Medicare mostly doesn't pay; Medicaid does, after spend-down 337 CONFIRMED
P052 H2: Medicaid pays for this, not Medicare — and only after the savings are gone 328 CONFIRMED
P053 Stamp: Indeterminate 343 CONFIRMED-INTERNAL
P054 $37.1B ARPA HCBS funding, 2021 345 OVERSTATED — MACPAC: $37.1B planned, state and federal funds
P055 "71% of it went to wages" 346 CORRECTED — $26.3B was "workforce recruitment and retention" (rate increases, bonuses, training excluded)
P056 710K+ on HCBS waitlists, 2024 347 CONFIRMED (KFF) — but see P063/C1 on composition
P057 36.5% home health aides foreign-born 348 CONFIRMED at secondary tier; 2019 ACS vintage, presented as current
P058 Best natural experiment available is the 2021 funding surge 350 CONFIRMED-INTERNAL
P059 "$37.1B, 71% routed to direct-care wages and hiring bonuses" 350 OVERSTATED (as P055)
P060 Wages moved substantially where states spent it that way 350 CONFIRMED at secondary tier
P061 Colorado direct-care wage 12.41→ 18/hour 350 UNVERIFIABLE at primary tier
P062 Vacancy and turnover did not improve to match 350 CONFIRMED (KFF 2023 50-state survey)
P063 National HCBS waiting list grew ~692,679 (2021) → over 710,000 (2024) 350 OVERSTATED — KFF: "roughly 0.7 million in most years since 2016," series not comparable over time, definition changed in 2023
P064 "the opposite of what capacity-targeted funding should produce if funding were the binding constraint" 350 OVERSTATED — inference rests on P063; only 5% of ARPA funds targeted waitlists
P065 Src: MACPAC ARPA §9817 (2026) · KFF 50-state HCBS surveys 351 CONFIRMED
P066 "That pattern fits the workforce theory." 352 CONFIRMED-INTERNAL
P067 Wage-inflation confound "nobody has ruled out" 352 CONFIRMED (V5: MACPAC states no evaluation requirement)
P068 The program carried no evaluation requirement 352 CONFIRMED (MACPAC, primary)
P069 The federal assessment that could settle this is still pending 352 CONFIRMED (GSA/AIR, report anticipated 2026)
P070 "The honest verdict is open, not proven." 352 CONFIRMED-INTERNAL
P071 More than a third of home health aides are foreign-born 353 CONFIRMED at secondary tier (2019 ACS)
P072 "over twice the rate for the workforce overall" 353 CONFIRMED as of the 2019 base (17%); STALE against 2024 (19.2% → 1.9×)
P073 Real 2025–2026 immigration-enforcement changes have hit this sector already 353 CONFIRMED at secondary tier
P074 So far fear-driven absenteeism reported locally, not a measured national decline 353 CONFIRMED
P075 H2: wages rose, vacancies didn't fall, nobody's proven which caused what 342 CONFIRMED
P076 CLASS Act, a federal LTC insurance program, enacted 2010 359 CONFIRMED (ACA Title VIII, 2010-03-23)
P077 Collapsed in 2013 without enrolling a person, opening a claim, or collecting a premium 359 CONFIRMED (repealed by Pub. L. 112-240 §642, 2013-01-02)
P078 HHS's own 19-month review found voluntary signup let only the sickest buy in 359 CONFIRMED at secondary tier
P079 No benefit design could price around it 359 CONFIRMED at secondary tier
P080 WA Cares removes that channel with a mandatory 0.58% payroll tax 359 CONFIRMED (RCW 50B.04.080; WA Cares Fund) — see log D2 on the 2026 rate-setting change
P081 $36,500 lifetime cap vs CLASS's proposed open-ended benefit 359 CONFIRMED (365 benefit units × $100, RCW 50B.04.010/.060)
P082 "the two programs are not a clean contrast" 359 CONFIRMED-INTERNAL
P083 Adverse-selection failure and inadequate-benefit failure could look identical from outside 359 CONFIRMED-INTERNAL
P084 "WA Cares' first-ever payment went out July 1, 2026" 359 OVERSTATED — statute makes benefits payable from that date; no payment event verified
P085 "there is no outcome data yet" 359 CONFIRMED
P086 Src: HHS actuarial review of CLASS · ATRA 2012 · Washington ESD, WA Cares Fund 360 CONFIRMED
P087 Stamp: Confounded natural experiment 361 CONFIRMED-INTERNAL
P088 Washington modeled a state reinsurance backstop instead of a public benefit 362 CONFIRMED at secondary tier
P089 Its 2015 feasibility study found "little potential to generate savings" 362 UNVERIFIABLE — Milliman study never retrieved (record's own disclosure)
P090 "suggestive evidence … from one state, reported secondhand, not a verdict on every reinsurance mechanism" 362 CONFIRMED-INTERNAL (correctly self-limited)
P091 Standalone individual LTC market shrank from 125 carriers (2000) to ~7–10 today 362 CONFIRMED at secondary tier; no census post-2016
P092 Plain talk: CLASS collapsed because only sick people signed up; WA's benefit is smaller and capped 363 CONFIRMED
P093 H2: the last federal program died before it opened; the new one hasn't paid a claim yet 358 OVERSTATED — conflicts with P084's own "first-ever payment went out"
P094 Nursing home care is a mandatory Medicaid entitlement 369 CONFIRMED at statutory tier (42 U.S.C. §1396a(a)(10)(A) + §1396d(a)(4)(A))
P095 HCBS runs through optional, capped waivers with waiting lists in most states 369 CONFIRMED at statutory tier (§1396n(c)(1), (c)(9)); 40 of 51 states, KFF 2024
P096 "backwards from the 'home care is cheaper' intuition" 369 CONFIRMED-INTERNAL
P097 Waitlist ~710,000 by 2024, about 40% above the seeded prior 369 CONFIRMED-INTERNAL (arithmetic: 710/500 = 1.42)
P098 Indiana waitlist grew 12,800 → over 17,000 "in eighteen months" 369 CORRECTED — record dates are July 2024 → Feb 2026 = 19 months
P099 Indiana sends two-thirds to three-quarters of Medicaid LTC dollars to institutional care 369 CONFIRMED at secondary tier (34% vs ~23% HCBS, both disputed in record)
P100 "But it is one state" 369 CONFIRMED-INTERNAL
P101 Texas is the one state where the relevant spending data is unknowable, excluded from two national trackers 369 CONFIRMED at secondary tier (V5: exclusion documented, not independently re-verified)
P102 Src: KFF Waiting Lists 2016–2025 · Indiana FSSA/DDARS · AARP LTSS Scorecard 370 CORRECTED — the KFF brief relied on is "2016 to 2024"
P103 Stamp: Consistent, not confirmed 371 CONFIRMED-INTERNAL
P104 Plain talk: money isn't obviously the problem; structure might be 372 CONFIRMED-INTERNAL
P105 H2: nursing homes are a legal right, home care is a waiting list 368 CONFIRMED
P106 Stamp: Prior broken 378 CONFIRMED-INTERNAL
P107 3.48 HPRD in the dead federal rule 380 CONFIRMED (89 FR 40876)
P108 "38+DC states already exceeded old floor" 381 OVERSTATED — the pre-2024 federal standard had no numeric HPRD floor; record's own finding is that 29 states require under 3.5 HPRD and 15 under 2.5
P109 +11% mortality at PE-owned homes 382 CONFIRMED (NBER 28474 rev. 2023 / RFS 37(4) 2024)
P110 The 2024 rule is dead, not merely under litigation risk 384 CONFIRMED
P111 Two federal courts vacated it 384 CONFIRMED (90 FR 55687 names AHCA v. Kennedy, N.D. Tex.; Kansas v. Kennedy, N.D. Iowa)
P112 Congress imposed a ten-year moratorium 384 CORRECTED — see P024
P113 CMS formally repealed it effective February 2026 384 CONFIRMED (eff. 2026-02-02)
P114 38 states + DC already exceeded the old federal floor before the rule existed 384 OVERSTATED — see P108
P115 "no state staffing law has been struck down" 384 UNVERIFIABLE (V5 negative; no docket search in the record or this pass)
P116 New York: ~400 violating facilities identified with almost no penalties for years 384 CONFIRMED at secondary tier (single case, trade press)
P117 Src: Federal Register (Dec 2025 repeal) · Center for Medicare Advocacy · "Health Affairs 2025, 22-state staffing panel" 385 CORRECTED — the panel published in the March 3, 2026 issue
P118 PE-owned nursing homes show an 11% causal increase in short-stay mortality 386 CONFIRMED — LATE for compliers; MTE analysis recovers a similar ATE
P119 "the most rigorous study run on the question" 386 OVERSTATED — no systematic comparison of designs in the record
P120 A design built specifically to correct for which patients end up at which homes 386 CONFIRMED
P121 "No published rebuttal exists." 386 OVERSTATED — record's own text is "no direct econometric rebuttal was located"; a JAMA COVID-era comparison is on record
P122 Src: Gupta, Howell, Yannelis & Gupta, RFS (2024) 387 CONFIRMED
P123 Plain talk: courts killed it, Congress made sure it can't come back for a decade 388 OVERSTATED — see P024
P124 Plain talk: PE homes have measurably higher death rates, nobody's found a real hole 388 OVERSTATED — see P121
P125 H2: the federal staffing rule is dead, twice over 377 CONFIRMED
P126 The inquiry assumed assisted living had overtaken nursing homes in resident count 394 CONFIRMED-INTERNAL
P127 ~1.24M nursing-home residents vs 1.016M in assisted living 394 UNVERIFIABLE at primary tier (NPALS and CASPER both blocked this pass)
P128 "the comparison is three years stale on one side" 394 CONFIRMED-INTERNAL
P129 NH census flat-to-declining while AL keeps growing, so the gap could be smaller or closed 394 CONFIRMED-INTERNAL (correctly hedged)
P130 "two federal watchdog reports eight years apart both found most states can't count AL safety incidents" 394 CORRECTED — GAO-26-107884 is Information on Federal Spending and Medicaid Coverage; it makes no incident-reporting finding
P131 The federal reporting fix won't take effect until 2027 at the earliest 394 CONFIRMED at secondary tier
P132 Src: NPALS (2022) · KFF/CMS CASPER (2025) · GAO-18-179 (2018) · GAO-26-107884 (2026) 395 CONFIRMED (both GAO numbers exist; 2018-01-05 and 2026-06-02)
P133 Plain talk: AL is still smaller than nursing homes by headcount 396 CONFIRMED-INTERNAL
P134 H2: still the smaller setting, and still uncounted 393 CONFIRMED
P135 Stamp: Kill condition fired 402 CONFIRMED-INTERNAL
P136 63M family caregivers, US 404 CONFIRMED at secondary tier (AARP/NAC 2025, all ages)
P137 234B1.01T range of "the" value 405 CONFIRMED-INTERNAL — but see P140
P138 19.6% caregivers with poor mental health 406 UNVERIFIABLE at primary tier (cdc.gov blocked)
P139 Unpaid caregiving's economic value cannot be reconciled to a single defensible figure 408 CONFIRMED-INTERNAL
P140 Estimates "for a comparable period" range 234B1.01T 408 CORRECTED — $234B is CBO 2013 (2011 data); $1.01T is AARP 2026 (2024 data). Not comparable periods
P141 Spread explained by population scope, wage-rate assumption, hours source 408 OVERSTATED — the record's own anchor 4 shows AARP's own series moved 350B600B→$1.01T over time; elapsed time is a fourth driver, unnamed
P142 "There is no 'the' number." 408 CONFIRMED-INTERNAL
P143 Src: AARP/NAC Valuing the Invaluable (2026) · CBO (2013) · RAND (2014) 409 CONFIRMED-INTERNAL
P144 63 million Americans provide unpaid care 410 CONFIRMED at secondary tier
P145 Over a quarter reduced work hours because of it 410 CONFIRMED at secondary tier (AARP/NAC 2025)
P146 CDC's national health survey found the caregiver mental-health gap widened over the last decade 410 OVERSTATED — record's own red team #19 downgraded this to pandemic-confounded and suggestive
P147 The one architecture anchored by an actual RCT — Cash and Counseling 410 CONFIRMED
P148 C&C lets Medicaid recipients hire their own caregivers, including family 410 CONFIRMED
P149 Measurably improved satisfaction and health 410 CONFIRMED at secondary tier (ASPE/Mathematica)
P150 "It did not save money" 410 CONFIRMED
P151 It delivered more of the authorized care rather than the same amount more cheaply 410 CONFIRMED
P152 Src: CDC BRFSS Caregiver Module (2021) · ASPE Cash and Counseling Demonstration 411 CONFIRMED-INTERNAL
P153 Plain talk: no trustworthy single dollar figure — that's a finding, not a gap 412 CONFIRMED-INTERNAL
P154 Plain talk: the one strong trial says paying them directly works, it just isn't cheaper 412 CONFIRMED
P155 H2: no honest dollar figure, one genuine trial 401 CONFIRMED
P156 Germany, Japan, Netherlands are the usual LTC-financing comparators 418 CONFIRMED-INTERNAL
P157 All three show real, current strain rather than a settled model 418 CONFIRMED at secondary tier
P158 Germany needed federal loans to hold its 2026 premium steady 418 UNVERIFIABLE — cited to NBER w31870, which contains no such content; record's actual source is a German consumer-finance blog
P159 German workforce vacancy rate rose 110% in a decade 418 CORRECTED (citation) — NBER w31870 (Long-term Care in Germany, Nov 2023) contains no vacancy statistics at all
P160 Japan projects a 570,000-worker shortfall by 2040 418 UNVERIFIABLE at primary tier (MHLW not reached)
P161 Japan "is already raising copays" 418 CORRECTED — record says copay increases are being proposed
P162 The Netherlands' 2015 reform works against its own cost-containment goal 418 CONFIRMED at secondary tier
P163 Score all three as donors of design fragments, not whole-system precedents 418 CONFIRMED-INTERNAL
P164 Src: NBER w31870 · MHLW 9th LTCI plan · PMC11984009 (Wlz) 419 CORRECTED — w31870 does not support either Germany claim
P165 Plain talk: all three are raising rates, tightening benefits, or short-staffed 420 OVERSTATED — "raising" for Japan, see P161
P166 H2: three systems, all mid-crisis, none a finished blueprint 417 CONFIRMED
P167 Table row: FEDERAL LTC INSURANCE 2/3/2/3/3 429 CONFIRMED-INTERNAL (matches reconciled matrix)
P168 Table row: HCBS DE-CAPPING 2/3/4/3/4 430 CONFIRMED-INTERNAL
P169 Table row: WAGE FLOOR 3/3/3/3/3 431 CONFIRMED-INTERNAL
P170 Table row: CASH & COUNSELING 1/3/3/5/4 432 CONFIRMED-INTERNAL
P171 Table row: CREDIT FOR CARING —/—/—/2–3/— 433 CONFIRMED-INTERNAL
P172 Table row: PACE EXPANSION all unscored 434 CONFIRMED-INTERNAL
P173 Table row: FEDERAL AL STANDARDS 3/3/4/3/3 435 CONFIRMED-INTERNAL
P174 Table row: OAA/NFCSP EXPANSION 3/3/3/3/3 436 CONFIRMED-INTERNAL
P175 Table row: PRIVATE LTC INS. REFORM 2/3/3/3/3 437 CONFIRMED-INTERNAL
P176 Table row: CASH (COMPARATOR) 3/3/3/3/3 438 CONFIRMED-INTERNAL
P177 Nine candidate architectures, scored 1–5 against five objectives 441 OVERSTATED — two of the nine are unscored; disclosed only in the honesty box
P178 Only C&C is anchored by an actual randomized trial 441 CONFIRMED
P179 It wins decisively on caregiver relief regardless of weighting 441 CONFIRMED-INTERNAL (26 vs 22 under CR×3)
P180 Its own trial data show a real cost increase, pulling its cost rank down 441 CONFIRMED
P181 It does not win "no matter what"; the first-pass scorecard overstated that it did 441 CONFIRMED-INTERNAL
P182 A second, harder re-score moved 25 of these 40 cells 442 CONFIRMED-INTERNAL
P183 The most important thing it found was how little the board knows 442 CONFIRMED-INTERNAL
P184 First re-score was told not to look; this one was structurally prevented 442 CONFIRMED-INTERNAL
P185 Scale defect: "no evidence found" was a low score while a high score required evidence 442 CONFIRMED-INTERNAL
P186 Twenty cells at 1 or 2 rested on no evidence at all, in either direction 442 CONFIRMED-INTERNAL (section B of the re-score log has exactly 20)
P187 Corrected, 30 of the 40 cells sit at dead neutral 442 CONFIRMED-INTERNAL (independently recounted: 30)
P188 The differences between most of these instruments disappear 442 CONFIRMED-INTERNAL
P189 Withdrawing "private LTC insurance reform ranks last under all three weightings, with no exceptions" 442 CONFIRMED-INTERNAL
P190 Four of that row's five cells were empty 442 CONFIRMED-INTERNAL
P191 It was the least stable finding on the board, not the most 442 CONFIRMED-INTERNAL
P192 The wage floor no longer leads on workforce; it ties the do-nothing cash comparator 442 CONFIRMED-INTERNAL (21 = 21; two rows at 24 lead)
P193 The sequencing argument gets no support from the scorecard 442 CONFIRMED-INTERNAL
P194 Src line: 9 × 5 × 3, two independent re-scores, reconciliation log committed 443 CONFIRMED-INTERNAL
P195 Stamp: Twice re-scored; one headline withdrawn 444 CONFIRMED-INTERNAL
P196 H2: the best-evidenced fix isn't the one anyone's proposing 425 CONFIRMED-INTERNAL
P197 Headline hypothesis unresolved; workforce-vs-funding came back INDETERMINATE 450 CONFIRMED-INTERNAL
P198 The best available test has an unruled-out macro-wage-inflation confound 450 CONFIRMED
P199 "Our scoring scale was broken, and it broke in one direction." 450 CONFIRMED-INTERNAL
P200 A cell could only score high on evidence, but scored low on its absence 450 CONFIRMED-INTERNAL
P201 Corrected, 30 of 40 cells sit at neutral 450 CONFIRMED-INTERNAL
P202 Most apparent differences were differences in how hard we looked 450 CONFIRMED-INTERNAL
P203 PACE is still unscored; it now has two independent readings in the record 450 CONFIRMED-INTERNAL
P204 Entering it needs criteria written too late; not backdating them 450 CONFIRMED-INTERNAL
P205 "Our best evidence has no row." 450 CONFIRMED-INTERNAL
P206 "an 11% mortality increase … across seven million patients" 450 CORRECTED — the analysis sample is 4.2 million unique short-stay patients
P207 It attaches to an architecture never put on the board 450 CONFIRMED-INTERNAL
P208 Texas's numbers are unknowable, not just unfavorable 450 CONFIRMED at secondary tier
P209 WA Cares has zero months of real performance data; first benefit paid weeks before this report 450 CONFIRMED (statutory start date); "paid" per P084
P210 "Twice re-scored, not final." 450 CONFIRMED-INTERNAL
P211 Two blind re-scores of a sibling filing moved entirely different cells 450 CONFIRMED-INTERNAL (GBMT-10)
P212 Receipts: "12-workstream protocol" 454 CORRECTED — the sources page and the record both say eleven (nine executed)
P213 Receipts: anchor table preserving corrected errors (superseded 70% vs current 56%) 454 CONFIRMED-INTERNAL
P214 Receipts: deviations log (31 entries) 454 CONFIRMED-INTERNAL (31 rows counted)
P215 Receipts: red-team pass, 22 resolved items, run as a parallel workflow 454 CONFIRMED-INTERNAL
P216 Receipts: two re-scores, second structurally blinded, reconciliation log publishes both readings and one withdrawn headline 454 CONFIRMED-INTERNAL
P217 Receipts: "all public" 454 OVERSTATED — the published deviations digest omits entries 22–29
P218 Next filing: Drugs (GBMT-8) 491 CONFIRMED-INTERNAL

B. Sources & data page — site/elder-care/sources/index.html

Located by record and anchor row rather than line number (the page is generated).

# Claim Where Verdict
S001 Scope: LTSS for older Americans and adults with disabilities — HCBS, AL/residential, SNF, unpaid family care, and the financing under all four R0 CONFIRMED-INTERNAL — and the disclosure the whitepaper drops (see log C1)
S002 "Imports the gubment method (M1–M8) in full" R0 CONFIRMED-INTERNAL
S003 Phase 0 verdict: GO R0 CONFIRMED-INTERNAL
S004 Phase 0 (2026-08-03) verified the anchor table's starred rows R0 CONFIRMED-INTERNAL
S005 Committed a first ACS-derived population baseline (65+ and 85+ by state) R0 CONFIRMED-INTERNAL
S006 Kill condition fired: caregiving valuation not reconcilable to a defensible single range R0 CONFIRMED-INTERNAL
S007 The 2024 CMS staffing rule turned out to be dead, not merely contested R0 CONFIRMED
S008 Assisted living does not out-census skilled nursing after all R0 UNVERIFIABLE at primary tier (see P127)
S009 Eleven workstreams (§2–§12) run to varying depth R0 CONFIRMED-INTERNAL — contradicts P212's "12-workstream"
S010 Red-team pass: Workflow, 11 agents, 22 items resolved, 3 scorecard cells reconciled R0 CONFIRMED-INTERNAL
S011 That re-score was instructed not to read the scorecard, not structurally prevented R0 CONFIRMED-INTERNAL
S012 A genuinely blinded pass ran 2026-08-10 and moved 25 of 40 ranked cells, withdrawing one headline R0 CONFIRMED-INTERNAL
S013 Anchor 1 prior: a majority turning 65 will need some paid or unpaid LTSS before death R1 CONFIRMED-INTERNAL
S014 ASPE 2022 (DYNASIM4): 56% will develop severe need R1 CONFIRMED at agency-primary tier
S015 ASPE 2019 (HRS-based): 70% R1 CONFIRMED at agency-primary tier
S016 Kemper/Komisar/Alecxih 2005: 69% R1 CONFIRMED (cited inside the ASPE 2019 brief)
S017 Majority claim holds either way R1 CONFIRMED
S018 The commonly-cited 70% is ASPE's "superseded" 2019 vintage; cite 56%, flag 70% as a citation trap R1 OVERSTATED — different model and different definition, not a superseding revision
S019 Anchor 2 prior: 1M+ additional direct-care jobs by early 2030s R1 CONFIRMED-INTERNAL
S020 PHI 2025: 772K+ new jobs from growth, 2024–2034 R1 UNVERIFIABLE at primary tier (BLS projections not re-pulled)
S021 9.7M total openings including replacement/turnover R1 UNVERIFIABLE at primary tier
S022 Corrected on red team: the "~10x undersold" comparison mixed growth-only against growth+replacement R1 CONFIRMED-INTERNAL
S023 Like-for-like is ~25% undersold, not 10x R1 CONFIRMED-INTERNAL (772/1000 = 0.77)
S024 Anchor 3 prior: insurers fell from 100+ to a small handful R1 CONFIRMED-INTERNAL
S025 Peak 125 (2000) → 104 (2002), AHIP survey, never repeated R1 CONFIRMED at secondary tier
S026 ~12–15 by 2014 (NAIC) R1 UNVERIFIABLE at primary tier
S027 Current ~7–10 standalone individual carriers (2024–25 trade lists) R1 CONFIRMED at trade-press tier, self-disclosed
S028 LIMRA corroborates >75% exit by 2012 R1 UNVERIFIABLE at primary tier
S029 No rigorous market census exists post-2016 R1 UNVERIFIABLE (V5 negative)
S030 Market partly shifted to hybrid life/LTC combo products ($4.2B new premium, 2024) R1 UNVERIFIABLE at primary tier
S031 Anchor 4 prior: caregivers in the tens of millions, value in the hundreds of billions R1 CONFIRMED-INTERNAL
S032 AARP/NAC 2025: 63M caregivers (all ages) R1 CONFIRMED at secondary tier
S033 AARP Valuing the Invaluable 2026 (2024 data): 59M adult caregivers, 49.5B hours, $1.01T R1 CONFIRMED at secondary tier
S034 Up from $600B (2021) and $350B (2006) R1 CONFIRMED at secondary tier
S035 CBO 2013 (elderly-only, replacement-cost): $234B R1 UNVERIFIABLE — cbo.gov DataDome-blocked (atlas UNOBTAINED.md §1–2 records the same block)
S036 RAND 2014 (elderly-only, opportunity-cost): $522B R1 UNVERIFIABLE at primary tier
S037 Not reconcilable to a single defensible figure — the headline finding R1 CONFIRMED-INTERNAL
S038 Red team clarified: three named methodological drivers, not true irreconcilability R1 CONFIRMED-INTERNAL
S039 The instruction to report a conditional range, never a point estimate, stands R1 CONFIRMED-INTERNAL
S040 Anchor 5 (asset limit ~$2,000) not attempted, queued R1 CONFIRMED-INTERNAL
S041 WA Cares rate 0.58% of gross wages, no wage cap, unchanged since enactment R1 OVERSTATED — RCW 50B.04.080(1): from 2026-01-01 the rate is set biennially by the Pension Funding Council, capped at .58%
S042 Cap $36,500 (2026), auto-inflation-adjusted R1 CONFIRMED (WA Cares Fund: "$36,500 (grows over time with inflation)")
S043 Premiums delayed Jan 2022 → July 2023 R1 CONFIRMED (RCW 50B.04.080(1): "beginning July 1, 2023")
S044 Opt-out window through Dec 2022 R1 CONFIRMED at secondary tier
S045 SB 5291 (2025) loosened vesting, no rate/cap change R1 UNVERIFIABLE at primary tier
S046 Benefits first became payable July 1, 2026 R1 CONFIRMED at statutory tier (RCW 50B.04.060(1))
S047 "Delayed and revised once" undersold the amendment history R1 CONFIRMED-INTERNAL
S048 Research lands weeks after the first-ever payout, essentially zero months of payout data R1 CONFIRMED
S049 2024 rule: 3.48 total HPRD, 0.55 RN, 2.45 NA, 24/7 on-site RN, phased 2026–2029; "Original (Apr 2024)" R1 CONFIRMED at agency-primary tier on every figure; the rule was published 2024-05-10 (89 FR 40876)
S050 Rescinded effective Feb 2, 2026 (CMS interim final rule, Dec 3, 2025) R1 CONFIRMED at agency-primary tier (90 FR 55687)
S051 Following vacatur in two federal courts R1 CONFIRMED (per the repeal rule's own text)
S052 Plus a statutory 10-year moratorium via H.R. 1 (2025) R1 CORRECTED — Pub. L. 119-21 §71111 runs to 2034-09-30
S053 Facility-assessment and Medicaid staffing-spend-transparency provisions survive R1 CONFIRMED at agency-primary tier (repeal touches only §§483.5, 483.35)
S054 MAJOR reframe: the rule is dead by both courts and Congress R1 CONFIRMED
S055 Anchor 8 prior: waiting lists ~500,000+ nationally with substantial double-counting R1 CONFIRMED-INTERNAL
S056 KFF tracker: ~692,679 nationally in 2021 R1 UNVERIFIABLE — the current KFF brief reports the series as "roughly 0.7 million in most years since 2016"; the 692,679 point is not in it
S057 Growing to ~710,000+ by 2024 (+2.6% 2023→2024) R1 CONFIRMED (KFF, both figures verbatim)
S058 Average wait fell 45→36 months (2021–23), rose back to ~40 (2024) R1 CONFIRMED (KFF, verbatim)
S059 CA, NM, TX together hold over half of all waitlisted individuals R1 CONFIRMED at secondary tier
S060 Texas's own 2015 data shows ~95% duplicate-listing inflation R1 CONFIRMED at secondary tier
S061 Seed undersold the scale — actual ~40% higher, still rising through 2024 despite ARPA §9817 R1 OVERSTATED — KFF states the series is not comparable over time and changed definition in 2023
S062 Anchor 9: not attempted in Phase 0; confirmed during full execution R1 CONFIRMED-INTERNAL
S063 Of KFF's $415B (2022) total, HCBS $284B vs institutional $131B R1 CONFIRMED (KFF, verbatim)
S064 HCBS spending crossed institutional spending nationally in 2013 R1 CORRECTED — a Medicaid-LTSS fact, not an all-payer one (see P047)
S065 Gupta/Howell/Yannelis/Gupta, RFS 37(4) 2024 (NBER WP 28474) R1 CONFIRMED
S066 ">7M Medicare patients", 2005–2017, within-facility IV design R1 CORRECTED — "Our analysis sample contains 4.2 million unique short-stay patients"
S067 "OLS: +10% mortality" R1 CORRECTED — the paper reports an OLS effect of 0.3 pp, "about 2% of the mean"
S068 Causal estimate: +11% R1 CONFIRMED (LATE; MTE recovers a similar ATE)
S069 ~20,150–21,000 excess deaths over the sample R1 CORRECTED — the Aug 2023 revision states "about 22,500 additional deaths" and ~172,400 lost life-years
S070 Also +19% billed spending R1 CORRECTED — IV estimates are +8% per stay and +6% per stay + 90 days
S071 ~50% higher antipsychotic use R1 CORRECTED — the paper contains no antipsychotic result; its three well-being measures are mobility, ulcers, pain
S072 "Confirmed and precise, largely uncontested" R1 OVERSTATED — the authors' own MTE finds "small beneficial effects for some patients"
S073 One JAMA COVID-era comparison found PE facilities in line with industry average R1 CONFIRMED-INTERNAL
S074 Confidence downgraded to medium-high; core estimate unrebutted R1 CONFIRMED-INTERNAL
S075 AL/residential care ~1.016M residents (NPALS 2022) R1 UNVERIFIABLE at primary tier (cdc.gov blocked)
S076 SNF ~1.24M residents (KFF/CMS CASPER, July 2025) R1 UNVERIFIABLE at primary tier
S077 Prior likely broken but vintage-mismatched and directionally unstable R1 CONFIRMED-INTERNAL
S078 No independent current AL count obtainable (NIC MAP paywalled) R1 CONFIRMED-INTERNAL
S079 The research/regulatory-attention-asymmetry half of the claim remains plausible R1 CONFIRMED-INTERNAL
S080 NFCSP FY2026 appropriation ≈ $209M (ACL) R1 UNVERIFIABLE — ACL's own NFCSP page publishes only FY2013–FY2016 (145.6M150.6M)
S081 Medicaid HCBS spending ≈ $116B (FY2020) to $162B (CY2020) R1 UNVERIFIABLE at primary tier
S082 Total Medicaid LTSS ≈ $257B (2023) R1 UNVERIFIABLE at primary tier
S083 NFCSP is ~three orders of magnitude below Medicaid HCBS (209Mvs 140B), not two R1 OVERSTATED — the ratio is ~670×, i.e. 2.8 orders, and both inputs are unverified
S084 NFCSP funds well under 0.1% of even the lowest independent valuation R1 CONFIRMED-INTERNAL (209M/234B = 0.089%)
S085 §2 headline: no single canonical LTSS spending number exists R2 CONFIRMED
S086 415B(2022) : Medicaid61253B) R2 CONFIRMED
S087 Out-of-pocket 17% (~$71B) R2 CONFIRMED (KFF, verbatim)
S088 Medicare + private LTC insurance ~21% (~$87B), not separable by any current source R2 CONFIRMED
S089 HCBS (284B)hasexceededinstitutional(131B) since 2013 R2 CORRECTED — see S064
S090 Urban DYNASIM and CBO diverge substantially from NHEA-based accounting R2 CONFIRMED at secondary tier
S091 The field has no single measurement convention R2 CONFIRMED
S092 §3 headline: money to wages, wages rose, neither vacancies nor waiting lists improved; INDETERMINATE R2 CONFIRMED-INTERNAL
S093 ARPA §9817 put 71% ($26.3B of $37.1B) toward workforce recruitment/retention R2 CONFIRMED at agency-primary tier (MACPAC Fig. 1) — the record is right where P055 is wrong
S094 Colorado's direct-care wage 12.41→ 18 R2 UNVERIFIABLE at primary tier
S095 Every state responding to KFF's 2023 survey still reported shortages R2 CONFIRMED at secondary tier
S096 National HCBS waiting lists grew ~692,679 (2021) → ~710,000+ (2024) R2 OVERSTATED — see S056/S061
S097 Red team downgraded the verdict to plain INDETERMINATE R2 CONFIRMED-INTERNAL
S098 Foreign-born share 28% nationally, 36.5% for home health aides — more than double the general workforce R2 CONFIRMED at secondary tier against the 2019 base; STALE against 2024
S099 The actual labor-supply effect is anecdotal-only so far R2 CONFIRMED-INTERNAL
S100 §4 headline: "Reinsurance was actually tried by a state, and it failed" R2 OVERSTATED — red team #8 already softened this to "suggestive negative evidence from one unread secondary-sourced state study"; the headline was not updated
S101 Washington's 2015 Milliman study found "little potential to generate savings" R2 UNVERIFIABLE — PDF never retrieved (record's own disclosure)
S102 The same study led Washington to build WA Cares instead R2 CONFIRMED at secondary tier
S103 Reinsurance shifts tail risk without touching adverse selection R2 CONFIRMED-INTERNAL (labelled the researcher's inference)
S104 Hybrid life/LTC products, dominant since ~2014, pitched at 500K5M net worth R2 CONFIRMED at secondary tier
S105 §5 headline: Indiana confirms the hypothesis; Texas can't be tested R2 OVERSTATED — "confirms" outruns the file's own "CONSISTENT WITH (n=1, correlational, confounded)"
S106 Indiana waitlist ~12,800 (2024) → over 17,000 (Feb 2026) R2 CONFIRMED at secondary tier
S107 34% HCBS per the state's report, ~23% per AARP's Scorecard R2 CONFIRMED-INTERNAL (discrepancy disclosed)
S108 Well below the ~53% national average R2 CONFIRMED at secondary tier
S109 Texas excluded from the two leading national trackers for data-accuracy reasons R2 CONFIRMED at secondary tier
S110 Its own 2015 data shows ~95% duplicate-listing inflation R2 CONFIRMED at secondary tier
S111 §6/§9 headline: rule dead twice over; state law more durable but enforcement lags R2 CONFIRMED
S112 Vacated by two courts, 10-year moratorium, rescinded eff. Feb 2026 R2 CORRECTED on "10-year"; rest CONFIRMED
S113 38 states + DC already exceeded the old federal floor R2 OVERSTATED — see P108
S114 A 2025 Health Affairs 22-state panel: mandates raised direct-care staffing ~5%, no closure effect R2 CORRECTED (date) — published in the March 3, 2026 issue; findings CONFIRMED at author-institution tier
S115 New York: ~400 violating facilities, essentially no penalties for years R2 CONFIRMED at secondary tier
S116 PE ownership: +11% causal mortality, >7M Medicare patients R2 CORRECTED on the sample size (4.2M); effect CONFIRMED
S117 §7 headline: two GAO reports eight years apart confirm the same undercounting gap; fix not until 2027 R2 OVERSTATED — the two reports document different gaps (incidents vs. spending data)
S118 GAO-18-179 (2018): 26 of 48 studied states couldn't report critical-incident counts R2 CONFIRMED (GAO, verbatim: "Twenty-six state Medicaid agencies could not report…")
S119 GAO-26-107884 (2026) confirms the gap persists R2 CORRECTED — that report is on federal spending and Medicaid coverage; it makes no incident-reporting finding
S120 It flags the $12B Medicare+Medicaid AL spending figure as "likely an undercount" R2 CONFIRMED (GAO, verbatim)
S121 A CMS rule requiring annual state incident reporting doesn't take effect until 2027 at the earliest R2 CONFIRMED at secondary tier
S122 Medicaid's services-only AL footprint ~18–20% of residents R2 UNVERIFIABLE at primary tier
S123 No clean time series exists to quantify AL's own growth R2 CONFIRMED-INTERNAL (V5 negative, self-disclosed)
S124 §8 headline: valuation can't be pinned to one number; health/labor evidence is solid and getting worse R2 OVERSTATED on "getting worse" — red team #19 downgraded the trend to pandemic-confounded
S125 AARP's own figure moved $350B (2006) → $600B (2021) → $1.01T (2024) R2 CONFIRMED at secondary tier
S126 Two independent estimates "for a comparable era" (CBO $234B, RAND $522B) diverge by more than 2× R2 CONFIRMED-INTERNAL — and the correct framing the whitepaper drops (see P140)
S127 Cash & Counseling (1998–2003) is one of the only true RCTs in the domain R2 CONFIRMED
S128 Treatment caregivers reported less strain and better health R2 CONFIRMED at secondary tier (ASPE/Mathematica)
S129 The mechanism was closing an unmet-need gap, not efficiency R2 CONFIRMED at secondary tier
S130 Costs ran higher in year one (AR +17%, FL +14%) R2 UNVERIFIABLE at primary tier
S131 CDC BRFSS: caregivers' mental-health gap widened from 2015–16 to 2021–22 R2 UNVERIFIABLE at primary tier (cdc.gov blocked)
S132 §10 headline: Germany, Japan, Netherlands all under real current strain R2 CONFIRMED at secondary tier
S133 Germany's contribution rate rose again for 2025; federal loans needed to hold 2026 rates steady R2 UNVERIFIABLE — sourced to a German consumer-finance blog
S134 Workforce vacancies up 110% against 45% demand growth R2 CORRECTED (citation) — attributed to NBER/Geyer; NBER w31870 has no vacancy data
S135 Japan projects a 570,000-worker shortfall by 2040 and is already proposing copay increases R2 UNVERIFIABLE at primary tier — but the record's "proposing" is right where P161's "raising" is wrong
S136 The Netherlands projects a 266,000-worker shortfall by 2035 R2 UNVERIFIABLE at primary tier
S137 Japan's favorable reputation traces substantially to one clustered research lineage R2 CONFIRMED-INTERNAL (flag retained, completeness disclaimed per red team #20)
S138 §11 headline: C&C is the only architecture with real trial evidence; the rest measured research effort R2 CONFIRMED-INTERNAL
S139 Nine architectures on five anchored objectives, across three scoring passes R2 CONFIRMED-INTERNAL
S140 C&C is the only RCT-backed architecture and wins decisively under caregiver-relief weighting R2 CONFIRMED-INTERNAL
S141 Its own RCT documented a cost increase, correcting cost-containment 3 → 1 R2 CONFIRMED-INTERNAL
S142 A third pass, structurally blinded, moved 25 of 40 ranked cells R2 CONFIRMED-INTERNAL
S143 High scores required cited evidence while low scores accepted its absence R2 CONFIRMED-INTERNAL
S144 Twenty cells at 1 or 2 rested on no evidence; corrected, 30 of 40 sit at neutral R2 CONFIRMED-INTERNAL (independently recounted)
S145 The private-LTC-reform "ranks last with no exceptions" claim is withdrawn R2 CONFIRMED-INTERNAL
S146 Four of that row's five cells were empty; last place survives only under the unamended scale R2 CONFIRMED-INTERNAL
S147 The wage floor loses its workforce lead under both readings, tying the unstructured-cash comparator R2 CONFIRMED-INTERNAL
S148 "Deviations log … 21 entries" R3 CORRECTED — the committed log has 31
S149–S167 The 19 published deviations digest rows (covering entries 1–21 and 30–31) R3 CONFIRMED-INTERNAL as summaries of their source entries; entries 22–29 are absent from the published page
S168 Red team: 22 attacks run as a Workflow (four attack/defend groups plus re-score, reconciliation, synthesis) R4 CONFIRMED-INTERNAL
S169 "This log is the authoritative correction layer" R4 CONFIRMED-INTERNAL
S170–S191 The 22 published attack/verdict pairs R4 CONFIRMED-INTERNAL against red-team-log.md; substantive claims inside them are carried at S085–S147 and in ledger-record.md
S192 A fresh scorer blind to the original numbers re-scored all nine architectures R4 CONFIRMED-INTERNAL — with the S011/S197 correction attached
S193 Three cells differed by ≥2 points and were reconciled R4 CONFIRMED-INTERNAL
S194 C&C cost-containment corrected 3 → 1 R4 CONFIRMED-INTERNAL
S195 Federal AL standard workforce-sustainability corrected 1 → 3 R4 CONFIRMED-INTERNAL
S196 OAA/NFCSP cost-containment corrected 4 → 2 R4 CONFIRMED-INTERNAL
S197 Correction, 2026-08-10: the earlier re-score was instruction-blinded, not structurally blinded R4 CONFIRMED-INTERNAL
S198 The structurally blinded pass "moved 26 of the 40 ranked cells" R4 CORRECTED — it differed on 26 and moved 25; every other statement on both pages says 25
S199 All three earlier corrections were independently re-derived, so they stand R4 CONFIRMED-INTERNAL
S200 "Private LTC insurance market reform still ranks last under every weighting" is withdrawn R4 CONFIRMED-INTERNAL
S201 Four of that row's five cells rested on no evidence either way R4 CONFIRMED-INTERNAL
S202 Both readings published side by side in ws11-rescore-log.md R4 CONFIRMED-INTERNAL
S203 Receipts: 11-workstream protocol (9 executed to findings) Footer CONFIRMED-INTERNAL
S204 Receipts: deviations log (31 entries) Footer CONFIRMED-INTERNAL — contradicts S148 on the same page
S205 Receipts: full red team applied (22 items, 3 reconciled cells) Footer CONFIRMED-INTERNAL
S206 Receipts: second structurally blinded re-score (25 of 40 moved, one headline withdrawn, both readings published) Footer CONFIRMED-INTERNAL
S207 Status: "Red-teamed; PACE unscored" Footer CONFIRMED-INTERNAL

Public-page coverage: 425 extracted · 425 verdicts · 0 unaddressed.

← All Elder care research documents Sources digest Read the whitepaper