18% of GDP, worst outcomes in the OECD
Coverage is half-solved and access is not. Roughly 92% of residents hold insurance, and raising that number does not reach the remaining problems: a plan can deny, a network can exclude, a bill can arrive after the fact, and in about a third of counties the care the plan is a claim on does not exist. Insurance is a claim on care that exists.
That spending runs $1 trillion/year more than any peer nation's, and the outcome gap spans life expectancy, infant mortality, maternal mortality, and chronic disease burden. 28M uninsured. 45M underinsured. Family premiums average $23,968/yr. Workers earning $60K spend 20%+ of income on healthcare once cost-sharing and denials are counted. Hospital consolidation has eliminated competition in 75% of metropolitan markets. Defensive medicine — tests ordered primarily to avoid malpractice liability — is widespread; published estimates of its national cost vary too widely to state one. Billing administration and coverage denial operate as tolls on care itself; the Accord retires both by design.
The incentive structure is wrong at every level: fee-for-service rewards volume, not outcomes. CMS projects 20.6% of GDP by 2034 under current law. Every other developed nation achieves better outcomes at 9–12% of GDP.
Prices are the fiscal case. Commercial plans pay hospitals an average of 254% of Medicare rates, and nonprofit status does not restrain that — tax status governs profit distribution, not pricing behaviour. Administrative simplification is real and is roughly one-sixth of the lever. Any proposal that reaches peer-nation costs on paperwork alone is arithmetically wrong.
Comprehensive across categories — $0 premiums for the floor
Enrollment is automatic, from birth, and deductibles are $0 as well — supplemental tiers are separately priced. Coverage spans medical, mental health, SUD, maternity, and prescription care; point-of-care cost sharing (copays, sliding scales, low-income waivers) is set by AHQB on the clinical evidence. The table states the boundary plainly:
- Hospital care — inpatient and outpatient
- Emergency care
- Primary care
- Specialty referral
- Maternity and newborn care
- Mental health and substance-use treatment
- Prescription drugs on the AHQB formulary
- Preventive care with $0 cost sharing
- Basic dental prevention + emergency extraction
- Emergency vision exam
- Hearing screening
- Capacity payments for rural trauma + pandemic readiness
- Reduced or zero point-of-care cost sharing
- Comprehensive adult dental
- Vision correction (exams, lenses, frames)
- Hearing aids
- Private room where available
- Faster elective scheduling — within the same wait-time ceiling
- Premium prosthetics
- Employer may pay it as a benefit — taxed as compensation, no new exclusion
One federal payer buys the floor through four payment methods, each matched to a cost structure rather than to a provider type. Any provider of any ownership form participates under whichever method fits each service line, so the mix in a region follows local capacity, geography and population — not who owns the building. Delivery stays plural: private practices, nonprofit systems, public hospitals, tribal systems, and a standing public arm of Regional Health Authorities built on the community-governed health centre model. The Veterans Health Administration remains a veterans' system clinically — while supplying the institutional chassis the Authorities run on: Title 38 personnel authority and national credentialing, which took decades to build and cannot be procured.
Every county is assigned to exactly one Authority. There is no unassigned territory, which is what ends deserts by definition: a desert is a county no system wants, and under universal assignment there is no unwanted county — only one whose assignment carries a capacity payment. Deserts are attached rather than built, and that is a far smaller programme than chartering an institution in every underserved county. It is also the anti-dumping rule, because an Authority with a standing catchment obligation cannot improve its position by moving a patient out.
Four tests decide what the floor carries. It pools cost nobody can bear alone, maintains the capacity to work and raise a family, mitigates losses it would otherwise pay for later, and maximizes what one generation owes the next. The tests run in both directions: an annual physical for a healthy adult satisfies none of them. Corrective lenses and hearing devices are paid as a set amount toward a working appliance at long intervals, so the family keeps shopping and can spend more if it wants — but the eye exam and the fitting are free, because the judgement happens at the appointment and the price discipline belongs on the object.
The government pays for nearly everything and operates a minority of it. At maturity the federal share of health spending is about 90%, and the federal share of delivery is 15–20% — concentrated where nobody else will go. Most hospitals never change hands; their economics change and their ownership does not. There is no purchase obligation and no bailout, which is what keeps this from becoming a bridge for holders of distressed assets.
A catchment is an accounting unit, not a fence. Any qualified provider can treat anyone, and care delivered outside your home region settles between Authorities at the national schedule — behind the counter, with nothing for the patient to do. Where a procedure has a strong volume-outcome relationship, referral to a high-volume centre is automatic and the travel is covered, which makes the choice a wealthy patient already makes the default for everyone with the same diagnosis.
How an Authority reaches its catchment is its own call, made continuously rather than scheduled from Washington. Telehealth, mobile units and permanent buildings are all available and all take time to stand up. A building takes longest — not only to construct, but to establish that a permanent structure is the right long-term answer for that location at all.
Variable, discretionary professional and ambulatory services, on a national schedule with regional adjustment and public maximums — private insurers today pay hospitals ~254% of Medicare rates (RAND). Site-neutral payment removes the premium a hospital can charge for care an office could deliver. The standup follows Maryland's all-payer model, the tested on-ramp.
Standby capability carries fixed costs that per-encounter payment cannot cover — ambulance service, obstetric readiness, trauma standby, air transport. It buys named capability: hours, staffing, equipment, transfer agreements, drills. It does not pay for an empty building. Where a local unit cannot be sustained, funded transport is the backstop.
Longitudinal primary care on a demographic base, with a restricted audited condition set, mandatory outlier reinsurance, standing coding-intensity adjustment, and blended payment — so neither stinting nor upcoding has a clean payoff, and no provider gains by dumping sicker patients.
Hospitals in concentrated markets receive predictable annual revenue adjusted for population, service mix, social risk, quality and access — removing the reward for volume expansion and the leverage behind market-power pricing. Maryland and the AHEAD model are the live precedents.
Optional supplemental — Medigap for every age, on the same clinical floor
Above the floor sits a regulated supplemental market on the pattern ~65 million Medicare households already know from Medigap. It buys two distinct things. On services the floor already covers, it buys money and convenience: reduced or zero point-of-care cost sharing, a private room where available, faster elective scheduling within the same wait-time ceiling. Above the floor, it buys benefits the floor does not yet reach — comprehensive adult dental, vision correction, hearing aids, premium prosthetics.
Same doctors and the same AHQB clinical standards throughout. On any service the floor covers, every patient gets the same standard of care, and no plan may sell priority for urgent or medically necessary treatment. Where supplemental reaches above the floor it is buying real care that the floor has not yet absorbed — and that list is precisely the ratchet's migration queue. The essential layer belongs to everyone; competition begins above it.
The market is community-rated and guaranteed issue with no medical underwriting: premiums vary only by age and geography within statutory bands, never by health status, and no one can be turned away. Plans are standardized into comparable tiers under AHQB supervision — a comparison table, not a confusion market. An individual can buy a tier directly, or an employer can pay for one as a benefit — taxed as ordinary compensation, with no new exclusion: rebuilding the employer-insurance tax exclusion would rebuild the distortion this architecture removes.
The one-way ratchet. The AHQB charter requires an annual review of whether any supplemental benefit should migrate into the floor — the test is falling unit cost or strengthened clinical evidence, and migration only ever runs toward the floor, never out of it. (Illustrative: a benefit like hearing aids becomes a migration candidate once reference pricing has cut its unit cost — whether and when is the Board's evidence-based call, made under its published, appealable process.) The ratchet is why the federal share of the total rises over time as the floor absorbs more benefits at lower unit costs, while total system spending stays inside the 15–16% of GDP planning range.
Independent authority over coverage, pricing, and clinical safe harbor
The American Healthcare Quality Board is structured like the Federal Reserve — independent, Senate-confirmed with staggered terms, insulated from political direction. It holds three authorities over the floor.
Safe-harbor clinical guidelines. Documented protocol compliance is a complete statutory defense against malpractice liability. That removes one important incentive for defensive testing: when a clinician follows and documents an AHQB evidence standard, the statute protects the evidence-based choice. How much the country saves as a result is uncertain — the literature on how far malpractice pressure actually moves imaging and testing volume is mixed — and no savings from this mechanism are separately credited to the central Distributed Healthcare score.
Reference pricing at 120% of an international benchmark, drawn annually from Germany, Canada, Australia, France, Japan, the UK and the Nordics. The VA has run this for pharmaceuticals since the 1990s at 40–70% below commercial prices.
Evidence-based coverage and cost-sharing decisions, with published rationale, comment periods, and formal appeal rights for every limitation.
The Board also supervises the supplemental market's tier standardization and rating rules, and runs the annual floor-migration review that operates the ratchet.
Two stages, both measured on access-adjusted spending
The Healthcare Cost Brake is one of the macrogovernors. Both its stages are measured on access-adjusted spending — cash outlay plus the value of care demanded and not delivered. That basis is the whole design: a cash-only trigger can be satisfied by care simply not happening, which lets a delivery failure register as a cost success.
A warning at 17.4% of GDP, or sustained drift off the legislated spending path, prompts diagnosis, price negotiation and correction of the affected payment method. Services already failing their access standard are protected from cuts.
An outer backstop at 19.61%, set above the worst modeled scenario, suspends the benefit ratchet and forces indication and procurement review.
Action is cost-side only. There is deliberately no automatic economy-wide fee clawback, because a broad cut deepens the workforce shortage that caused the overrun in the first place. Only if containment fails does the Debt Sunset Governor fire.
The ramp phases who enrolls when, at one unchanging rate
What phases is enrollment. The rate never does. When a tranche goes live, its workers move onto the floor and its payroll levy replaces FICA plus premiums at the full 28% (10.5% employee / 17.5% employer) from its first day — there is no reduced introductory rate, for anyone, ever. Tranche assignment runs inside the ten-year phase-in on employer size and average payroll and on region (enrollment follows delivery capacity — VHA and Post Office 2.0 site density plus the COMPASS shortage score — with a quarterly published readiness map); the NSB and AHQB hold rule-making authority to add a benefits-level axis. Three size tranches:
Self-employed, gig workers, employers under 20 workers, and every currently uninsured American.
Employers of 20–500 workers.
Employers over 500 workers, plus the federal workforce.
The enrollment-priority order is statutory — deliberately the inverse of the usual protect-the-comfortable-incumbents pattern, and built so no governor can game it:
By benefit composition, the floor launches comprehensive across categories as listed above, and widens only through the one-way ratchet. Every region-tranche cell is gated on four capacity metrics — a failed gate pauses that region and tranche only, never the national rollout:
- Provider availability per 10,000 enrollees
- Wait-time stability — no more than +10% over any 90 days
- No increase in emergency-department diversion
- AHQB quality-metric stability
What is measured, what is modelled, and what is neither
Every claim below carries its evidence status, because a modelled number and a measured one should not be read in the same voice. Observed is existing US, VA or CMS data. External model is an estimate from outside the Accord, cited as an independent comparator. NAA model is our own output. Unscored mechanism means the mechanism is supported and its national magnitude has not been established — nothing is credited to the score.
2024 national health expenditure, $5,283B.
CMS National Health Expenditure Accounts, 2024
VA Clinical Resource Hubs recorded over 1.2 million encounters across 46 clinical services in FY2024, up from roughly 482,000 in FY2021, and served 83% of VA medical centers. The hubs exist to share clinicians across sites with local shortages.
Limit: The rural share of these encounters is not published here. A widely repeated 'nearly 40% rural' figure could not be verified at source and is not used.
Among roughly 11,800 veterans with chronic kidney disease (2,147 receiving telenephrology, 9,678 not) over 2021–2024, telenephrology management was associated with 15% lower mortality than primary-care management alone, and with greater use of guideline-directed medications. Hub-and-spoke design.
Limit: OBSERVATIONAL, not randomised. Selection into the programme can produce a mortality difference on its own. It supports the proposition that networked federal clinical capacity converts geographic coverage into clinical benefit. NO Distributed Healthcare mortality figure inherits from it.
Clinical Journal of the American Society of Nephrology, May 2026
Pandey, Wells, Ye, Fitzpatrick and Galvani model the Medicare for All Act from a 2024 baseline of $5,278.6B and project $4,237.4B, a reduction of $1,041.2B or 19.7%.
Our conversion: The paper states no percent of GDP for its projected system. Against 2024 GDP of about $29.35T, $4,237.4B is roughly 14.4%. That conversion is ours, and it should never be quoted to a second decimal — the figure moves with the denominator.
Limit: It models INSURANCE ARCHITECTURE AND PAYMENT ONLY — no change in care delivery, clinical staffing or organisational structure. So it is a genuine independent comparator for the cost side and says nothing about the Accord's delivery architecture. It is also more aggressive on prices than the Accord is, assuming a 51% pharmaceutical reduction and including universal dental, so part of the gap to its lower endpoint is assumption rather than mechanism. It corroborates; it does not validate.
Pandey et al., medRxiv, posted 2026-07-24, version 1 — PREPRINT, NOT PEER-REVIEWED
Two independently constructed Accord models bracket 14.5–17.5%. The planning range is externally corroborated near its lower edge by a separate 2026 model built on different assumptions.
Limit: A planning range, not a guaranteed outcome.
Following and documenting an AHQB evidence standard is a statutory defence, which removes one incentive to order tests for legal cover. Measured behaviour supports the incentive being real: a hospitalist study found 28% of medical orders placed defensively, and 93% of surveyed Pennsylvania physicians reported practising defensively.
Limit: Published estimates of the national dollar cost vary across several multiples and the empirical literature on how far malpractice pressure moves testing volume is inconclusive. NO dollar figure is published and NO saving is credited to the central score unless it is independently scored later.
The mechanism is supported by the VA evidence above. The national magnitude has not been established for the Accord.
Limit: Nothing is credited to the score.