Optimal Care America

Final sponsor paper · Revised September 9, 2026
Evidence cutoff: August 16, 2026; sponsor scenario dated September 8, 2026
Status: final policy recommendation with a provisional planning envelope

When a person needs care, the system owes them 3 things: a door they can open, a path they can follow, and a human being who owns what happens next. Optimal Care America is built around that promise.

The proposal establishes a comprehensive national care floor and finances it through one public payer. National law would secure the benefit, patient rights, safety rules, portability, and accountability. Communities would decide how to deliver care within those rules. Patients and clinicians would remain at the center. A voluntary, separately governed data layer could support personalized medicine when the evidence, consent, security, and human oversight are strong enough.

This paper selects that architecture, called A1. It also sets a provisional 2024-dollar planning envelope of $5.0–$6.2 trillion per year, with $5.5 trillion as the base case. At the Centers for Medicare & Medicaid Services (CMS) resident denominator of 341.1 million, the range is $14,658–$18,176 per resident; the base is $16,124. The base sits $221.4 billion, or 4.2 percent, above the official 2024 National Health Expenditure (NHE) anchor of $5.2786 trillion.[5][10]

The numbers are a sponsor planning judgment, not a bottom-up actuarial result. The underlying program remains at 38 of 39 supported criteria. Original criterion 24 remains incomplete. The project still lacks qualified, independent acceptance of comparable B0 and A1 resource quantities, unit costs, covariance, and rerun evidence. Nothing here is legislation, sovereign consent, actuarial certification, a guaranteed saving, a final tax rate, or permission to implement.[6][7][10]

The recommendation is direct: build A1, budget for the base case, and use the high case to test capacity. Count savings only after patients can reach care, workers can deliver it safely, and the rights in this paper survive contact with the real system.

1. A promise the system cannot evade

An insurance card does not deliver care. A usable guarantee must connect each person to preventive, primary, urgent, specialty, hospital, behavioral, oral, pharmacy, rehabilitative, home, long-term, palliative, diagnostic, and emergency services. It must survive a transfer between institutions. It must work for a person who needs an interpreter, an accommodation, a supported decision, or help getting to the appointment.

The accepted Optimal Care America corpus supplies the operating discipline for that guarantee. It defines an outcome-first architecture, 11 patient-journey families, a rights and outcome constitution, a symmetric resource-accounting method, and a controlled evidence frontier.[1–7] These are checked designs. They do not prove that the country can yet staff A1 safely, that A1 will improve national outcomes, or that it will cost less than the current system.

The outcome constitution fixes the order in which the system must judge itself. Rights and authority come first. Survival and protection from catastrophic harm follow. Then come health, function, patient-defined outcomes, access, continuity, experience, equity, caregiver effects, resilience, and learning. Comparable resource cost comes last. A good national average cannot cancel a serious rights violation or hide harm inside a required subgroup. The constitution contains 39 metrics in 13 domains and leaves missing, incomparable, or underpowered findings open.[2]

The national care floor would turn those commitments into enforceable program conditions. It would protect bodily autonomy, informed choice, supported decision-making, refusal, privacy, dignity, accommodation, and accessible communication. It would require timely entry, safe handoffs, a named receiver, appeal, correction, and remedy. It would track avoidable death, serious harm, abandonment, delay, health, function, cognition, comfort, patient goals, caregiver burden, and unjust subgroup gaps.

A benefit that exists on paper but cannot be reached does not meet the floor. Neither does a referral that closes without a receiving clinician, or a discharge that leaves the patient to discover the next step alone.

The legal form remains open. Congress and the relevant governments would need to decide how statutory entitlement, administrative appeal, an independent ombuds function, injunctive relief, damages, and a private right of action fit together. Legal, ethical, disability, patient, clinical, labor, federalism, territorial, and Tribal review would shape that work. The obligation is settled here. The complete remedy structure is not.

2. Accountability follows the patient

Optimal Care America does not create one national delivery company. It creates one guarantee and connects 4 layers of care. Responsibility stays with the sending person or organization until a named receiver accepts it.

2.1 The person and the personal care team

Every person would have a named lead clinician, a care steward, a team identity, and a usable longitudinal plan. A name in a database is not attachment. The person must have a working route to the team, the team must accept responsibility, and its workload must remain inside a safe envelope.[3]

The personal care team would handle prevention, new symptoms, chronic illness, medication reconciliation, care-plan maintenance, and navigation. Patients could change teams, seek a second opinion, designate supporters, and record their preferences and communication needs. Emergency care and other services that become unsafe behind a primary-care gate would remain directly accessible. Where no safe permanent panel exists, a local access team would hold temporary responsibility and keep the capacity failure visible.

2.2 The Community Health Commons

The Community Health Commons is a set of local capabilities, not a prescribed building or ownership model. Its services include comprehensive primary and preventive care, 24-hour access and acute evaluation, behavioral and substance-use care, oral health, pharmacy, common diagnostics, maternity and newborn support, rehabilitation, home and mobile nursing, palliative care, interpretation, accessible communication, transport coordination, and social- care navigation.[3]

A Commons could be a public clinic network, a nonprofit system, a group of independent practices under contract, mobile services, a rural hub-and-spoke network, or a local combination. The national program would specify what the community must be able to do and who remains answerable. The community would choose the physical form.

Local operators would track patient clocks, unanswered test results, open referrals, staffed capacity, transport, supply constraints, and people at risk of being lost. A licensed bed without a nurse is not capacity. A referral without an acknowledgment is not a handoff.

2.3 The Regional Care Grid

The Regional Care Grid would connect emergency medical services, hospitals, specialists, procedural services, critical care, maternity, rehabilitation, long-term services and supports, laboratories, imaging, pharmacy, blood, and medical transport.[3] Regional operating bodies would maintain a live view of staffed capacity and coordinate transfers, scarce resources, disaster response, and recovery across organizations.

Regions should follow clinical travel times, referral patterns, population need, workforce, and redundancy. Existing corporate borders would not decide the map. Before a transfer, the sending team would identify a receiving clinician and site. The sender would remain responsible until the receiver accepts.

2.4 The National Expert Mesh

The National Expert Mesh would connect rare and complex care, reference diagnostics, evidence synthesis, scarce capabilities, and multi-region response.[3] It would move expertise to patients and local clinicians whenever possible. National consultation would strengthen the team closest to the person; it would not turn ordinary care into a remote command center.

The 11 machine-readable journey families cover prevention, new symptoms, urgent and emergency care, hospital episodes, maternal and newborn care, behavioral health, rehabilitation, medication management, oral health, long-term care, and serious illness. Every journey uses the same contract: usable entry, named ownership, acknowledgment, escalation, recovery, patient rights, and subgroup non-regression.[4] Payment, technology, and institutional ownership may change along the route. The clock and the duty do not.

3. One public payer for the care floor

One national public payer would hold sole financial responsibility for all services inside the guaranteed floor.[8][9] The policy would separate access from employment, insurer networks, household claims fights, premiums, deductibles, and point-of-care cost sharing. Payment administration would remain, but it would support care instead of organizing the clinical relationship around competing liability.

One payer does not mean one owner. Independent practices, public systems, nonprofits, academic centers, and other qualified organizations could deliver care. Local administration could help operate the program. Patients and clinicians would not have to determine which payer owes for the same floor service.

Mission-specific and sovereign systems may retain their own governance, delivery, and statutory duties. Their financial and operational connections would require lawful negotiation and, where applicable, government-to-government Tribal consultation. This paper assigns no authority to a Tribal Nation, the Indian Health Service, the Veterans Health Administration, military health, a territory, or any other government.

Existing federal and state health appropriations would enter the national pool or receive credit against it. An employer health contribution would replace most employer premium spending. A progressive household health contribution would replace premiums and cost sharing, with no contribution below a protected income threshold. High-income, capital-income, and health-harm tax bases would supply additional revenue. Credited public revenue plus new contributions must equal the enacted budget and reserve. Distributional and macroeconomic scoring must set the rates; this paper does not guess them.

Care teams would receive mainly prospective operating budgets adjusted for risk and need, with protections for access, quality, equity, and workload. Hospitals and regional networks would operate under global budgets. Public need review would fund major capital separately. Blended salary, episode, and limited fee payments could remain where they protect access and professional independence. The payer would negotiate medicines and devices against evidence, with explicit safeguards for shortage and useful innovation.

Billing should become a small public utility. Fraud control should pursue identity theft, false eligibility, conflicts, corrupt procurement, phantom services, and compromised data. It should not make routine obstruction of necessary care its business model.

The architecture is selected. Statutory language, contribution rates, incidence, intergovernmental credits, sovereign agreements, and provider- payment calibration still require public analysis. The $5.5 trillion base is a planning target, not a certified revenue requirement.

4. National rights, local care

The national minimum would guarantee comprehensive, clinically appropriate, medically necessary care across the full life course.[8][9] It would join 5 parts that fail when separated:

1. a capability floor covering preventive, primary, urgent, emergency, specialty, hospital, behavioral, substance-use, oral, reproductive, maternity, newborn, pharmacy, diagnostic, rehabilitative, home, long-term, and palliative care;
2. rights to accommodation, language access, consent, refusal, continuity, appeal, and remedy;
3. safety and evidence rules with independent review, protected reporting, and explicit treatment of uncertainty;
4. information standards that let patients and accountable teams use and correct records across settings; and
5. public reporting of availability, waiting, failed handoffs, safety, outcomes, subgroup performance, and recovery.

The standard would define what every community must provide. It would not force every community into the same building, employment model, or workflow. Services with uncertain evidence could use a controlled evidence-development path or an individualized exception. Every exclusion and medical-necessity decision would expose the rule, evidence, responsible reviewer, deadline, and independent appeal route. Urgent cases would use an expedited path.

Local delivery may vary when it meets or exceeds the floor. A public, time-limited variance could test a different method, but it could not waive a noncompensable right or reduce the benefit. Tribal Nations and sovereign health authorities are not local exceptions. Their powers and program interfaces require lawful sovereign arrangements.[9]

Supplemental private financing could purchase amenities, non-displacing convenience, and noncovered elective preferences. It could not purchase faster or preferred access to a scarce, clinically necessary floor service.[8][9]

The rule has operational teeth. A supplement could not pull a service out of the floor, waive safety rules, override allocation by clinical need, recruit enough workers to make the floor unavailable, or shift its cost and risk to the public pool. Providers working in both layers would report staffed capacity, waiting times, workforce movement, cross-subsidy, adverse selection, and subgroup effects. Regulators could suspend supplemental activity, order capacity restoration, and remedy patient harm when it weakens the floor.

Some lines will be contested. Amenity, convenience, elective preference, and clinical benefit do not always separate cleanly. Later legal and evidence work must set access thresholds, monitoring powers, remedies, and an expedited appeal. Allowing supplements does not prove that the public floor has enough workers or facilities.

Clinical judgment stays visible

Patients and accountable clinicians would make clinical decisions under transparent standards. The payer could audit eligibility, payment integrity, and program compliance. It could not become an invisible clinical gatekeeper. When medical necessity is disputed, the system would identify the governing rule, evidence, reviewer, appeal route, and deadline. Safety rules would keep urgent care moving while the dispute proceeds.

Institutions and tools may support a decision. They cannot dissolve ownership. A specialist who accepts a question owns it until answering it or returning an acknowledged plan. A receiving organization owns a transfer after acceptance, not before. A person leaving the hospital keeps a named follow-up owner.

Clinical-standards power would be divided. Multidisciplinary panels inside an operationally independent institute would draft standards through public evidence review. A public National Care Standards Board would approve binding floor standards under duties of rights, evidence, equity, feasibility, transparency, and conflict control. A separate surveillance and learning office would watch outcomes and trigger scheduled or urgent revision.[9]

Patients, caregivers, practicing clinicians, workers, disability and civil- rights bodies, communities, states, territories, and sovereign governments would have defined challenge routes. Meetings, funding, rationales, material conflicts, recusals, and minority reports would be public. An independent patient ombuds or administrative-review office could order fast reconsideration or a temporary safety stay.

Federal, state, territorial, professional, and sovereign authority cannot be settled by an organization chart in a policy paper. Appointments, removal, funding, emergency procedure, preemption, and judicial review require qualified legal and governmental work. Representation on a national body does not replace government-to-government consultation or consent by a Tribal Nation.

5. Personalized medicine must earn trust case by case

The long-term clinical aim is AI-enabled personalized medicine: rapid multimodal assessment, diagnostic support, and individualized treatment that serve patients and physicians rather than an opaque claims intermediary.[8]

A patient and care team could choose to assemble a governed clinical profile from history, examination, medicines, laboratory results, imaging, pathology, genomics, symptoms, goals, and selected device or home observations. An AI tool could organize the record, surface possible explanations, identify missing information, connect findings to evidence, and compare reasonable treatment paths. It would support a decision. It would not issue an invisible denial, make a final diagnosis without accountable review, or replace consent.

The tool should leave behind something the patient and physician can inspect: the data it considered, important gaps, plausible alternatives, confidence and uncertainty, relevant evidence, known limits, the human decision owner, and a monitoring and escalation plan. Sometimes the right output is “not enough information.” The system must be able to say it and find appropriate help.

The accepted corpus supports named accountability, consent, correction, audit, safe downtime, independent safety review, second opinions, subgroup non- regression, and an uncertainty-aware decision process.[1–4] It does not prove that a national personalized-AI layer will improve care, reduce cost, lighten workloads, keep data secure, or perform evenly across populations.

Participation is voluntary

People who contribute the data needed for personalization could use that layer. People who decline would keep the complete standard care floor.[8][9] Direct care would use only the minimum data reasonably necessary under lawful authority. Personalization, research, and model improvement or training would require separate affirmative choices. Public-health use would require specific, necessary, proportionate, auditable legal authority.

Refusal could not reduce eligibility, timeliness, clinician attention, diagnostic diligence, appeal rights, or the quality of standard care. Consent would distinguish direct care, personalization, research, model improvement, public health, and commercial use. Patients could inspect and correct source data, see material uses, and stop future optional participation without losing the ordinary record required for safe care.

The program would prohibit the sale of identifiable health data, health-data advertising, penalties for refusing an optional use, and reuse outside the stated authority or permission. A commercial partner could not borrow consent given for direct care or research. Any permissible commercial use would need its own affirmative opt-in, independent review, and an unconditioned standard- care route.

Withdrawal would stop new optional collection and reuse. Optional copies would be deleted or de-identified when law and technology permit. Records needed for safe continuity and legal duties would remain. Consent would disclose limits involving backups, completed research, and trained models before participation, then document what happened after withdrawal.

Data generated during ordinary care differs from genomic data, continuous device feeds, cross-domain linkage, secondary research, and model-training reuse. One permission cannot honestly cover all of them. Access also requires more than a consent form. People who lack money, broadband, accessible tools, language support, geographic access, or well-measured data should not receive inferior personalization by default.

A human owns every consequential decision

A named, qualified human would review and own any AI-influenced decision that materially affects serious diagnosis, treatment, triage, transfer, discharge, eligibility, coverage, authorization, denial, risk classification, sensitive data, or a use that is uncertain, changed, degraded, disputed, or outside its validated purpose.[9] AI could not make the final decision on diagnosis, treatment withdrawal, coercive action, discharge, transfer, denial of a floor benefit, or an appeal.

The responsible person or team would decide whether the tool fits the case, review its output, explain material uncertainty, document disagreement, watch the result, and escalate possible harm. Patients could learn when an AI output materially shaped care, demand human review, seek a second opinion, correct the data, and challenge an adverse decision.

Clinicians need not reproduce a model's mathematics. They do need reliable information about its intended use, limits, validation, changes, known failure modes, and output provenance. Care must remain safe when the tool or network is down. Human review also needs testing. Workload, competence, independence, disagreement, and rubber-stamping determine whether oversight is real.

Purely clerical tools could omit case-level notice only when they cannot change clinical meaning, priority, eligibility, access, rights, or the responsible human decision. They would still remain inventoried, auditable, monitored, and subject to escalation.

6. Health data carries a duty of repair

Health data can make care continuous and personal. The same data can expose a person for years. Governance must begin with purpose, authority, and remedy, then build the security program around them.

A national data charter would separate the minimum record for safe direct care from permissions for personalization, research, model improvement, public health, and commercial activity. It would give patients access, correction, provenance, segmentation, delegation to trusted supporters, and a usable record of material disclosures and model uses. It would require role-based access, encryption, logs, independent testing, incident response, recovery, and safe offline workflows.

The charter would also govern retention, deletion, vendors, model changes, subcontractors, audit access, conflicts, and exit rights. Its authority would remain jurisdiction-specific, including Tribal sovereignty and sovereign data governance.

The sponsor retains a narrowed security safe harbor.[8][9] An organization could receive protection from private class damages for an external intrusion only if it proves, with evidence created before the incident, that it met a current risk-based national standard. It would also need independent assessment, continuous monitoring of material controls, data minimization and segmentation, prompt response, honest notice, and practical mitigation. “We followed industry custom” would not be enough.

The safe harbor would not cover unauthorized use, sale, disclosure, or processing; intentional, reckless, grossly negligent, deceptive, or discriminatory conduct; overcollection; broken consent; concealed or delayed notice; false certification; conflicts; an ignored material weakness; repeated failure; or a failure to mitigate. It would preserve regulatory and sovereign enforcement, statutory duties, individual actual damages, class-wide injunctive or declaratory relief, judicial review, correction, monitoring, deletion where possible, and security improvement.

A no-fault compensation route, funded by covered data holders, would support people harmed by qualifying failures. Statutory minimum recovery would remain available for defined high-risk violations. A public body would update the national standard independently and scale it to data sensitivity and foreseeable harm.

This liability direction is contested. It is not settled law, settled ethics, or accepted project evidence. Actual damages can be hard to prove when the harm is probabilistic, delayed, dignitary, collective, genomic, or biometric. Qualified legal, ethical, cybersecurity, civil-rights, patient, and sovereign review must define protected claims, statutory recovery, the compensation fund, medical monitoring, certification, preemption, and federal, state, territorial, and Tribal powers.

7. Build capacity before moving patients

The dates below are targets, not predictions. Public readiness gates control every irreversible move. Political deadlines do not waive rights, continuity, or safe workloads.[6][7]

2027: write the promise into law

Draft the National Care Charter, benefit floor, appeals and remedies, standards institutions, payer authority, financing bases, transition authority, and reporting duties. Score how the $5.0–$6.2 trillion envelope affects households, employers, governments, and the economy. Begin formal state, territorial, labor, disability, civil-rights, patient, clinical, and government-to-government Tribal processes. The first product is law and negotiated authority, not a technology contract.

2028–2030: build capacity before moving people

Create the payer and standards bodies. Map every community and region against the Commons, Grid, and Expert Mesh capabilities. Expand primary, behavioral, oral, home, rehabilitation, maternity, pharmacy, and long-term-care capacity. Fund training and capital. Establish portable records and safe downtime. Publish staffed-capacity gaps. Existing coverage remains until the receiving system can own the work.

2031: open the national floor where readiness is proved

Begin enrollment and payment conversion only in regions that pass public gates for usable entry, team responsibility, emergency access, specialty handoffs, medicines, continuity for active treatment, workforce load, accessible communication, appeals, cybersecurity, and cash flow. Patients keep clinicians who are willing and able to continue safely. Organizations keep plural ownership. The payer changes before the clinical relationship does.

2032–2034: complete the universal transition

Expand region by region under independent stop, repair, and rollback authority. Insurance enrollment alone does not qualify a region. Graduation requires measured access, outcomes, rights, subgroup performance, safe workloads, reliable supplies, and closed handoffs. Legacy administration ends only after claims, appeals, payments, and continuing care reconcile.

2035 and after: earn the personalized future

Operate the universal floor first. Release voluntary AI-enabled uses one clinical purpose at a time after lawful authority, external and local validation, subgroup safety, human ownership, consent, security, accessibility, vendor control, remedy, monitoring, and downtime all pass. Maintain a complete standard-care path for people who decline and for every outage or suspension.

The floor and the optional data layer have separate release gates. They may share infrastructure, but floor eligibility never depends on optional data. A failed readiness test stops the launch date, not the patient's rights.[9]

8. Budget for the care promised

Final 2024 National Health Expenditures were $5,278.6 billion, or $15,474 per resident, using the CMS accounting boundary and a population denominator of 341.1 million. NHE records expenditure. It is not a clean bottom-up count of labor, goods, capital, and reserve resources.[1][5][6]

The sponsor planning envelope keeps that denominator and 2024 dollars. It sets the mature annual cases at $5.000 trillion, $5.500 trillion, and $6.200 trillion. Per-resident figures are $14,658, $16,124, and $18,176. Compared with the B0 accounting anchor, the cases are -5.3 percent, +4.2 percent, and +17.5 percent. The low case's minus sign is an accounting comparison between unequal evidence bases. It is not a savings promise.[10]

The base case refuses the cheapest headline. It funds universal entry, safer staffing, behavioral and oral health, home and long-term care, public-health resilience, capital renewal, information and safety work, and a reliability reserve. It books no prevention dividend. It reinvests part of any administrative simplification in access and accountability.

Where the base case goes

These are planning allocations, not forecasts. They force the design to fund quiet but necessary work instead of hiding it in a promise.[10]

The amounts total $5.5 trillion. Administration does not disappear. Eligibility, payment, reporting, audit, appeals, standards, safety, and daily operations still take people and systems. The reform replaces duplicated claims friction with visible public administration and clinical support.

What moves the estimate

Clinical demand and service intensity can move the base by about ±$330 billion. Labor compensation and realized productivity move it from -$275 billion to +$385 billion. Universal-access take-up contributes -$165 billion to +$275 billion. Medicines, devices, and supplies contribute -$220 billion to +$275 billion. Long-term-care scope moves it from -$110 billion to +$220 billion; administrative productivity, from -$165 billion to +$110 billion; and capital and reserve needs, from -$55 billion to +$165 billion.[10]

Those movements cannot be added mechanically. The low and high cases are joint stories with correlated assumptions. The high case is wider because unmet need, worker scarcity, long-term care, and reserve requirements are more likely to reveal missing costs than to make them vanish.

The published model reproduces every displayed number. Reproducible judgment is still judgment. Original criterion 24 remains incomplete because B0 lacks a comparable bottom-up resource range, A1 quantities and unit costs have not won independent acceptance, covariance remains provisional, no qualified reviewer has rerun the model, and nationwide outcome and rights non-regression have not been observed.[6][7][10]

This paper selects A1 as the policy frontier. It does not claim an independently certified numeric frontier. The envelope is strong enough to design legislation, workforce and capital plans, and validation tests. It is not strong enough to waive those tests before an irreversible transition.

9. Ten decisions now govern the design

The sponsor-design phase closes with 10 decisions. The August 23, 2026 record keeps the rationale, tradeoffs, dependencies, and replacement instructions.[9]

1. Care floor. Guarantee comprehensive medically necessary care across the life course, with evidence-development, individualized-exception, and appeal routes.
2. Payer. Give one national public payer sole financial responsibility for the floor. Preserve mission-specific and sovereign governance through lawful arrangements.
3. Supplemental care. Permit amenities, non-displacing convenience, and noncovered preferences. Do not sell a faster path to scarce, necessary floor care.
4. National and local authority. Make entitlement and rights uniform. Allow local delivery to vary when it meets or exceeds the floor. Preserve sovereign authority.
5. Clinical governance. Divide drafting, public approval, challenge, and revision. Publish the evidence, conflicts, representation, and dissent.
6. Consent. Use the minimum necessary data for direct care under lawful authority. Require separate opt-ins for personalization, research, and model improvement. Constrain public-health and commercial use.
7. AI accountability. Put a named human behind every consequential AI- influenced decision. Give patients notice, explanation, correction, human review, and challenge. Do not automate final action over core care or rights.
8. Breach remedies. Limit the security safe harbor to a proved, strong program facing an external intrusion. Disqualify misconduct. Preserve enforcement, injunction, and compensation.
9. Sequence. Give the universal floor and optional personalized layer separate legal, operational, and readiness gates. Build useful shared infrastructure in parallel.
10. Evidence. Use the provisional $5.0–$6.2 trillion envelope for planning. Keep original criterion 24 open. Make no guaranteed-savings, certified-tax- rate, or implementation-readiness claim.

A later sponsor can change any decision by number. The revision must preserve the superseded record, reconcile every dependent section and model input, and pass the builder and independent checks again.

Conclusion

The United States already spends roughly $5.3 trillion a year on health care. The honest planning judgment is not that a new system will pay for itself. It is that the country should be prepared to spend about $5.5 trillion to turn coverage into care people can reach.

If mature A1 costs less, the public earns that dividend only after outcomes, rights, safe workloads, and reserve capacity pass. If it costs more, the shortfall should appear in a public budget before it appears in a patient's life as a queue, an abandonment, or unpaid care forced onto a family.

Money can be counted at the center. Care cannot be delivered there. It happens between people, in rooms and homes and ambulances, during ordinary visits and terrible nights. National scale should guarantee that relationship, move help toward it, and refuse to let responsibility disappear when the patient crosses an institutional line.

Build the floor first. Make the next responsible person impossible to lose. Let AI earn narrow authority one clinical use at a time. And never call it a saving when someone else paid with health, dignity, or time.

Notes and sources

1. U.S. Centers for Medicare & Medicaid Services, National Health Expenditure Data, final 2024 historical tables.
2. Optimal Care America outcome constitution, metric dictionary, and non-regression rules; summarized in the public technical record accompanying this release.[11]
3. Optimal Care America A1 operating architecture; summarized in the public technical record accompanying this release.[11]
4. Optimal Care America patient-journey contracts; summarized in the public technical record accompanying this release.[11]
5. Optimal Care America observed-baseline evidence and source ledger; CMS is the primary source for the national expenditure anchor.[1][11]
6. Optimal Care America symmetric resource-model and frontier controls; summarized in the public technical record accompanying this release.[11]
7. Optimal Care America release and closeout checks; summarized in the public technical record accompanying this release.[11]
8. Sponsor directions dated August 23, 2026.
9. Sponsor default decision record dated August 23, 2026.
10. Final sponsor planning envelope, September 8, 2026. All displayed scenario arithmetic is reproducible from this machine-readable file.
11. Public technical record, including artifact summaries, verification status, and integrity hashes.