Appendix B: How America trains, licenses and limits its doctors and nurses

This appendix supports the sections "Doctors" and "Nurses and the rest of the workforce" in the plan. It explains the four bottlenecks the physician program targets and the constraints the nursing program targets.

The shape of the shortage

The United States has about 2.6 practicing physicians per 1,000 people. The average across the OECD's wealthy countries is about 3.7. Closing that gap would take roughly 370,000 more doctors, and the gap is not evenly spread: it is worst in rural counties, in poor urban neighborhoods, and in the specialties that do not pay well, above all primary care and psychiatry. Roughly a third of American physicians practice primary care, against half or more in peer countries.

The shortage is not a natural condition. It is the product of a pipeline with four valves, each of them set by a law or a private body, and none of them set with the country's need in mind.

Bottleneck 1: residency positions

A medical degree does not make a doctor. After medical school, a graduate has to complete a residency, a period of three to seven years of supervised practice inside a hospital, before practicing independently. The number of residency positions therefore sets the number of new doctors, whatever the medical schools produce. Two separate constraints limit how many positions exist.

The money. Residents are paid, and their supervision costs money, and since 1965 Medicare has been the largest funder of that training, through what is called Graduate Medical Education payment. In the Balanced Budget Act of 1997, Congress capped each hospital's Medicare-funded resident count at whatever it had been in 1996. The logic at the time was that the country faced a physician surplus, a view that was widely held in the 1990s and has since been abandoned by almost everyone who held it. The cap was written as permanent, and it has barely moved. Hospitals may train residents above their cap, but they must pay for those positions out of their own revenue. Congress added about 1,000 Medicare-funded positions in the 2021 appropriations act, phased in over five years, and roughly 200 more in 2023. Those are the first meaningful increases since the freeze, and they amount to a few percent of the annual intake.

The permission. The Accreditation Council for Graduate Medical Education, the ACGME, is a private non-profit organization that decides which hospitals may train residents at all, approves each individual residency program, and must sign off on any permanent increase in a program's size. It is not a government agency. Its standards are the standards; most state licensing boards require ACGME-accredited training, and the private specialty boards require it for certification. The government has no power to overturn its decisions. This is why the plan's first approach is to work with the ACGME, and its fallback is a second federally recognized accreditor with Congress conditioning federal funds on states licensing its graduates.

What neither constraint controls. The Medicare cap counts heads, not specialties. A hospital under a fixed cap chooses what to train, and because residents are paid far less than fully trained doctors while generating revenue in the procedures they assist, the incentive is to fill positions in high-revenue specialties such as radiology and orthopedics rather than in primary care. The cap therefore not only limits the number of doctors but tilts the mix toward the fields the country has enough of.

The one program that trains where it is needed. The Teaching Health Center program, created in 2010, funds residencies based in community health centers rather than academic hospitals, in primary care, in underserved places, and its graduates stay in those places at far higher rates than hospital-trained residents. It funds around 1,100 residents across roughly 80 programs at about $160,000 each, and it has to be reauthorized by Congress every appropriations cycle, so it has lived under the threat of lapse for its whole existence. The plan makes it permanent and scales it to a third of all new primary care positions.

Bottleneck 2: immigration

About one in six people entering residency each year is a foreign national who needs a visa, roughly 6,600 physicians out of some 37,700 entering positions in 2025. Most arrive on a J-1 exchange visitor visa sponsored by the Educational Commission for Foreign Medical Graduates, and the J-1 carries a condition written into immigration law: when training ends, the physician must return to their home country for two years before applying for most work visas or for a green card. The country pays through Medicare to train these doctors, they finish residency in American hospitals, and then most of them must leave.

The main way around the two-year rule is the Conrad 30 waiver, under which a state may sponsor a physician to stay if they commit to three years in a federally designated shortage area. Congress created it in 1994 with 20 waivers per state, raised the number to 30 in 2002, and has not changed it since. Thirty per state, in states of any size, has meant that some states exhaust their slots in weeks while others never use them.

A second, larger group of immigrant doctors arrives already trained, having practiced for years in another country. In most states such a physician cannot get a license without first completing an American residency, which means competing with new graduates for a scarce position and repeating three to seven years of training. A board-certified surgeon with a decade of independent practice is, for licensing purposes, a medical school graduate. Tens of thousands of foreign-trained physicians live in the United States working as nurses, technicians, or outside medicine entirely. A handful of states, beginning with Tennessee in 2023, have created alternative pathways with supervised practice in place of a repeated residency, and the plan builds a national version and ties residency money to states adopting one.

Bottleneck 3: medical school seats

Medical school enrollment has grown substantially since the early 2000s, after the Association of American Medical Colleges reversed its own earlier position that the country had too many doctors and called for a 30 percent expansion. New schools have opened and existing ones have grown, and the number of American graduates now exceeds the number of first-year residency positions available to them, which is why the residency cap is the binding constraint today.

Even so, a universal system needs more. MD programs are accredited by the Liaison Committee on Medical Education, jointly sponsored by the American Medical Association and the Association of American Medical Colleges; osteopathic programs by the Commission on Osteopathic College Accreditation, an arm of the American Osteopathic Association. A school must notify the LCME before any class-size increase that cumulatively exceeds 10 percent or 15 students, whichever is smaller, since its last full survey. The accreditor cannot forbid growth, but a school that grows faster than its faculty and clinical sites can support risks its accreditation, so in practice class sizes move at the accreditor's pace. The plan works with both accreditors and funds the faculty and sites that growth requires, siting new public schools where the shortages are.

Bottleneck 4: the license

There is no national medical license in the United States. Licensure is a state function, exercised by about 70 state medical and osteopathic boards, and each license is valid only within its own state. A physician licensed in Ohio cannot treat a patient in Indiana, in person or by video, without an Indiana license. The Interstate Medical Licensure Compact, launched in 2017, lets a physician who qualifies obtain licenses in participating states through an expedited process, but it is an agreement among states, not every state participates, and it produces multiple licenses rather than one. A surplus in one state does nothing for a shortage in the next.

There is one important exception. The Department of Veterans Affairs has for decades allowed a physician licensed in any single state to practice at VA facilities in every state, on the theory that a federal system needs a federal workforce. The plan extends that model: a federal license, recognized in every state, for physicians practicing in shortage areas under federal programs, for the relief corps, and for remote care.

The nursing bottleneck

Nursing has a different pipeline and a different constraint. There is no residency cap and no accreditor limiting growth. There is a shortage of teachers and of places to learn.

In 2025, according to the schools' own reporting, nursing programs rejected 93,176 qualified applications, nearly 17,000 of them to graduate programs, which is where nursing faculty come from. The reasons the schools gave were a lack of funding, not enough faculty, not enough experienced nurses available to supervise students on the ward, a shortage of clinical sites, and not enough classroom space. Every one of these is a resource problem rather than an applicant problem.

Faculty. A nursing faculty position now usually requires a graduate degree and often a doctorate, and a nurse with that credential earns substantially more in clinical practice than in a classroom. Faculty positions go unfilled, and an unfilled position caps the class. The plan pays faculty salary supplements tied to enrollment growth and funds graduate education for nurses who commit to teaching.

Preceptors. Clinical training happens on the ward under the supervision of an experienced nurse called a preceptor. That supervision is unpaid, and it is added on top of the preceptor's regular shift. Hospitals limit how many students they will host because every student is a cost to a nurse who is already busy. The plan pays a per-student fee to the facility and a stipend to the preceptor, and the same fix unblocks nurse practitioner, physician assistant, and midwifery training, which depend on preceptors too.

Simulation. Simulation means practicing on realistic mannequins and staged scenarios in a lab rather than on a ward. The National Council of State Boards of Nursing ran a multi-site randomized study, published in 2014, that found replacing up to half of traditional clinical hours with high-quality simulation produced the same outcomes on licensing exams and in the first months of practice. Many state boards still cap simulation well below that. The plan funds simulation centers on the condition that the state board raises its cap to what the evidence supports.

Placements. Every nursing school negotiates its own clinical placements with every hospital, one relationship at a time, and slots go unused because nobody can see them. The plan funds a placement exchange in every participating state, a registry of available clinical slots by unit and specialty with a scheduling system.

Staffing ratios

California passed the first statewide law setting minimum nurse-to-patient ratios in 1999, and the ratios took effect in 2004, at one nurse to every five patients on a medical-surgical ward and tighter in intensive care. Hospitals predicted closures. The research since has found that California nurses care for fewer patients than nurses in comparable states, that patient mortality is lower where staffing is better, and that nurse burnout and turnover are lower. The plan sets enforceable minimum ratios nationwide as a condition of Medicare participation, with funded transition time, and it treats the ratios as an ownership safeguard as well as a workforce measure, since a provider cutting quality cuts staffing first.

Scope of practice

Nurse practitioners, physician assistants, certified nurse-midwives, and pharmacists are trained and licensed to do more than many states let them do. About half the states allow nurse practitioners full practice authority; the rest require a supervision or collaboration agreement with a physician, which in a rural county with no physician means no practice. Pharmacists in most states may not prescribe for the minor conditions, contraception, or routine renewals they are trained to handle. These restrictions are set state by state, usually after lobbying by the physician societies, and the evidence that they protect patients is thin.

The plan uses three federal levers at once: it pays the same rate whoever delivers the service, so there is no financial reason to route a visit through a physician; it preempts contrary state scope-of-practice law for clinicians participating in the national program; and it conditions federal funding on states aligning their rules. This is the fastest capacity in the plan, because every one of these clinicians is already trained, already licensed, and already working.

Why the timeline is what it is

Nothing in the physician pipeline produces a practicing doctor in less than seven years from the day the money moves: four years of medical school and at least three of residency, and the residency positions have to be created and accredited first. That is why the plan starts training in year one and expects the first program-trained physicians in year eight, and why everything it can do sooner, telehealth at parity, re-entry for clinicians with lapsed licenses, full practice authority, pharmacist prescribing, foreign-trained physicians, and the relief corps, is scheduled for the first year. Those are the levers that add capacity from people who are already trained.