This is a beta. Every person named in this document is real, and every story comes from published reporting or from a book we cite by name in the text. Nothing is a composite and nothing is invented. Where a figure comes from a survey or a study, we say which one. If you think we have a fact wrong, tell us.

Arguments about American health care are usually arguments about who pays. This document is about something else: what is actually wrong with the system, for the people who need it and the people who work inside it.

There are nine problems. They are not a list of grievances; they are connected, and they compound. Insurance that does not cover much, in a country that has not built enough hospitals or trained enough clinicians, staffed by people who are prevented from doing the job properly and punished for saying so.

We have left out one large subject on purpose. This is not about how much America spends in total, or what share of the economy health care consumes. That matters, and we will publish it separately. Every problem below would still be a problem if the national totals were half what they are.

1. Insurance that does not cover you

The familiar number is the uninsured. The more important number is the insured who cannot afford to use their insurance.

The Commonwealth Fund's 2024 biennial survey found that 23 percent of working-age American adults are underinsured: they had coverage all year, and still faced out-of-pocket costs or deductibles large enough, relative to their income, that the coverage did not do what coverage is for. The great majority of them are not in some marginal individual-market plan. They are in employer plans, the kind most Americans think of as good insurance.

The consequences show up immediately in the same survey. Fifty-seven percent of underinsured adults said they had gone without needed care because of the cost. Forty-four percent were paying off medical or dental debt.

This is the part that surprises people who have not been through it. A deductible is not a discount; it is a bill you pay in full before the insurance begins. A network is not a directory; it is the set of doctors whose care your plan will pay for, and every other doctor in the country is a financial hazard. Someone with a $4,000 deductible and $600 in savings does not have health coverage in any sense that matters in March. They have a promise that becomes useful in November, if they are unlucky enough to reach it.

The delay this produces is not a failure of willpower, and it does not look like one from the outside. It looks like negligence.

Brian Alexander's book The Hospital follows a man named Keith Swihart in Bryan, Ohio, and the sequence is worth understanding because it is ordinary. Swihart had a job, a 401(k), a wife, a daughter, and insurance. He also had type 2 diabetes. He sneezed, the vessels in his left eye ruptured, and he lost most of the sight in it. The specialist who could treat him was in Toledo, and he could not drive to Toledo because he could not see. By the time he was properly examined his blood sugar had reached 900, roughly 300 points past the level that can cause a diabetic coma. He had a bone infection and gangrene running down his foot.

The staff treating him were kind, and Alexander records what they were also thinking: how could he let this go so long, why won't he comply. Swihart believed he was complying. He had seen some black on his foot and thought it was a bruise. It did not hurt, because the nerves had already gone.

Being insured, employed, and careful was not enough, and from the outside his catastrophe looked like his own fault.

2. Care that is not there to be covered

Insurance is a promise to pay for care. It is not a promise that the care exists within reach. In much of the United States, it does not.

The clearest measurement is maternity care, because birth cannot be postponed. In the March of Dimes' 2024 report, 35.1 percent of American counties are maternity care deserts: no birthing facility, no obstetric clinician, none. Half of all US counties have no hospital that delivers babies at all. More than five million women of childbearing age live somewhere with no maternity care or nearly none, and the trend is the wrong way, because more than a hundred hospitals have closed their obstetric units.

The economics driving that are not hidden. Obstetrics is expensive to staff around the clock and badly paid by Medicaid, which covers a large share of births. A service that loses money and cannot be scheduled is the first thing a struggling hospital drops.

The same logic empties whole categories of care out of whole states. Alaska has no burn center. A person badly burned in Anchorage is flown to Harborview in Seattle, 1,445 miles away, and reaching Anchorage from elsewhere in Alaska can be another 800. Alaska has no Level I trauma center; its two Level II centers are both in Anchorage. No one decided Alaskans should not have burn care. A burn unit simply cannot generate the volume to pay for itself there, and nothing in the system pays for things that cannot pay for themselves.

Rural hospitals are closing on the same calculation of volume against margin: 14 shut in 2018 and 18 in 2019. Ellwood City Hospital in Pennsylvania missed payroll while its board was still debating a strategic plan.

A hospital that is not there cannot be brought back by giving people better insurance cards.

3. A workforce that was deliberately kept small

The shortage of American clinicians is often described as though it were a demographic accident. It is not. In its two most important bottlenecks it is the direct result of policy, and the policy is still in force.

Doctors first. A medical graduate cannot practice without completing a residency, and residencies are funded chiefly by Medicare. The Balanced Budget Act of 1997 capped the number of residency positions Medicare would fund at each teaching hospital, freezing them at that hospital's 1996 level. The reason Congress did this is the part worth sitting with: it had been advised to expect a surplus of physicians. Medical schools have expanded considerably since. The number of federally funded slots to train those graduates has barely moved, and a few hundred positions have been added since in small tranches.

Nurses second, and here the bottleneck is teachers. American nursing schools turned away 92,672 qualified applications in 2025, and 80,162 the year before, according to the American Association of Colleges of Nursing. These were not weak candidates; they met the requirements. The schools lacked faculty, clinical placement sites, preceptors, and classroom space. There were 1,588 unfilled full-time nursing faculty posts in 2024, and more than a third of current nurse faculty are expected to retire within a few years.

So the country has a shortage of nurses, a surplus of qualified people who want to become nurses, and not enough instructors to convert one into the other. That is a solvable problem that has not been solved.

Read the two together and the shape becomes clear. The United States is not short of people who want to do this work. It is short of the training capacity to let them, and it has been short of it by decision rather than by accident.

The debt is the third turn of the screw. A doctor now typically leaves training owing a sum that would have been unimaginable a generation ago. Wendy Dean and Simon Talbot, in If I Betray These Words, describe a primary care physician who left the University of Cincinnati in 1976 owing $4,500, about $22,000 in today's money, and who spent years unable to recruit a younger partner into his practice, because the young doctors he met were carrying debt that made a small primary-care practice financially impossible to take on.

4. Permission to be treated

Between an American patient and the treatment their doctor has ordered sits a review process, and it is worth being precise about what that process is, because it is not what most people assume.

Prior authorization began in the 1990s as a check on a small number of genuinely expensive procedures. Elisabeth Rosenthal, a physician who now edits KFF Health News, traced what it became in An American Sickness. Insurers hand the reviewing to outside contractors. Some of those contractors charge the insurer nothing, because they are paid out of a share of the money they save by making treatment difficult to obtain or refusing it. The incentive is not to determine whether care is necessary; the incentive is to reduce approvals, because approvals are the cost and denials are the revenue.

Having established that machinery, insurers extended it downward, to the ordinary and the cheap. Rosenthal records a New York cardiologist, Dr. Barry Lindenberg, receiving a prior-authorization request for a patient's refill of an old generic heart medicine, and faxing back: ARE YOU KIDDING ME?

Two features of this process do most of the damage. The first is that the reviewer is a stranger who has not examined the patient and often does not practice in the relevant specialty. The second is that delay works as denial: care postponed long enough stops being useful, and a decision that takes three weeks has decided the matter regardless of what it says.

The case that made this concrete for the public involved a seventeen-year-old named Nataline Sarkisyan, and it is documented from an unusual angle, because the man running communications for her insurer later wrote about it. In Deadly Spin, Wendell Potter describes CIGNA's refusal to cover a liver transplant that UCLA was ready to perform. CIGNA's own transplant case manager had reviewed the file and recommended approval within three days. A CIGNA medical director, who had never met or examined the girl, wrote to her doctors that the transplant was experimental. Potter's account of his own calendar in those weeks includes preparing the company's Investor Day at a Manhattan hotel, where CIGNA would tell analysts it expected to earn more than a billion dollars that year.

For clinicians, the same system is a tax on the working day. Rosenthal followed the Manhattan cardiologist Dr. Richard Hayes seeking approval for a stress test for a patient with chest pain: the insurer's website said Hayes was in network, the contractor said he was not, and Hayes ended up on the phone himself for more than half an hour reciting a dozen codes and findings to get authorization for a test that would pay him between $600 and $1,200. After more than 20 years in solo practice dealing with 21 insurers, he closed it in 2014.

5. Prices that no one will explain

American hospitals do not have prices in the sense that other businesses have prices. They have a chargemaster, an internal list of many thousands of items with numbers attached, and those numbers bear no stable relationship to what anything costs to provide. Insurers negotiate discounts off the list, and the size of the discount depends on the relative bargaining power of the insurer and the hospital rather than on the care. A patient with weak coverage, or none, can be billed the undiscounted list price, which is the highest number in the building.

Because the list is not public in most states and not explicable anywhere, the same procedure at the same hospital has several different prices depending on who is asking, and no one involved can say why any of them is the number it is.

Steven Brill documented what this does to an ordinary person in America's Bitter Pill. Emilia Gilbert was sixty-one and drove a school bus for about $1,800 a month. She slipped in her own yard in Fairfield, Connecticut, went to the emergency room at Bridgeport Hospital with a bleeding nose, waited most of the night, and left with half a dozen stitches. Her three CT scans were billed at $6,538; Medicare would have paid about $825 for all three. She was billed separately for instruments, bandages, and IV tubing, which are supposed to be covered by the emergency-room facility charge, itself $908. The bill came to $9,418, and because her insurance had strict limits, no discount applied. A judge required her to pay it off in weekly installments for six years.

When Brill asked a senior vice president at the hospital's parent system about those charges, the executive said the chargemaster was fair, and then, after checking, said he had no way of explaining how any of the numbers had been arrived at.

At the other end of the scale the same opacity produces decisions no family should make. Brill followed a couple through 11 months of cancer treatment that generated $902,000 in bills, including 28 separate invoices from a single hospital. The wife, whom he calls Alice, paid about $30,000, still owed $142,000 a year after her husband's death, and sold an inherited family farm to clear it. Somewhere in the middle they began weighing treatment against cost, and she described the end of it plainly: the doctor said the next round of chemotherapy might extend his life by a month and would be very painful, and her husband gave up.

6. Collection practices that do not even pay

When patients cannot pay these bills, hospitals pursue them, and the methods are harsher than most Americans realize. The physician and historian Luke Messac cataloged them in Your Money or Your Life. The industry term is extraordinary collection actions, and they include garnishing wages, emptying bank accounts, foreclosing on homes, and asking a court to order the arrest of a patient who fails to appear. Messac's point is that none of this is exotic any more.

It also lands on people the hospitals' own rules say should never have been billed. Roughly half of nonprofit hospitals send bills to patients who qualify for charity care under the policies those hospitals filed themselves. Johns Hopkins sued its own low-wage employees and garnished their pay when they could not afford care at the hospital where they worked.

The defense offered for all of this is financial necessity. That defense can now be checked. A 2017 study of Virginia hospitals that garnish wages found the practice returned, on average, one tenth of one percent of hospital revenue. The hospital that sued more patients than any other in the state recovered two tenths of one percent. Hospitals that took their patients to court collected less in a year than a health system chief executive is typically paid.

The study's senior author was Marty Makary, a surgeon at Johns Hopkins, and his conclusion was that the argument that hospitals must do something this ugly to stay afloat is not supported by the data.

This is the one part of the system that cannot be defended as a hard trade-off, because it does not accomplish anything. Families lose houses and wages and occasionally their liberty so that institutions can recover a rounding error.

7. Care delivered by people who are not given enough time to deliver it

Nursing is where the shortage of clinicians becomes a shortage of attention, and the mechanism is simple. Nurse staffing is a hospital's largest controllable cost. Adding a nurse to a shift costs money immediately and produces savings that are real but diffuse and delayed. So staffing is set close to the minimum that the ward can be argued to tolerate, and the difference is absorbed by the nurses.

Theresa Brown, an oncology nurse, wrote a book called The Shift about a single 12-hour day on her floor. The most useful sentence in it is about the difference between three patients and four:

With four patients they sometimes become human to-do lists. Three allows me to treat my patients as people. Instead of rushing from room to room I can move at a human pace and also be on top of everything going on with them.

One extra patient is the difference between a person and a list. That is the entire staffing argument, made by someone doing the work, and it explains why ratios are fought over so bitterly.

Rosenthal traces where the pressure came from. A veteran hospital administrator described watching head nurses become clinical nurse-managers through the 1980s, setting staffing by statistical analysis of payment. What was lost, she said, was the head nurse who fiercely protected the patients on her ward and did not give a damn about the financials.

The smaller economies run underneath. Brown describes the unpaid 30-minute meal break that nurses are scheduled for and rarely take, because no one is free to cover their patients; class actions have been filed over it, and by her account essentially every nurse on her floor should legally be paid for lunch on every shift.

Outside hospitals the same squeeze is worse, because the work is barely paid at all. More than 20 million Americans care for a family member without pay, and about 2.2 million home care aides work on the books, nearly all women and disproportionately women of color. In Who Will Care for Us?, Paul Osterman describes an aide called Monica who was the only person who knew one patient across his home, his hospital admission, and his nursing home, and who arrived at his discharge with his pill bottles and the observation that he seemed short of breath at night. She was the continuity in his care, and she is the part of the system that no record counts and no budget pays properly.

The standard explanation is that nobody wants these jobs. Wyoming tested it: between 2001 and 2004 the state raised total compensation for home care aides in one program from $9.08 to $13.19 an hour, and turnover fell from 52 percent to 32 percent. North Dakota and California saw the same. It is a wage problem that has been mistaken for a workforce problem.

8. Clinicians who know what is needed and are not permitted to provide it

For 20 years the profession's distress was labeled burnout, and that word placed the fault in the doctor: insufficient resilience, poor self-care. Institutions built programs on that premise. Burnout inventories were used to identify individuals and manage the risk they represented, and those identified were directed to mindfulness and wellness courses that had no evidence behind them. Declining could be recorded as being disruptive.

Wendy Dean and Simon Talbot, both physicians, borrowed a better description from military psychiatry. Moral injury was defined by the psychiatrist Jonathan Shay, writing about Vietnam veterans, as the betrayal of what is right by someone who holds legitimate authority in a high-stakes situation. Applied to medicine, in their book If I Betray These Words, it produces the plainest statement of the problem we have found:

As doctors, we know what our patients deserve but often cannot provide it because the business of medicine gets in the way.

The distinction is not academic. Burnout is a condition of the worker, and the remedy is to fix the worker. Moral injury is an injury inflicted by the institution, and the remedy is to change what the institution requires.

What the institution requires is mostly documentation, and the record system that was meant to relieve that has intensified it. Arnold Relman edited the New England Journal of Medicine for over a decade. In 2014, writing in the New York Review of Books about his own hospital care after a fall, he reported that neither of his physicians seemed to be in charge of his care or spent much time at his bedside, and that both left long notes in his computer record that he described as repetitious boilerplate and lab data with no coherent account of how he was actually doing. Conversations with his doctors, he wrote, were infrequent, brief, and hardly ever recorded.

A man who had edited the most influential medical journal in the country could not get his own doctors to talk to him.

9. And clinicians who cannot safely say so

The last problem is the one that keeps the others in place. The people best positioned to describe what is wrong are employees, and they can be dismissed.

Most American physicians now work for hospitals, systems, or staffing companies rather than for themselves. Studies in 1998 and again in 2013 confirmed that they could be dismissed with little or no stated reason and without meaningful due process. Many institutions instruct staff that speaking to a reporter is a terminable offense. In 2020, two physicians in Mississippi and one in Washington State were fired after pressing for stronger safety measures at the start of the pandemic.

What that produces is a system whose failures are visible only to people who cannot afford to describe them.

The exception is instructive. Ray Brovont ran the emergency department at Overland Park Regional Medical Center in Kansas, a hospital owned by HCA and staffed by EmCare, a physician staffing company owned by private equity. From 2012 he told EmCare that the arrangement was unsafe: for 18 hours a day one physician covered the emergency room and was simultaneously responsible for emergencies everywhere else in the building. The hospital expanded in 2014 without adding staff, and in 2016 opened a 10-bed pediatric emergency room that would have no dedicated pediatric emergency physician overnight. One doctor was now responsible for every adult emergency, every overnight pediatric emergency, and every cardiac arrest in the building.

The predictable thing happened. The single physician on duty was called upstairs after midnight to resuscitate a patient who had stopped breathing. While he was there, a man in the emergency room waited 45 minutes with a major heart attack. From arrival to opened artery took over two hours against a 90-minute standard, and the patient died three days later.

Brovont pressed harder and was fired. He sued for wrongful termination, litigated for more than four years, and in 2021 a Missouri court of appeals upheld an award of $26 million. Before he was dismissed, an EmCare executive explained his obligations in a sentence that needs no gloss: you cash the check every month to be a corporate representative, and there is a responsibility to support the corporation's objectives.

A doctor had to win a multi-year lawsuit to establish that he was allowed to say one physician cannot be in three places at once.

What this adds up to

The nine problems are one problem seen from different positions. Coverage that does not cover, in places where the care is missing, staffed from a workforce the country declined to train, gated by reviewers paid to say no, priced by numbers nobody will defend, collected by methods that recover almost nothing, delivered by people given too little time, who are prevented from practicing properly and cannot safely complain.

None of this is caused by a shortage of money. The United States spends more on health care than any country in history. It is caused by spending that money on the wrong things: on deciding whether to pay rather than on care, on pursuing people who have nothing rather than on staffing the ward, on prices that cannot be explained in daylight rather than on building the hospital that is not there.

The care these people needed was not merely unaffordable. Much of the time it did not exist.

That is a problem you fix by building. Here is how.