Last year, America’s total medical costs hit a new record of $3.4 trillion, according to the federal government. That’s about 18 percent of the country’s total GDP, meaning that one out of every six dollars we spent in 2016 went to health care. The national doctor bill dwarfs anything else we spend money on, including food, clothing, housing, or even our mighty military.

If that $3.4 trillion were spread equally throughout the population, the bill would come to some $10,350 for every man, woman and child in the country. But fortunately –for most of us, anyway—the cost of health care is not equally distributed. Rather, a small number of Americans run up most of the expense. The biggest medical costs are concentrated on a fairly small segment of the population—people with one or more chronic illnesses, plus victims of accidents or violent crime. The cost is so concentrated, in fact, that an estimated five percent of the population accounts for 50 percent of total medical costs.

For the purposes of this project, we’re calling these people The Platinum Patients—they’ve also been described as “super-utilizers” or “frequent fliers.”

This concentration of total cost on a small segment of the total population is reflected in another common aspect of medical spending: the concentration of treatment, and cost, in the end of a life span. For most people, the vast majority of all the health care they’ll ever get comes near the hour of death. Hundreds of billions of dollars each year are spent treating Americans who are in the last weeks, or days, of life.

The old Marx Brothers’ joke—“I wouldn’t dare go to the hospital—people die there all the time”—is essentially true. Many people die in the hospital—in many cases, just after they’ve incurred a hugely expensive round of surgery, treatment, and medication. About one-third of Americans undergo operations in the last month of life.

If these issues were subject to hard, cold economic theory, a health-care system would probably distribute spending differently. The large sums it costs to keep a sedated cancer patient with dementia alive in a hospital bed from age 94 to 95 could presumably be directed instead to provide, say, a kidney transplant for a 40-something victim of renal failure, or a young woman who is too depressed to care for her baby. That money could be used for pre-natal care for uninsured mothers, setting the stage for both mother and child to have a healthier and happier life. Or, those funds could be used to provide health insurance at reasonable cost to the 29 million Americans who have no health coverage today.

One famous, or perhaps notorious, advocate of limiting late-in-life medical spending is former Colorado Gov. Richard Lamm, who was given the nickname “Governor Gloom” in the 1980’s for his argument that the elderly have a “duty” to avoid costly care when the end is near. There’s only so much money available for medical care, Lamm noted, so it ought to be used in the most efficient way. In the face of bitter criticism, Lamm stuck to his guns. Just this spring he told the Denver Post: “When I look at the literature, and there are such things as $93,000 prostate operations at some stage of prostate cancer that might give two extra months of life, it is outrageous.”   

The problem with these straightforward economic calculations is that they involve real human beings who have friends and relatives. That 94-year-old cancer patient, after all, may have loving children or grandchildren at the bedside; hardly anybody is willing to let Grandpa die just to save money for the overall health-care system.

The issue of allocating medical spending is most acute in the United States, because we spend far more on treatment and medication than any other country. All the other developed democracies on the planet guarantee health care for everybody (citizen or alien), and yet they spend, on average, about half as much per capita as the U.S.

But all over the world, health systems are struggling with the same concentration of cost that plagues the U.S.

So Britain created an organization to make rules for how its healthcare money is spent. It’s formally called the National Institute for Health and Clinical Excellence, but everyone knows it by its acronym: “NICE.” This outfit issues guidelines to the regional medical authorities on what should be covered, and what shouldn’t. Should a 94-year-old get a hip replacement? Should a terminal cancer patient be given a course of medication that costs $40,000 and extends life an average of four months? (In Britain, the answers are, generally, “No.”)

In one widely-reported case, the NICE guidelines said that a pub waitress—a mother of three—who contracted breast cancer should not receive the drug Herceptin. After all, NICE noted, the medication costs about $36,000, and doesn’t usually help with that woman’s particular form of cancer. Since there is only a finite amount of money in the National Health Service budget, the agency said, it would be smarter to spend those thousands on a treating another patient with a better chance of recovery.

As pure economics, this made sense. As politics, it was a disaster. The waitress’s case became a national scandal. The tabloid headlines savaged the agency: “Not so NICE—Mum Left to Fight Cancer Without a Pill.”

For systems that are looking for smarter ways to allocate limited funds, health-care economists have created a pair of measures to determine which treatments or drugs are worth paying for. The “Quality-Adjusted Life Year,” or QALY (pronounced “quolly”) and the “Disability-Adjusted Life Year, or DALY (“dolly”), are supposed to steer health-care dollars in the direction that provides the greatest quality of life. These ratings would say, for example, that spending money to keep an aging, asthmatic Alzheimer’s patient on life support for 9 months is not as useful as spending the same money for 9 months of pre-natal care for a poor, uninsured mother-to-be.

Americans who are not health-care economists tend to resist the concept of QALYs and DALYs because they lead the system not to pay for one person’s health care in order to pay for another’s. This is considered “rationing” of health care, and rationing is generally condemned under a variety of names, most memorably as “death panels.”

In fact, though, every nation rations health care every day. No country—not even the richest oil sheikdoms—can afford to pay for every advanced surgical procedure and every costly drug that modern medicine knows how to provide. Accordingly, health-care systems are constantly making choices—rationing—about which treatments to pay for.

The United States, too, rations health care. Just ask any of the 29 million uninsured Americans who generally can’t see a doctor or pay for a prescription until they’re sick enough to go to the emergency room. But the U.S. does its rationing in a different way.

In other rich countries, there’s a basic floor of care that everybody gets, which means there’s a ceiling as well—the system simply won’t pay for certain drugs or procedures. In the U.S., millions of people have no floor except the emergency room, and others have no ceiling. With the right insurance plan, there’s almost no limit to what money can buy in American health care, regardless of the age or condition of the patient. And so we continue to spend huge sums on that small, generally elderly segment of the population with chronic illnesses, while millions have no health insurance.

One approach to this quandary that seems promising, both for the individual patient and for the health-care system overall, is the concept of “death with dignity,” as reflected in the Hospice movement. Hospice was initially a British idea that has spread to France, the U.S., and other advanced democracies. It’s a system that emphasizes caring, not curing, that replaces the all-out battle against death. In essence, the surgeries and the IV tubes and the breathing machines are replaced with a calm acceptance that one’s time is coming.

A patient in Hospice avoids the operating room and the hospital ward, spending the final weeks or days of life at home or in a quiet facility, often with a regimen of drugs to control the pain of disease. In the U.S., most of the people who commit to Hospice are elderly, but it’s a path that terminally ill patients sometimes choose in their 20’s or 30’s. For the ailing individual, and for friends and relatives, it provides a more tranquil opportunity to reflect and say goodbye than the hurly-burly and confusion of a major hospital. For a health-care system, it can be a massive money-saver.

Accordingly, the two big government health-insurance plans, Medicaid and Medicare, both provide payment for hospice services. One section of the Affordable Care Act (“ObamaCare”) says that physicians can now be paid for an office visit in which they discuss end-of-life options such as Hospice; this was the provision that Sarah Palin famously denounced as “death panels.”

Still, most people facing serious illness avoid Hospice and place their bet instead on the marvels of modern medicine. The physical result is often positive; doctors today can cure diseases that were considered terminal just a few years ago. However, the fiscal impact of these miracle cures is increasingly painful for national treasuries.

As other countries have found, there’s no simple solution to the problem of concentrated health-care costs. But one step that could clearly help in the U.S. would be a commitment, at long last, to provide health care for every American. All of the rich countries that guarantee health care for everybody have better health outcomes at much lower cost than the U.S. This is not a coincidence: a comprehensive system of universal care will always be cheaper and more effective than the haphazard, crazy-quilt network of overlapping and costly payment systems America is stuck with today.

And when everybody is covered, the health-care system can probably make fairer decisions about where the money should be spent. If America is going to pay $3.4 trillion for health care, after all, we ought to make sure that every American benefits from that colossal expenditure.