Skip to content

The Power and Peril of the ICU

Reevaluating critical care in the wake of Covid-19

“Next week,” I recall an infectious disease colleague prophetically stating: that was when Covid-19 would hit us. It was March 2020, and I was on service as the attending physician in the ICU of the Boston-area hospital where I still work today. We were chatting with a group of medical residents about the latest pandemic news coming out of New York, gathered unmasked in the small workroom where the residents wrote their notes, made phone calls, ate lunch, gossiped. It was business as usual in our ICU that day, but hospitals in New York City were already seeing waves of critical coronavirus patients. I remember feeling a bit skeptical at the precision of my colleague’s prediction for when ours would crash to shore.

But they were right. The onslaught was right ahead of us, and I soon found myself intubating our hospital’s first patient with failing lungs due to Covid-19. We didn’t know the diagnosis at the time because back then all our SARS-CoV-2 tests were sent to the Massachusetts Department of Public Health, where they would be processed over the course of several painstaking days. Though I had not yet encountered the condition, my suspicion was high. This patient had all the typical signs: acute flu-like symptoms followed by shortness of breath, both lungs bright-white with pneumonia on X-rays, and very low oxygen levels in the blood. My colleagues and I donned head-to-foot PPE and followed the newly instituted precautionary protocol meant to limit the dispersion of viral particles during an intubation, keeping only a few staff members in the room at a given time. The procedure went smoothly: I put the patient to sleep and paralyzed their muscles, then advanced the plastic breathing tube down between the vocal cords into the trachea. The respiratory therapist turned on the ventilator; the nurse started a drip of heavy sedatives. Then, we waited with hope that their lungs would begin to heal with each passing day.

Of course, what began with one patient soon became a deluge. In the coming weeks, patient after patient came to the ICU with similar telltale symptoms. Our ICUs quickly filled up; like many hospitals, we converted a post-operative recovery area into an ICU to expand capacity. The days were long and sweaty under our zipped-up white suits, face masks, and N95 respirators. Patients were often desperately sick. Many did not survive. And we had to deliver terrible news to families again and again.

Yet, despite the intensity and the tragedy, I cannot say that I have ever felt so professionally useful and, indeed, satisfied in my clinical work as I did then. After all, fighting a pandemic of severe pneumonia is what the ICU does best: even as so many patients could not be saved, still others made their way through the most arduous phase of this illness and survived, some with many years ahead of them. Each time that happened, I experienced the opposite of burnout—I was energized by the realization of just how miraculous intensive care medicine could be. Some of these patients I have followed up with, many months later, in my outpatient clinic, and while their lungs have not always fully healed, their lives are full and underway. In a strange sense, Covid showed us what the ICU was made to do. But today, as we envision health care in a post-pandemic world, we need to grapple more honestly with another reality; the ICU has a dark side too.

Breath of Life

The clinical innovation that has saved so many people from dying of Covid-19 is arguably less than seventy years old. Its story begins, in the version usually told, one late summer day in Copenhagen, when a twelve-year-old girl named Vivi came down with a headache and stiff neck. Her arms were weak, and she developed a fever. Soon, her breathing became labored, so she was taken to Blegdam Hospital for an evaluation.

They had reason to worry: the year was 1952, a historic epidemic of polio was sweeping the city, and Vivi was its latest victim. At 9:45 a.m. on August 27, a physician at the hospital examined her. Her condition was clearly worsening. The physician observed that her chest was barely rising and falling with each breath. An X-ray showed the collapse of her left lung. Practicing in twenty-first century Boston, I’ve never treated a patient with polio, but poring over a translation of Vivi’s medical records, her course of physiological deterioration is familiar; she could have stopped breathing at any moment.

Contrary to what pop culture would have you believe, there are vanishingly few eureka moments in medicine, but this was one of them.

But she didn’t: at the very last moment, an anesthesiologist named Dr. Björn Ibsen made a desperate intervention to save the young girl’s life. A surgeon cut a small slit in her neck and placed a breathing tube directly into her trachea. Then, Ibsen attached a bag to the end of the tube, pumping air into her ailing lungs by hand. He kept at it for hours and hours, but it was not easy; Vivi was observed to be “kicking,” “retching,” and suffering “spasms and agitation.” At nearly 1 a.m., her skin was turning blue from reduced oxygen intake. But she was still alive, unlike the dozens of patients whose lungs had failed in the days before her. Ibsen administered powerful sedatives and muscle relaxants so that he could continue mechanically breathing for her, increasing her oxygen and expelling carbon dioxide from her body. Eventually, the crisis passed.

Contrary to what pop culture would have you believe, there are vanishingly few eureka moments in medicine, but this was one of them. Dr. Ibsen and his colleagues used so-called positive pressure ventilation on every polio patient with respiratory compromise that came into the hospital. To accomplish this, they needed a massive quantity of labor because each patient had to be hand-ventilated, around-the-clock, by someone squeezing a bag. So, they recruited some one thousand medical students—about two hundred each day—to act as methodical bag-squeezers until each patient perished or survived. With the passage of time, “spontaneous breathing,” as we call it, could begin again for some.

Ibsen’s innovations are typically described as the birth of the ICU; although in this sense of the word, the ICU is not any one thing. It’s more or less a logistical arrangement—a scenario bringing together specialized personnel and life-support technologies into one physical space so they may treat the most critically ill patients with greater ease. As the translator of Vivi’s medical records, Louise Reisner-Sénélar, put it in Intensive Care Medicine, Ibsen’s greatest contribution was to recognize that the hospital had to be “reorganised to treat respiratory failures in a multidisciplinary centralised unit.” In other words, getting patients like Vivi out of the woods required a ready supply of people to administer breathing tubes and drugs, monitor her physiological response, pump the bags, and determine when to change course. It required a unit.

Manual ventilation has long since been automated by machines, but the setup pioneered by Ibsen is remarkably similar to today’s equipment. So, too, are its inadequacies: as Covid-19 swept the nation, we, like the doctors at Blegdam, worried about personnel shortages—a gap some hospitals went so far as to plug with volunteer medical students, who, perhaps, were delegated with tasks slightly less monotonous than squeezing bags. Just as it was in mid-century Denmark, many patients died despite strenuous efforts: one study found that mortality rates among mechanically ventilated patients with Covid-19 were nearly 50 percent. And, both then and now, for those who survive, the path ahead is not easy.

Vivi never recovered from quadriplegia and did not regain the ability to breathe on her own; she required mechanical ventilation for the rest of her life. She remained hospitalized for some seven years and eventually left to live with her parents in 1959. She returned to Blegdam in 1971 with severe pneumonia—a common complication in those with chronic respiratory failure—and died two and a half days later, at the age of thirty-one. But her resuscitation by Dr. Ibsen during her childhood gave her many more years of life, some of it in her home. Indeed, as a result of the innovation at Blegdam, patient outcomes from severe polio dramatically improved: mortality fell from 87 to 40 percent; the new approach saved approximately one hundred lives in Blegdam alone. Of course, for the 40 percent who could not be saved, tracheostomy and ventilation merely prolonged suffering—the kicks, agitation, and spasms would all be for naught. But that was, and is, the price that must be paid when practicing critical care medicine without a crystal ball.

In other words, the core purpose of an ICU—as Ibsen once conceived it, and as I do today—is to buy time. To buy enough time for the lungs to heal and the body to recover, so that the joys of life can resume. But that makes the ICU a gamble that does not always pay off: sometimes, time is not enough.

Capital-Intensive Units

Inspired by the success at Blegdam, and the development of the modern ventilator and other technological medical advances, ICUs began to proliferate in hospitals across the world, bringing patients back from the brink of once universally fatal conditions. But in the United States, the story was a bit more complicated. ICU growth in this country has been driven by clinical need and innovation, but also by our unique, often profit-oriented health care financing system.

As the historian Gabriel Winant discussed in his book The Next Shift: The Fall of Industry and the Rise of Health Care in Rust Belt America, both private and public insurance in the United States financed the historic transformation and expansion of the nation’s health care infrastructure in the latter half of the twentieth century. In particular, when the obvious problems with tethering insurance to jobs were partially addressed by the passage of Medicare and Medicaid in 1965, hospitals lobbied for—and ultimately won—a major perk: the ability to bill the federal government for “capital expansion,” allowing them to build seemingly endless new wings containing endless new beds and expensive equipment, practically for free. At the same time, technology and treatment steadily improved, driving more demand for critical care, which was increasingly used for more health conditions and post-surgery care. As Winant puts it, “hospitals could pass through their costs—including capital investments—to the third parties that footed the bills, Blue Cross and the federal government most significantly. The existence of a larger and more advanced hospital plant, in turn, encouraged more use.”

Sometimes providers are put in a position of providing invasive care that they believe to be not merely fruitless but unethical.

In the United States, the expansion of the ICU was explosive: in the late 1950s, only a quarter of the nation’s largest hospitals—and only a fraction of smaller hospitals—had an ICU. But by the mid-1970s, ICUs were found in virtually all of the nation’s mid- and large-size hospitals. In the early 1980s, profit-driven medicine exploded: more than ever before, hospitals began behaving like bona fide businesses. When Medicare overhauled its payment system and pivoted from paying costs per day to a lump sum based on diagnosis (as I previously described in Baffler no. 52), high-tech and capital-intensive treatments became the most lucrative. Hospitals that took the most cues from the corporate world brought in the profits that allowed them to gobble up their competitors and invest in moneymaking tech and equipment. Even as the overall number of hospital beds fell, ICU care was on the rise: from 1985 to 2000, the number of ICU beds rose by 26 percent, from 69,300 to 87,400.

However, with the shift to lump-sum hospital payments in the 1980s, hospitals found that ICUs, which contained some of the most technologically advanced services offered, could run afoul of their optimal calculus for resource expenditures and return. Bluntly put, the longer a patient stays, the more it costs. The function of an ICU, you’ll recall, is to buy time, and as the old adage goes, time is money. One consequence has been the rise of LTACHs, or long-term acute care hospitals: off-site, mostly for-profit facilities that provide long-term critical care. The LTACH industry grew in leaps and bounds after its emergence in the 1980s: from 1997 to 2006, the number of LTACHs increased by about 9 percent annually. Over time, LTACHs became lucrative pickings for for-profit corporations, which now dominate the sector. One of the largest such corporations, Kindred Healthcare, was recently acquired by private equity firms, mirroring trends in the larger health care economy.

Today, decades after their proliferation across the globe began, the United States has more ICUs per capita than almost anywhere else on Earth. Notably, that supply came in use during this pandemic: without it, our death toll from Covid-19 may have been even more abysmal than it is. But outside the context of a pandemic, too often that overabundance has influenced the ways that care is delivered: Americans are more likely to spend time in an ICU compared to those in other nations. A dated study found that our ICU admission rate is fivefold higher than Canada’s and ninefold higher than the UK. In part, this may be because we sometimes overuse the ICU for patients who are not sick enough to benefit from its care: one study found that about a quarter of patients were in the ICU for monitoring but were “otherwise receiving care that could be provided outside of the ICU.” That’s not a bad thing per se—closer nursing attention is valuable for many patients. At the same time, in another study, among patients with four common medical conditions, those admitted to the ICU were more likely to receive invasive procedures relative to those treated in regular hospital wards yet had similar outcomes. So there could also be downsides for some.

Perhaps more concerning, however, is the use of the ICU and invasive life support for patients too ill to benefit from their application. During their last six months of life, 40.3 percent of elderly American patients with cancer undergo one or more ICU stays, compared to only 8.2 percent of such patients in Germany and 15 percent in Canada. We are also far more likely to die in the ICU, even though it is a place where few people say they want to spend their final hours: while the ICU is the site of death for about one in twenty English patients, it is where nearly one in five of us Americans go to die. As for patients transferred to LTACHs, the outlook is even more grim: as of 2006, about 70 percent of Medicare patients transferred to LTACHs with ongoing need for mechanical ventilation were dead in less than a year.

In other words, sometimes ICU care doesn’t buy time; it imposes burdens on what time we have left. Unfortunately, though, this is often something we cannot predict. Just like the doctors at Blegdam Hospital, we do not have a crystal ball; we do not know who will be able to survive an ordeal of life support for Covid-19 just as they could not for polio. Sometimes, however, the writing is on the wall, and doctors are put in a position of providing invasive care that they believe to be not merely fruitless but unethical.

Substituted Judgment Day

Many health care providers have been demoralized by their work in the ICU due to what they view as the harmful provision of ongoing invasive measures. In 2010, for instance, doctors at a New Jersey hospital came to believe that the critical care they were providing to a seventy-two-year old man named Ruben Betancourt, who had suffered irreversible brain damage after his breathing tube was accidently displaced during a surgery, was inappropriate. As the Wall Street Journal reported, following this terrible event, Mr. Betancourt’s lungs and kidneys failed, and he “developed severe, infected ulcers around his body.” Only mechanical ventilation, dialysis, and artificial feeding kept him alive but in a state of either unconsciousness or—perhaps more frighteningly—partial consciousness. “Mr. Betancourt is dying,” the hospital insisted in a court document quoted by the Journal, “and that dying is being prolonged by the treatment rendered.” Thus far, this is a largely unremarkable clinical situation, with the exception that the case wound up in court after Mr. Betancourt’s physicians moved to pursue palliative measures and stop dialysis against the wishes of his family. Almost always, in the face of an intractable dispute over such decisions, physicians will acquiesce; the patient may be transferred to an LTACH, and that will be that. In this case, however, his physicians felt strongly enough to push back in court. Even though the court required the physicians to reinstate dialysis, their understanding that the patient was terminal was correct as Mr. Betancourt died during the legal wrangling.

But that makes the ICU a gamble that does not always pay off.

Like all critical care doctors, I have seen many very similar cases, albeit without the legal disputes. Such cases can often take a mental toll on health care providers; in my experience, much more on nurses than doctors, as it is the former who take care of patients’ bodies throughout the day, painstakingly bandaging breaking-down skin, sometimes inadvertently causing discomfort with every turn needed for cleaning and care. Of course, the real problem is what such an ordeal means for patients who are often, if not typically, in states of delirium that prevent them from understanding what is happening and why they are suffering. Living, in such a situation, can mean being endlessly and uncomprehendingly prodded and turned, receiving intermittent invasive procedures, and being poked with needles—not to mention being unable to talk, eat, breathe, or meaningfully interact with others.

Needless to say, this is not the norm of critical care—far from it. For one thing, we do manage pain and anxiety with medications, although these measures can sometimes fall short. And typically, when it becomes clear that invasive life support is only prolonging suffering, patients—or, more usually, their families, given that incapacitation is commonplace in such situations—elect to forgo or discontinue invasive life-support measures. Sometimes less invasive measures are used instead, such as antibiotics and fluids, but not mechanical ventilation or chest compressions. Other times, however, a complete reorientation of care toward “comfort measures” is the course of action. Only those treatments that improve comfort, like pain medications, are continued; most or all of those that do not—including measures such as mechanical ventilation and artificial nutrition—are avoided or discontinued altogether.

All this is to say that the system typically “works” with honest and ample communication: patients, or their loved ones, learn about prognosis and treatment options from the providers; reasonable measures are continued; unreasonably invasive or unhelpful therapeutics are eschewed. Indeed, near the end of life, patient (or family) decisions to limit the use of some interventions, like CPR, have long been the norm. But like all systems, this model can fail. Sometimes, physicians do not clearly communicate the prognosis and medical options to patients or families. Other times, there may not be an available surrogate for an incapacitated patient, which can create legal and logistical hurdles when it comes to changing “goals of care,” as they are called. In some cases, conflicts erupt among a patient’s family members who cannot come to a consensus around them. In these scenarios, maximal, invasive forms of care may be continued until agreement is reached, assuming there is not a designated medical decision maker willing to buck the opinions of others. In other instances, I have seen family members overwhelmed by the magnitude of the decision: letting go might make sense (and they may realize it and even say so), but making the decision feels impossible because it threatens a lifetime of guilt over having been the one to say: “Stop the machines.” So, we don’t.

It’s difficult to quantify this problem because there is no straightforward medical definition of “medical futility”—indeed, it is a thorny concept. One 2013 study in JAMA Internal Medicine surveyed critical care physicians at one health system every day for three months and ascertained whether they believed that any of their ICU patients were receiving care they considered “futile,” without providing a definition of the word. The doctors reported that 11 percent of their patients received care that was “futile,” and another 8.6 percent received care that was “probably futile.” The outcomes of those whose care were deemed futile were dismal in this study: 85 percent had died within six months, and virtually all of the survivors were transferred to other facilities (mostly LTACHs) in “severely compromised health states.” One such patient’s condition was described as: “Anoxic brain injury, no meaningful communication, bedridden, requiring tube feeding.” A second: “Persistent vegetative state, requiring mechanical ventilation.” A third: “Bedridden; nonresponsive; requiring mechanical ventilation, hemodialysis, and tube feeding.”

A paper published in 2019 in the medical journal Lancet Respiratory Medicine examined the causes of what the authors call “irresponsible use of intensive care unit resources in the U.S.A.” One of these causes, they argue, was the expansion of ICU resources I outlined earlier. In health care, supply can create its own demand: machines tend to be used, and that includes those in the ICU. In contrast, limited supply can sometimes appropriately lead us to more carefully consider which machines are right and when to use them: one Canadian study, for instance, found that tightness in the supply of ICU beds led to more discussions and shifts in patient’s goals of care prior to an ICU transfer in the setting of clinical deterioration. The authors of the Lancet Respiratory Medicine study also describe the radical post-World War II pivot from a “paternalistic” medical culture, in which doctors would unilaterally withhold interventions—like intubation—they felt were not beneficial for a patient, to what they call “autonomy without context.” Critical medical decisions, they note, have increasingly been pushed to family members who aren’t always given adequate medical context or an understanding of the ramifications of their decisions, much less a clear recommendation as to the best next step. Not only can this lead to confusion; it can impose a scarring, unbearable psychological ordeal on loved ones.

Autonomy is rightly a sacrosanct value in medicine, of course. Sovereignty over our bodies, and authority over what is done—or not done—to them, is not a principle to be lightly violated. But therein lies the Catch-22 of critical care ethics: given the incapacitating effects of critical illness, so many patients do not have the ability to participate in goals-of-care discussions in the ICU. That is to say, decisions must by definition fall to others—whether to family or friends, medical providers, or someone else.

As a result, when a conflict about proper decision-making emerges, as it sometimes does, it is typically not between patients and providers, or between such abstract ethical principles as “autonomy” and “beneficence.” Rather, the conflict is among other human beings operating under emotional, stressful situations with different viewpoints and understandings. It is a conflict among the patient’s family members, or between their family members and physicians, or a combination thereof. Needless to say, of course, family members and friends almost universally want the best for their loved ones—and increasingly, patients designate a “proxy” decision maker who will be responsible for their medical decisions after incapacitation, although this is far from universal. Moreover, such proxies or “surrogates” are instructed to make decisions as they imagine the patient would: what is referred to as “substituted judgement.” Still, invariably, some patients will be in the position of having critical decisions about end-of-life care made for them by estranged family members, potentially even by individuals who may have mistreated them, or with whom they have fundamentally different philosophical perspectives (on religion, for instance)—or, simply, by loved ones for whom the emotional burden of life-and-death decision-making is too much to bear. The fundamental limitations of the “substituted judgment” concept become obvious when one considers how frequently a patient’s closest family members diverge when it comes to end-of-life decision-making, even when presented with the same facts.

Whatever the specific causes, conflicts—whether among family, or between family and providers—can sometimes end with medical providers believing that the care they are providing is inappropriate, harmful, or unethical because the opinions of the surrogates are deemed a form of autonomy that ultimately takes precedence. Consequently, in ICUs and LTACHs across the country, many patients are receiving highly invasive care every day from providers who believe it is wrong—a troubling thought. And that is the dark side of critical care medicine.

Do No Harm

Many physicians have a favorite condition to treat. Mine is acute respiratory distress syndrome, or ARDS. This life-threatening illness results when some insult—often a viral or bacterial infection, but sometimes trauma (the causes are numerous)—provokes overwhelming inflammation in both lungs. Fluid seeps out of the tiny blood vessels of the lung, drowning the tiny air-filled sacs called alveoli, which impedes breathing and causes oxygen levels to plummet. ARDS is what patients with severe Covid-19 have today: it is why their lungs fail.

Polio can also cause the lungs to fail, as in cases like Vivi’s, but for a different reason: it causes weakness of the diaphragm, which we need to pull in each breath, and of the muscles that we use to safely swallow and cough. In both cases, mechanical ventilators are useful because they buy time—time for the respiratory system to heal. That is to say, ventilation does not “treat” ARDS in the typical sense of the term: it doesn’t suppress viral replication, reduce inflammation, or promote healing. In fact, the ventilator is best avoided when possible. Intubation can predispose patients to new lung infections. Because it can be very uncomfortable to have a tube in the throat and a machine breathing for you, patients often require deep sedation or even medications that totally paralyze their muscles, causing immobility which often leads to severe weakness.

Yet even though ARDS hits the lungs like a hurricane, leaving a trail of inflammation and destruction in its wake, for many who survive the acute phase of the disease, the lungs heal: white blood cells clear the cellular debris, and the fluid in the alveoli seeps back into the blood. But that can only happen if a patient lives to see it. If their blood oxygen level falls too low and their heart stops, the possibility of future healing is foreclosed. So, while ventilators can sustain ARDS patients, shepherding them forward until healing can begin, the time that life support buys is not always enough. When the lungs are simply too sick, as they can be in severe cases of Covid-19, the time you buy—whether two days or five weeks—is for naught, because the damage is overwhelming and cannot be reversed. When there is a chance of recovery, however—and of life and pleasure and joy ahead—that is a risk worth taking.

Countless lives have been saved in the ICU through the intensive application of invasive medical technology. That first patient with Covid-19 I intubated survived after a long ordeal of critical care; Vivi, at Blegdam Hospital, lived for many years, even if she still died terribly young. This is why we have ICUs: the invasive application of physiology and procedures for a purpose—the extension of life and joy, which, needless to say, can encompass a wide spectrum of experiences for each of us. But we sometimes employ invasive measures when the likelihood of benefit is scant, and suffering is certain. And so, one hopes at the beginning of the end of the pandemic, we should begin to think more carefully about the role of the ICU in the post-Covid era.

For one thing, we need to acknowledge that endless, market-driven ICU supply is not the answer to the pandemics of the future. We should recognize that there even could be an excess of ICU bed supply in the United States, at least in some places, or that there have been downsides to their endless expansion at the expense of other hospital functions such as primary and mental health care and substance abuse treatment. Admittedly, this is an odd thing to say in the wake of the worst respiratory pandemic in a century—a time where fears of scarcities of ventilators and ICU beds was a bigger concern than excessive provision of critical care. But in many ways, the real problem was less often inadequate aggregate national ICU supplies and more about their distribution. As one analysis found, nearly half of poor communities lack a single ICU bed, compared to almost no high-income ones. And shortages of workers—often due to falling ill with Covid-19 themselves—always posed a far more serious limitation than insufficiencies in the number of ventilators, an issue with implications far beyond the ICU. Still, it is worth thinking about ways to create backup ICU capacity through various architectural innovations, for use during pandemics and natural disasters, that would otherwise lie dormant until the next emergency comes along.

ICU growth in this country has been driven by clinical need and innovation, but also our unique, often profit-oriented health care system.

Needless to say, there are also cultural, ethical, and legal questions about end-of-life decision-making that do not stem entirely from issues of health financing and infrastructure: How do we as a society reduce the provision of highly invasive care that extends life minimally—or not at all—at the expense of great suffering? There is no easy answer here. For decades, some have pinned their hopes on the ascertainment of “advanced directives”—or, legal documents detailing decisions to accept or reject medical treatment according to one’s personal wishes, values, and beliefs—from all of us before we are incapacitated. No doubt, such directives have their uses: individuals with very advanced states of disease may feel ready to say they do not want certain interventions (say, for instance, intubation) under any circumstances. But for most of us, such specific decisions cannot be made so far in advance. Some such interventions might make sense under certain circumstances, so the question is: When should we intervene and, as importantly, when should we stop?

Indeed, despite the fact that I engage frequently in thorny discussions around end-of-life in my work in the ICU, I have no advanced directives. The problem is that the matrix formed by various medical states on the one hand, and every possible medical intervention that could be needed in each such state on the other, is impossibly vast. More controversially, I also believe that we have perhaps overstated the subjectivity of many such questions. The longer I’ve practiced intensive care medicine—and have heard from patients and their families before, during, and after ICU stays—the more I have felt that what almost all of us “want” when it comes to invasive life-sustaining measures is more universal than what some medical ethicists would contend. I think it is fair to say that nearly all of us want medical treatments to extend our lives for as long as possible, until we reach the threshold when the pain and suffering imposed by such treatments exceeds potential future gains in life and the pleasure and joy it would bring. Obviously, the difficult part is knowing when that threshold has been crossed. But the risk is that a system so heavily weighted towards intensive care can make us forget that such a threshold exists at all.

The Covid-19 pandemic has made clear the uses of the ICU as well as our need to overhaul health care infrastructure according to communities’ needs, not their means. Yet the potential for medical care to improve our lives depends on our ability to access it when we need it, but also to be spared from unhelpful interventions when we don’t. In any radical reimagining of our health care system, we must take seriously both the power and the peril of intensive care medicine.