The 1960s and early 1970s saw the advent of two struggles within medicine that exist to this day: an all-out battle against death and disease and a countervailing search for ways to limit the collateral damage to the dying of that battle. They are struggles that work against each other, reflecting a deep conflict about the goals of medicine. The struggles are hardly confined to the United States, but they have taken on a distinctive life of their own because of some long-standing American values and the many pathologies of American health care.

The opening salvo of the battle was the much celebrated arrival during that era of a wide range of new medical technologies that greatly expanded the power of medicine to keep the human body alive: intensive care units (ICUs), organ transplantation, kidney dialysis, ventilators, improved tubing to provide nutrition and hydration to critically ill patients and a widening array of antibiotics to deal with infections. Along with those exciting developments was an escalation of the struggle against death, perfectly captured in President Richard Nixon's 1970 call for a ‘national crusade’ to conquer cancer, often characterized as a ‘war’ against it. No less important was a rapid expansion of the research budget of the National Institutes of Health (NIH). Optimism was in the air. Infectious disease, many declared, had been vanquished and medical research was confidently poised to conquer the chronic diseases of aging societies, fully expecting victory.

The parallel struggle was to improve the care of the dying. One of the casualties of the war against death was that medicine forgot its ancient history of care, its acceptance of death as an inevitable part of life, and the ethical duty of the physician to do everything possible to bring about a peaceful death. Complaints about the care of the dying mounted rapidly in the late 1960s and 1970s, and they can still be heard frequently some 50 years later. Doctors evade talking with patients about their prognosis, patients want to be given hope however poor the odds, and steady medical progress and an enhanced technological possibility of extending the lives of even the sickest patient has in many ways made decision making harder, not easier; there are just more things that can be done to forestall death. What came to be called the ‘technological imperative’ was widely understood by doctors to mean an obligation to do everything possible to continue the life of the dying and to aggressively use the available technology to do so. To that, technological imperative was added a less known ‘research imperative’: a supposed duty to pursue unlimited medical progress, but with no specified final outcome in mind. It aims to extirpate whatever diseases kill us (Callahan, Reference Callahan2003). Fatalism about death was rejected and medical triumphalism was put in its place.

The antidote to the collateral damage was initially taken to be three reforms: greater choice by patients at the end of life (by means of living wills and/or the appointment of a surrogate), by better physician training in doctor–patient communication and by the development of the hospice movement and improved palliative care. Those reforms have been vigorously pursued over the years and they have been a modest success. Some 25% of the American population has an advance directive or has appointed a surrogate, well over one million people die each year in the care of hospice (about 38% of all deaths), palliative care is now a growing medical specialty, and most medical schools have worked to improve physician training in care of the dying.

But I call all of that a ‘modest success’. One does not have to scratch far below the surface to hear stories from families, or in the media, or among doctors, that sound painfully like those we first began hearing about 50 years ago. It is good that 25% of people have an advance directive but that means 75% do not. For every story about wonderful doctors caring for the dying, there is another about physician insensitivity, overtreatment and a prolonged dying. The development of the hospice movement (inspired by the United Kingdom) is wonderful, but the fact that a large portion of those receiving such care receive it only a short time before they die is not. As for the triumphalist talk of the coming cure of chronic disease, well, the best that can be said is that we are still waiting. Some 20% of patients (heavily the old) account for an estimated 64% of our health care costs and with a large percentage of that cost incurred in the last year or so of life (Orzag and Emanuel, Reference Orzag and Emanuel2010). Not one of the chronic diseases has been cured and not one is anywhere near a cure. Meanwhile, the celebrated conquest of infectious disease four decades ago turned out to be premature, thanks to human immunodeficiency virus disease, antibiotic resistance and hospital infections. Nature is fighting back and is more than holding its own.

To be sure, I am painting a broad picture here, making large-scale generalizations. As always with such efforts, counter-examples can be found: physicians who talk well with their patients and offer accurate prognoses, who decrease aggressive interventions with their oldest patients (often without asking) and patients who have made clear in advance the kind of care they want at the end of life, and get it, and families who do not try to intervene and overrule a loved one's expressed desires. I have no way of knowing what the actual balance is between good and bad care, realized or thwarted desires, careful and careless use of high-technology interventions. All I can say is that, after many decades of following this issue, the number of news reports and research papers is, if anything, more common now and the reports of bad deaths even more pronounced than ever. Whether that phenomenon is a function of reality or simply more public interest and consequent media attention is unclear.

For a long time it was thought that the law stood in opposition to end-of-life care, tying the hands of physicians, and it was often possible to read state laws to that effect. But a number of crucial court decisions managed to get most of the early legal obstructions, if not all anxieties, out of the way (Jonsen, Reference Jonsen1998). The first case to attract attention was that of Karen Ann Quinlan, brought unconscious to the emergency room of a New Jersey hospital in 1975. She was in a deep coma, induced by a combination of barbiturates, alcohol and Valium and, some weeks later, diagnosed as being in a persistent vegetative state and put on a ventilator. Her family requested that she be taken off it and allowed to die. Her physician refused to do so and the case eventually was taken up by the New Jersey Supreme Court, which decided in 1976 that she could be taken off the ventilator. The Joseph Saikewicz case in 1976 in Massachusetts decided that, in the case of a severely retarded and institutionalized man who developed leukemia, that had he been competent, he would have refused therapy. Thereby the court legitimated the principle of ‘substituted judgment’, allowing the authorities at his institution to decide for him. The judges in a 1986 California case decided that Elizabeth Bouvia, a paralyzed patient being tube fed, did have the right to have the feeding stopped and to be allowed to starve to death. A much publicized and politicized case, running from 1995 to 2005 and moving through a variety of courts, pitted the husband of Terri Schiavo, in a permanent vegetative state, and who wanted tube feeding stopped, against her parents, who wanted it continued. A court finally agreed with her husband and the tube was removed.

There were other court cases and their net impact has been to remove any serious legal obstacles to end-of-life care, for competent, as well as incompetent patients. There has, however, been a continued worry on the part of many physicians that, unless they are very careful, they may be sued for their treatment of dying patients. But lawsuits pertinent to end-of-life care have been rare and, I believe, no physician has been successfully sued because of his end-of-life care. But the persistence of anxiety about that possibility probably plays some role in overtreatment. A careful examination of various legal beliefs about end-of-life care determined that most of them are myths, but because of variations in state laws “sometimes there is a grain of truth” (Meisl et al., Reference Meisl, Snyder and Quill2000). For many physicians, myths or not, it seems to be a worry that does not go away.

If, then, the law does not stand in the way of appropriate care, in fact easing the way, why may there be some substance to my contention that end-of-life care is becoming harder, not easier? It is important to distinguish between the various social efforts to change patient attitudes, on the one hand, and technological innovations in the medical care of the patient, on the other.

One possible reason for the success of the social efforts is that the educational progress to date has worked reasonably well with many patients because they are the ones typically more educated and medically sophisticated than the average person. It is now routine for lawyers who help clients prepare wills to suggest that they should also appoint a surrogate for end-of-life decisions. But most Americans will not get such help, sometimes because they cannot afford it, but sometimes because they want to evade thinking about death. It is also well known that minorities are less likely to have surrogates and, for that matter, to want more rather than less care treatment when they are dying. There is some good evidence that most people do not act on the knowledge they already have. A 2012 study of 1700 adults by the California HealthCare Foundation found that 80% believe it is important to have their end-of-life wishes in writing, but less than 25% have actually done so. Why was that? The main reason most gave for not even talking about the subject was that they had ‘too many other things to worry about now’, and 25% said that they did not want to talk about death and dying at all (California Health Care Foundation, 2012). There were also some interesting anomalies in the study results, and here is a striking one: ‘Almost 80% say they definitely or probably would like to talk with a doctor about end-of-life wishes, but only 7% have had a doctor speak with them about it’. Of late, death is more openly talked about in the media and stories about both good and bad deaths have almost become a staple topic; even so, it is still a subject of fear and uncertainty for many, probably most, people.

And no less so for physicians. American physicians are assiduously trained to embrace medical technology in the care of patients, and a heavy majority of medical students go into the sub-specialties, where the use of technology is most predominant. Medical students in the United States typically graduate with a heavy debt, pushing them toward the higher-paying specialties (Bodenheimer and Berenson, Reference Bodenheimer and Berenson2007). The proportion of those going into primary care is down around the 20% level and declining. International evidence has long shown that medical costs are lower when primary care is at least equal to specialty care. And then, once out in medical practice, physicians are for the most part paid on a fee-for-service basis, and paid much better for the use of technology than for talking with patients. The hospitals they make use of for their patients often engage in competition, aiming to have the latest and best technology, and with every incentive to use it to make up for its start up costs and to use it to maintain a good revenue stream. The evidence base for many common procedures carried out by physicians, in or out of hospitals, and particularly for treatment of patients with multi-organ failure, is notoriously weak. An uncommonly interesting study of physician behavior with elderly dying patients by medical anthropologists – “Risk, Life Extension and the Pursuit of Medical Possibility” – found that many such patients are coaxed and seduced into risky behavior in the name of life extension (Shim et al., Reference Shim, Ross and Kaufman2005).

To further complicate the story, American patients are increasingly demanding, or at least most physicians tell me they are – even if they also tell me that more patients want to die a ‘natural’ death. Stories about family members interfering at the last moment in fulfilling patient wishes are common in medical circles, but then so are stories about loving families gathered at the bedside of someone dying a peaceful death. Yet families and patients too often expect cures, and complain when they are not forthcoming. That kind of faith should not surprise us. Americans are constantly bombarded by direct-to-consumer advertising in the media, touting the benefits of various pharmaceuticals old and new – pharmaceutical companies now advertise expensive cancer drugs with minimal survival value in leading newspapers. The public is no less deluged by media stories on alleged coming miracles, new medical breakthroughs and imminent cures and all possible ways to improve their health. Although he is Canadian, Dr Allan S. Detsky (Reference Detsky2011) reports in a provocative study findings that reflect the attitudes of both Canadians and Americans: “Even if patients are in a health state for which cure is exceedingly unlikely, they want to have hope and be offered options that might help…although many patients prefer not to ‘know’ or ‘try,’ the majority of those who seek health care prefer active strategies. An extra test or two, ‘just to be sure’, is often preferred to possibly missing something”. Ironically, many doctors do not want to die the way their patients do, often conveying to the latter hopes that they know to be unlikely and would not entertain in their own case (Murray, Reference Murray2012). What they think about privately does not carry over in their willingness to be candid with patients. Telling patients that they are dying is, as one nice piece of reporting put, “A Conversation Many Doctors Won't Have” (Spam, Reference Spam2011).

Some comparative international studies have shown that Americans have a much greater interest in new technologies than citizens of European countries do. One survey found that some 34% of Americans (with comparable Canadian figures) believe that medicine can cure any illness they have, a figure far above that reported elsewhere (Kim et al., Reference Kim, Blendon and Benson2001). Some 86% of Americans believe that they should have access to the most advanced medical tests and procedures, and 66% say they are ‘very interested’ in new medical discoveries, well ahead of most (even though not all) European countries.

I earlier used the phrase “research imperative”. One of the great American government success stories is that of the rise and flourishing of our NIH, with a budget now of over $30 billion a year. The NIH has flourished because it has in effect promised the public that, with more and more funding, it will find cures to the major diseases that kill us. If the ancient ethic of medicine required physicians to inspire hope in even the most critically ill patients, the modern research enterprise has no less found it beneficial to inspire hope. That kind of hope is not only for the cure of disease but also that more research will eventually lower the cost of health care.

What have we the public received from the research investment? There have been few cures, but much success in keeping sick people alive for longer periods of time, lengthening both life expectancies and the process of dying. That kind of progress can be called a benefit, though of a two-edged kind: worse deaths and higher costs for them. As for controlling costs, most economic evidence shows historically and consistently that research expenditures and health care costs rise in tandem (Cutler, Reference Cutler2004). And it is reasonable to see not just a correlation but also causation when the research serves mainly to expensively extend life.

Two features of this form of progress are particularly worth noting: the difficulty of finding any bright line between living and dying, and the gradual demise of the earlier notion of the inevitability of death. The steady infusion of new and improved technologies into medical care means that there are more and more ways to keep dying patients from dying, usually only in small increments. There is always something more that can be done to add a few more hours, or days, or even weeks to the life of a dying patient. It gets harder and harder to say what is truly futile care. One provocative study of medical residents found that the physicians considered a patient to be dying when nothing more technologically could be done to keep him alive (Muller and Koenig, Reference Muller and Koenig1998). Death, in other words, was seen as a function of technological capacity, not something happening to the patient's body. It is hardly any wonder that many physicians feel that a patient's death is somehow a failure on their part; they just might have done better. A common argument for euthanasia is that most patients now die because of a physician's decision to discontinue life-sustaining treatment. And if that is the case there is no logical difference, some argue, between killing a patient by letting him die and directly killing him with an injection. I reject the argument but it is out there.

Although it may seem extreme, I believe that the fundamental ambivalence toward thinking about the place of death in human life and in medicine is at the core of many of the most vexing features of end-of-life care. Perhaps this is unique to the United States. That I cannot judge, but some 100 trips to Europe over the past 30 years has persuaded me that our ambivalence is at least more pronounced here than there. Between the 1940s and 1960s there were a number of books written by British authors satirizing American attitudes toward death. Aldus Huxley reportedly said that Americans reject the idea of death, and Evelyn Waugh, in The Loved One: An Anglo-American Tragedy (1948) and Jessica Mitford, in The American Way of Death (1963) had good sport with the American funeral and cemetery industry.

There are also some present peculiarities concerning explanations about why we die. Taken as a whole, those explanations implicitly make death avoidable and could someday be put out of business. Why do people die? Here are some common explanations. People die because of an unhealthy lifestyle, smoking, lazing about and letting themselves get fat. They die because of their bad luck in having parents who passed along to them lethal genes, or because they grew up in poverty or immersed in unhealthy environments. Or, if they die in old age, it is not their age that kills them but just the (eventually) curable chronic diseases of aging that do them in. Those who die of cancer in their younger adult years have the misfortune of dying before a (surely certain) cure is found: if only they had their cancer 20 years from now! And of course deaths from wars, household accidents, murders and auto fatalities are all known to be preventable. And is it not well known that good preventive medicine and public health measures can save lives? That is good news, though it is rarely specified for how long.

I have tried so far to show how much of American medicine and health care works to create an environment almost designed to make end-of-life care difficult, and at least not medically getting any easier. But it has not been designed to have that kind of impact. It is more the case that a variety of variables work together to make it all hard. They build on a foundation of an American belief in progress (going back to Thomas Jefferson), in the value commercial and otherwise of technological innovation, of economic rewards for those who work hard and who believe that doing good is not in conflict with doing well (including well-rewarded physicians), and in a full acceptance of a highly commercialized health care system, bringing well-deserved profit to medical entrepreneurs, pharmaceutical and device manufacturers, for-profit hospitals and insurance companies. It is now one of the few major economic sectors steadily offering new job openings in the midst of a major recession. Is that really good news in the face of an urgent need to slow the growth of health care costs?

The main strategies to deal with end of life issues, noted at the beginning of this article, are essentially managerial and educational: educate people about their rights and the need to appoint a surrogate, enhance the education of physicians and improve hospice services and palliative care. Efforts to understand the background culture of American medicine, as I have emphasized, have been less common. But there have been some new additions of importance on the managerial side. In recent years, efforts to coordinate care of the chronically ill and to bundle payments for their care have been added to that mix. The cost of care for the dying has been a recurrent theme in the reform efforts. The US Congress recently passed an ambitious health reform plan. In intent and on paper, it will insure some 32 million of the uninsured, rectify many failings of private insurance, dramatically increase comparative effectiveness research and put in place a number of efforts to control the rise of health care costs. But if it is far from universal health care in its added access to care, it is also at risk of dismemberment and dilution at the hands of a future conservative Congress. The potency of its cost-control efforts are still heavily hopeful rather than assured. A side bill, recently introduced in Congress, will increase reimbursements to physicians for time talking with patients about end-of-life care. Because of the excessive financial rewards for providing often excessive end-of-life care, the control of costs will be pivotal and, within that category, the control of technology costs. Poor end-of-life care is symptomatic of the larger failings of the American health care system. It is a sick system treating sick patients, an altogether bad combination.

But for me the deepest problem, and with it a challenge, is that of the way our American medical culture embodies a set of once useful and compelling but now distorted values, those of progress, suffering and death. I will call their combination the infinity syndrome. It exists in all developed countries but the American experience is notable for pushing it about as far as it can go, and we are now suffering the consequences of having done so.

At the heart of that syndrome is the conviction that suffering of any kind and to any degree (both mental and physical) is an evil, that death is the greatest evil of all, and that nothing should be allowed to obstruct or slow the ethical imperative of unlimited medical progress to hinder the war against them. That syndrome is the Enlightenment in action, but intensified by two key American core ingredients in its practice of medicine. One of them is the embrace of market values as a cherished part of its health care system (expand individual choice and commercial competition in the provision of care; Callahan and Wasunna, Reference Callahan and Wasunna2006). The other ingredient is, as a matter of principle, to set no boundaries or limits to the pursuit of medical progress and technological innovation (science as the ‘endless frontier’). Many Americans are resistant to universal health because they realize that it will entail some national limits on the provision of care (e.g. rationing), and – as the medical–industrial community well understands and lets the public know – it will no less dampen technological innovation and the medicalization of any and all human complaints, one of the fruits of innovation. It has long been known that the royal market road is one that turns desires into needs.

End-of-life care does not easily find a place for itself in this syndrome. How can it? It does not logically belong there. It requires a willingness to accept death as part of the human condition, to understand that progress has its limits. That dual requirement entails physicians who understand that accompanying patients on the way to death becomes as important as their earlier duty to lead them along the path to health. It requires patients and their families who understand that medicine cannot, in the end, save us from death, and thus not to expect medical miracles. And it will require a medical industry that understands that new and updated technologies that do no more than extend life for a few months (or maybe even for a year or so) at a high cost offer no real benefit. They not only can add to unsupportable health care costs but also serve to add a reason – a tantalizing seduction in fact – to put off good end-of-life care: do not give up just yet, fight the good fight.

Let me return to the hospice statistics to point out some possible routes to what I concede are difficult prescriptions. The average length of stay in hospice care is 69 days, with a median length of 21.1 days, meaning half received care for less than three weeks, and the other half more than that. As the majority of patients in hospice have cancer, it is highly unlikely that it is not known they were dying until three weeks before they did or, for that matter, before the 69-day average. Why not? The most likely reasons are that either the patient, the family or the physician in charge of the treatment plan (or all three together) could see the future well enough but chose to evade or ignore it. The aggressive treatment is thus continued for too long and the patient receives essentially useless treatment. A study of patients dying of heart disease found that over time a greater use of expensive intensive care occurred at the same time that hospice services were also rising (Unroe et al., Reference Unroe2010). Although there is good evidence that hospice care can reduce costs, it is the costs incurred before hospice that have to be counted as well. Improbable hope and self-protective denial are just a millimeter apart and the patient is often the loser.

How can that pattern be changed? One promising new development has been research on developing better disease prognosis and urging physicians to talk more frankly and earlier with those patients who are not yet dying but whose prognosis is poor. I would contend that it is an obligation of the physician to tell the truth even if it seems that sometimes patients do not want to hear it. This is obviously difficult for many physicians, and various efforts are underway to change that reluctance (Smith et al., Reference Smith2011; Workman, Reference Workman2011; Yourman et al., Reference Yourman2012).

But buried in the premise of my line of argument is a conviction that, if a risk is to be run, it should be on the side of stopping life-sustaining treatment too early rather than too late, and surely with the elderly. That policy would be particularly sensible when the difference in life expectancy was not great, but it could be plausible even if the difference was longer.

Although it is easy to find horror stories about cancer victims who could have been diagnosed earlier and their life saved, but for one reason or another were not, it has been just about unknown in my experience to hear another kind of story: that of a stage 3 or 4 cancer patient for whom plausibly beneficial treatment of a significant kind was deliberately declined by the patient or noted and communicated but not recommended by the physician – and yet was not regretted later by anyone. Far more regret is the case with patients treated too long and too aggressively for cancer and other lethal conditions, something that eventually became clear when the patient died a bad or miserable death.

Why was this allowed to occur? A tellingly powerful story of the death of her father after 10 days in the Stanford University Hospital, and $323,658 in costs, by a California health writer, Lisa Krieger. Her account brought out tellingly a recurrent point in many such stories. Her father was 88 years and suffering from advanced dementia, had ‘carefully prepared end-of-life instructions’, and was admitted to the hospital's ICU for septicemia. He went downhill thereafter, and the ‘medical nightmare started, as they so often do, incrementally’. Again and again, as his illness worsened, the ICU team came to the rescue. At what point, she asked herself, “should we have quit”? Again and again, however, the physicians held out hope, and one intervention after another was initiated. “Who was I”, Ms Krieger asked, “to summon his death”? So, she hesitated, and wondered, after still another crisis and rescue, “Now Was It Time To Stop?” Finally, ‘one last powerful tool’ was offered, immunoglobulin, a new class of drugs known as ‘biologics’, at a cost of $48,000. It did not work, and she finally asked the attending physician “Please, tell me what's ahead of us”? His response: “It's not black or white; its gray… A long and bumpy recovery with no guarantee of success.” And “now, finally, it was time to stop,” she decided. And the physicians did stop. Her father's case is one I would call ‘death by medicine’.

I began to think this way a few years ago when a friend of mine, a well-known teacher of ethics in a medical school, and close to death from recurrent pancreatic cancer, had trouble talking with me as he was clearly at the edge of death. He could not talk well, he said, because he had agreed to test another round of chemotherapy, and his mouth was full of sores. When I asked why he let that happen – particularly as he had taught courses on the care of the dying – he said, sheepishly, “they talked me into it”. It may well have prolonged his life, but that was a source of misery not gratitude.

All it will take for most people to let go too early rather than too late will be for them to realize that the so-called progress represented by possible ways of keeping them alive have a better than even chance of bringing them misery – as is almost always the case with a treatable but rapidly progressive metastasizing cancer. Of course it will also depend upon physicians who are inclined to see a patient die too early rather than too late, matched by families who feel the same way.

I can do no better here than to end with a quotation from a British physician Iona Heath (Reference Heath2010): “all clinicians caring for older people have the experience of treating one disease process only for another to take its place…when one cause of death is curtailed, others must inevitably come forward to fill the gap. Everyone is obliged to die from something”.

Of course to believe that it might be better to let someone die too early from a lethal disease rather than too late and – especially with elderly patients dying of multi-organ failure – to allow a patient to die from one of the organ failures rather than waiting for another to kill them requires for an American a certain willing suspension of a number of beliefs. There is the belief that all medical progress and research are good and, more than that, morally obligatory; that death is and always ought to be the chief enemy of medicine; that the whole point of medical technology is to use it; and that it is better for physicians to do all they can to preserve life when still more could be done, save for the last two minutes, even though all realistic hope had been given up days or weeks ago; and that in the tension between peacefully accepting death and courageously fighting to the last breath, it is best to admire the latter. Such a constellation of beliefs is slowly being eroded, chipped away at by education, sad stories, different attitudes among physicians and, maybe, a recognition that the war against death is ultimately unwinnable, and surrender to its inevitability is the course of wisdom. The US NIH, and the researchers it supports, do not appear ready to end the war, but that day may some day come, at least shifting the focus to achieve a better balance between care and cure. I would call that genuine progress. Just a little shift would help.