Good for Nothing?

               For over thirty years, we’ve been arguing about when medical care is “futile.” The debate began in the 1980s, when the most common question faced by hospital ethics committee stopped being, “do doctors have to do this procedure, even if patients or their families don’t want it?” to “do doctors have to do this procedure if patients or their families demand it, even if the doctors don’t want to?” One case that attracted a great deal of attention was that of Helga Wanglie, a woman in her late eighties who was in a vegetative state after hip surgery and whose physicians wanted a “DNR” order. Helga’s husband, speaking on her behalf, disagreed. Ultimately, the case ended up in the Minnesota courts and was settled based on a technicality—the hospital sued on the grounds that the husband wasn’t an acceptable surrogate, but the court found that he was. The issue of what qualified as “futile” treatment was largely skirted.

         The bioethics community then struggled to define futility, with Schneiderman, Jecker, and Johnson proposing what appeared to be a quantitative definition: if the treatment in question has failed in the last 100 cases, then it’s futile. There were several problems with this approach. First, it sounds nice and scientific, but nobody was collecting data on what worked in whom. Even if we had information about 100 people, they were rarely identical in what might be crucial ways—is an 85-year-old with advanced dementia, diabetes, and heart disease who has a cardiac arrest equivalent to a 75-year-old with cancer? Secondly, what exactly was meant by a treatment “failing?” Helga Wanglie’s doctors said a ventilator was “failed” treatment because it could not restore the patient to her condition before she was hospitalized—it wasn’t going to allow her to walk out of the hospital and go home. Her husband said a ventilator was “successful” treatment because it sustained Helga’s life. The answer to whether a treatment would be futile turned out to be, “it depends.” It depends on what the treatment was supposed to accomplish. The net effect was that in most of the country, the word “futility” was abandoned in discussing treatment options.

         The exception was Texas, which passed an “Advance Directives Act” in 1999, widely known as the Texas “Futile Care Law.” It spells out a process by which physicians can assert that a treatment is futile and need not be given, as well as a process for adjudicating any disputes between physicians, patients, and families. Ultimately, the physicians have the final say.

         Despite the consensus in the bioethics community and the near consensus among state legislatures that discussing futility was unhelpful, physicians have continued to believe that various treatments should not be provided in certain circumstances because it is morally wrong. To address the persistent discomfort physicians and nurses have felt administering what they regard as “futile” care, the American Thoracic Society recently came out with a position paper called “Responding to Requests for Potentially Inappropriate Treatment inIntensive Care Units.” This policy statement was endorsed by four other professional organizations. Another guideline, this one published by the Society for Critical Care Medicine, addressed what it called “non-beneficial” treatment. The originators of the quantitative futility concept, Schneiderman, Jecker, and Jonsen, continue to believe that “futility” is the best way to describe the class of treatments under consideration.

         So, which is it: Futile treatment? Potentially inappropriate treatment? Non-beneficial treatment? None of the above?

         The issue, it seems to me, is whether the problem is fundamentally a technical one or an ethical one. “Futility” advocates want to define the problem (perhaps surprisingly, since its principal supporters are ethicists) as a technical one. The proponents of “potentially inappropriate care” seem to acknowledge that the issue is ethical (again surprisingly, since its main advocates are physicians). It would be lovely to make the dilemma vanish with a technical solution! If only we could state precisely when CPR or a ventilator or dialysis or chemotherapy cannot work to achieve whatever specific goal it is supposed to accomplish. Then we wouldn’t have to ask whether it’s the right approach. Surely, we can all agree that if something cannot work, like antibiotics for a viral infection or apricot pits for cancer, providing such a “remedy” is inconsistent with professional standards. 

           Unfortunately, such cases of what has been called “physiologic futility” are relatively rare. More common is the situation where an intervention is deemed by medical professionals to be inappropriate because, in their estimation, the burdens outweigh the benefits. This conclusion rests on an understanding of the clinical situation, but also on a moral judgment.

         The desire to convert what is quintessentially an ethical dilemma to a technical one is understandable. It relies in large measure on the recognition that many medical decisions do have a technical and an ethical component, and that often, it is up to the patient (or the patient’s surrogate) to address the ethical piece. But that awareness does not preclude professional ethical standards playing a role as well. The debate about physician assisted suicide, while complicated because of varying legal standards, is to a large extent about whether physicians regard ending the life of a suffering, terminally ill patient as consistent with their ethical norms. The widespread recognition of the responsibility of physicians to care for patients who are criminals or other individuals they may find personally objectionable is based on the ethical code of conduct of the medical profession.

         Recognizing the patient’s values and shaping medical care to be consistent with those values is desirable—when the patient’s ethics and those of the medical profession do not irreconcilably conflict. It’s important to point out that I’m not just talking about the idiosyncratic views of a given physician—the personal ethical values that have been put forward as a justification for failing to provide contraception. I’m talking about an ethical standard of care set by the profession as a whole. The American Thoracic Society statement on “potentially inappropriate care,” while not perfect, is an effort to define what the profession regards as appropriate—acknowledging, through its caveat about “potentially inappropriate” care, that there could occasionally be circumstances where the profession’s norm is not applicable.

Making Dollars and Sense with MediCaring

For roughly 20 years, a proposal to offer a new Medicare benefit for frail older people has gradually been building momentum. Named Medicaring, it's the brainchild of the pioneering geriatrician and palliative care physician, Dr. Joanne Lynn. Now, the Altarum Institute, where Dr. Lynn is the director of the Center for Elder Care and Advanced Illness, has launched "MediCaring Communities." Lynn and colleagues describe this new, revised model and offer a simulation to show how it could play out in practice in an article in the Milbank Quarterly. The bottom line is that the concept not only makes sense, but it also makes dollars and cents.

The basic premise is that we still do a poor job of caring for frail elders, exactly what I repeatedly argue in my blog posts: we prescribe the wrong medications, subject frail elders to too many tests, and provide treatment that makes sense for people with just one medical condition, not for people who are very old, have multiple chronic conditions, and suffer from impairments in their daily activities. As a result, frail elders are often hospitalized, where they develop delirium and falls and endure incontinence and iatrogenesis. If this state of affairs wasn’t bad enough, all this costs a fortune.

In response, the proponents of MediCaring Communities argue that we should instead address frailty by providing “reliable, supportive services and a care plan that reflects the frail person’s situation and priorities.” This is best achieved by integrating supportive services (transportation, social services, and housing) and medical care and by replacing some of the most expensive and burdensome medical services with additional support services. Those who would be eligible for this program would either be over 65 and have at least 2 areas of physical limitation (or dementia) or over 80, regardless of limitations. In addition to incorporating health, social, and supportive services, a community board would monitor and revise the program. And each participant in MediCaring would have a longitudinal care plan that reflects his or her personal goals of care.

The authors simulated how this program would unfold in 4 different communities, reflecting diverse environments: Akron, Ohio, Milwaukie, Oregon, Queens, New York, and Williamsburg, Virginia. Working with leaders in each area, 4 separate plans were developed that varied in the details of their operation but were similar in the overall strategy. What they found was that all 4 communities would generate substantial savings every month for each enrolled beneficiary. These savings would arise from reductions in inpatient hospitalization, decreased use of skilled nursing facilities, and lower use of long term nursing home care. In Queens, for example, the monthly savings per enrollee would total $250 in the first year, but rise to $537 by the third year. Cumulative net savings from the 4 communities was projected to be about $11 million by the end of the second year and $31 million by the end of the third year for the 15,500 eligible patients.

For MediCaring to work in practice, CMS would need to grant waivers to bypass various regulations that govern current Medicare provider organizations. But most important—and most challenging—it would require older people to be identified as frail (an estimated 10 percent of the population over age 65 is frail, with that percentage rising with age) and it would require that frail patients and their families accept the tradeoffs that underlie MediCaring. Is that a realistic possibility in this era where people don’t want to think about declining health and don’t want to limit their options?

The good news is that 42 percent of Medicare beneficiaries are enrolled in hospice at the time of their deaths--and hospice, like the proposed MediCaring program, demands tradeoffs. It's taken 30 years to reach the current level of participation: in 1987, 4 years after the hospice benefit was first introduced, only 7 percent of older patients were enrolled in hospice at the time of their death. But still today, many people only enter hospice in the very last days of life, too late to derive maximal benefit. Nonetheless, Americans gradually came to understand that their needs changed as they entered life's last phase and that hospice provides more of what they need and less of what they don't than conventional Medicare. We need to help older people and their families reach the same conclusions about MediCaring.

A Rose by Any Other Name

While casting about for something to discuss in my blog, I stumbled on a short article that advocates renaming the “death panel” the “good planning panel.” The authors point out that family meetings involving physicians, patients, and their loved ones talking about future medical care are generally well-received. Moreover, this kind of advance care planning prevents depression and anxiety in both patients and their families, and when patients have these conversations, they typically end up undergoing fewer invasive procedures in their final weeks of life, procedures that most patients say they don’t want. Allowing Medicare reimbursement for such meetings would be a very positive step in the direction of improving the care for patients with advanced illness. Whether calling it a “good planning panel” would transform the way people think about these kinds of discussions, in light of the lingering association with the “death panels” born of the right wing media’s imagination, is another matter. Moreover, “panel”  is a poor choice of word, evoking the image of a jury delivering a verdict. But it led me to think about the power of words and the role of euphemisms in medicine.

When the Center for the Advancement of Palliative Care commissioned a market survey a couple of years ago, they learned that most people either had no idea what the term “palliative care” meant or assumed, incorrectly, that it was the same as “hospice,” which they in turn associated with imminent death. (Palliative care is an approach to care for anyone with advanced illness: it neither assumes the patient is close to death nor does it in any way limit treatment, but rather provides treatment focused on improving quality of life; palliative care can be given alongside of life-prolonging medical therapy). When the public were asked if they were interested in having “an additional layer of support” from their health care team, as palliative care was defined, they were uniformly enthusiastic. Similarly, many physicians were reluctant to broach the topic of “palliative care” with their patients because they thought it would be too frightening; they preferred to offer “supportive care.” So is “supportive care” a more useful name because patients understand that term correctly, or is it a misleading euphemism, designed to make patients think it is something that it isn’t? 

And what about the evolution of the “DNR” (do-not-resuscitate) order? Some years back, the phrase “DNAR” (do not attempt resuscitation) was introduced. Since I’m someone who likes to tell things as they are, I favored that substitution. After all, the implication of DNR seemed to be that if only the physician did perform CPR, the patient would be perfectly fine. Usually, the reality is quite different: no matter whether CPR is performed or not, the patient with advanced illness whose heart stops beating will almost certainly die. But more recently still, some physicians have replaced “DNAR” with “AND,” which stands for “Allow Natural Death.” Instead of focusing on whether a particular technological procedure (CPR) will or will not be tried, this formulation seeks to tell patients that what is at stake is having a “natural” experience. Natural, like organic, conjures up something good, unlike, presumably, something that is unnatural or inorganic. “Allow Natural Death” also adds the word “allow” to imply that if you don’t opt for this course, that is if you choose CPR, you will be obstructing or preventing something natural from occurring. Never mind that this is precisely the point—what is “natural” in this instance is to die, and CPR is intended to prevent that most unfortunate reality, just as taking insulin to treat diabetes or having bypass surgery to alleviate the symptoms of heart disease are very unnatural but often extremely desirable medical interventions.

So are these verbal permutations a good thing or they a kind of sleight-of-mouth, designed to deceive and manipulate? What if the original term—DNR or palliative care, for example—evokes such disgust that patients immediately reject it, whereas the new term—AND or supportive care—has far more positive resonance? I used to buy the bioethics argument that truth-telling is one of the cardinal virtues and that it’s a key ingredient of moral medical practice; that failing to tell the patient his diagnosis or his prognosis engenders fear and distrust, not to mention that it is profoundly disrespectful of a person’s autonomy, his individuality, his “right” to know about his own body and his own future. But I’ve been reading some behavioral psychology lately, and I’m not so sure that people make decisions based on calmly and systematically weighing the pros and cons of the various alternatives; they seem by contrast to rely heavily on their intuitions. What this perspective suggests is that there is no truly neutral way to present information, that words are powerful (though sometimes images are even more powerful), and that the best we can do is to avoid deliberately misleading patients. 

So both “death panels” and “good planning panels” are out because they are not panels and they are not about death; “advance care planning discussions” are more accurate. “DNR” and “AND” are out because they mislead; DNAR is more objectively correct, though it may well have positive associations for some patients and negative associations for others. And I’ll stick with calling what I do providing “palliative care” rather than “supportive care,” though I’m quite willing to define palliative care—if I’m asked—as providing support to patients and families through symptom management, psychosocial support, and advance care planning.

No Sense, Lots of Dollars

Twenty-five years ago, discussions of medical futility were the rage in bioethics circles. The discussions petered out when it became clear that futility was in the eye of the beholder: physicians and patients often had very different ideas about what futility meant, depending on what they hoped medical treatment would accomplish.
   
In one case that generated considerable publicity, physicians sought to turn off the ventilator that was keeping 86-year-old Helga Wanglie alive. They argued that the ventilator was futile treatment since it would never allow Mrs. Wanglie, who was in a persistent vegetative state, to regain consciousness. Mrs. Wanglie’s husband, however, argued that keeping his wife alive—supplying the oxygen that her heart needed to keep on beating—was the goal of treatment. And by that standard, the ventilator was performing admirably. The court to which the physicians presented their case did not address whether the treatment was futile; it merely ruled that Mr. Wanglie was the rightful spokesperson for his wife and his wishes should be followed.

A second problem with futility is that it is a good deal easier to identify after the fact—the patient died, ergo the treatment didn’t work—rather than in advance. Because futility was proving elusive, medical ethicists stopped talking so much about it and focused instead on ascertaining the patient’s goals of care. The prevailing wisdom came to be that doctors should provide any treatment that was consistent with those goals. Ethics consultations were used to mediate disputes between families and physicians over whether particular treatments could achieve the desired goals. But physicians continued to be bothered by the nagging feeling that at least some of the treatments they provided were morally wrong: they caused needless suffering as well as outrageous costs without much, if any, benefit. A new study just out puts the futility debate back on the table.

The authors of the study used a focus group of 13 doctors who work in intensive care units, the site of 20% of all deaths in America, to agree on a definition of futility. They came up with four reasons for assessing a treatment as futile. The patient was imminently dying, the patient would not be able to survive outside an ICU, the burdens of treatment greatly exceeded the benefits, or the treatment could not possibly achieve the patient’s explicit goals. They then asked physicians at a large medical center in Los Angeles to evaluate each of their ICU patients every day and indicate whether the care they were providing was futile, using these four criteria. In one fell swoop, the authors got rid of the two problems with previous futility studies—sort of. They used a prospective design, asking for evaluations in real time, not after the fact. And they defined futile care, albeit by unilateral decree.

Over a 3-month period, the investigators collected data on 1125 patients cared for in one of 5 different ICUs by a total of 36 critical care doctors. They found that 123 patients (11%) were perceived by their physicians to be getting futile treatment at some point during their ICU stay. Another 98 patients (8.6%) got “probably futile treatment.”

What characterized the 123 patients whose doctors were convinced they were getting futile care? Their median age was 67 and 42% were on Medicare. They were more likely to be older and sicker than the rest of the group. The majority (68%) died before hospital discharge; another 16% died within 6 months; almost all the remainder were transferred to a long-term care facility, dependent on chronic life support. The total cost of futile hospital care for these 123 patients was $2.6 million.

In light of these results, it may be time for critical care specialists to convene a consensus conference to see if they can agree on the criteria for futility. Agreement by the majority of doctors who care for ICU patients would carry far more weight than the focus group comprised of 13 physicians whose opinions formed the basis of the current study. If a majority of the nation’s critical care experts came up with criteria for futility, whether the same ones used in this study or some modification, then Medicare would be in a good position to decide to pay only for clinical care that met the newly defined standard of care. 

Medicare would not be dictating what is appropriate care; it would not be interfering in the practice of medicine. Medicare would merely be restricting payment to services of established benefit, just as it does when it pays for a cardiac pacemaker or an implantable defibrillator only if patients meet standard clinical criteria. Patients could still opt for treatment their doctors deemed futile if they were willing to pay for it. At an average cost of $4004/day for ICU care, I wonder how many people would pursue this route.