Planning Ahead

Maybe nobody paid much attention to National Healthcare Decisions Day because it fell on the first day of Passover or because it came right after the anniversary of the Boston Marathon bombing. Or maybe Americans are suffering from holiday overload. But April 16 was officially the day Americans were supposed to remember to complete an advance directive—a document that states who is supposed to make medical decisions in the event of our incapacity or what kinds of choices we want made—and the major media didn’t breathe a word about it. Are advance directives passé? Are they still toxic after Sarah Palin associated any attempt to plan for future medical care with a “death panel?” Or are people blasé about them—either because they already have one or because they don’t do any good?

A new study published this month in the Journal of the American Geriatrics Society looks at one aspect of the question: how many people have an advance directive and how has this changed over the past 10 years? Drawing on data from the Health and Retirement Study, a national sample of people over age 60, and using interviews conducted with next-of-kin after death, the authors conclude that the proportion of people who died with an advance directive in place increased markedly: from 47% in 2000 to 72% in 2010. And these numbers are dramatically higher than the earliest reports. Studies done just after the Patient Self-Determination Act first went into effect in 1991—a law that requires all health care institutions that receive federal money to ask patients if they have an advance directive and to offer them one if they don’t—found completion rates for the population as a whole barely made it to the double digits. But does this improvement matter?

The new study has some major limitations—it only includes people who died and it relies on the testimony of a relative or friend to determine if the person had drawn up a living will or designated a health care proxy. Moreover, it was confined to people over age 60. A Pew Research poll done a few years ago found that only 29% of the general public had a living will (up from 12% in 1990). But more important than how many people have advance directives is what kind of advance directive they have and whether these documents change what happens to people when they get sick and cannot speak for themselves about the kind of medical treatment they want.

As far as the kind of document goes, we know from the current study that 45% of people who died had a living will, some sort of statement about the type of treatment they would want. These range from very general assertions such as “no heroic treatment” to specific statements about ventilators or attempted cardiopulmonary resuscitation, but typically apply only if “death is imminent.” A total of 57% had a durable power of attorney for health care, or health care proxy—they had designated someone, usually an adult child or spouse, to make decisions on their behalf if necessary. And 38% of those older individuals who died had had both a living will and a health care proxy. We don’t know anything about what fraction of people had done any other kind of advance care planning: we don’t know, for example, if they completed a POLST form (a physician order for life-sustaining treatment, which is a medical order indicating which of various specific possible treatments such as CPR, a ventilator, or dialysis, a prospective patient would want. We don’t know if they indicated their goal of treatment, a decision that is supposed to precede completion of a POLST form, and which gives physicians guidelines to use in deciding about any given intervention: is the technology that is under consideration likely to be effective in prolonging life? In keeping the person independent? In promoting comfort? We don’t know if the 6005 people who form the basis of the new study had ever spoken to their families or their physicians about what mattered to them or what approach to treatment they would want if they became very ill.

We do know that whether or not someone had an advance directive failed to affect their chance of being hospitalized in the last 2 years of life and had no effect on whether they died in the hospital. That is, we know that the risk of hospitalization in the last 2 years of life went up considerably between 2000 and 2010 (from 47% to 72%), but having an advance directive did not affect the rate one way or another. We also know that the odds of dying in the hospital as opposed to a nursing home or at home fell (from 45% to 35%), but that fall was also unrelated to advance directive status.

So do advance directives achieve anything? The authors of the current study showed in an earlier study that having an advance directive appears to be associated with a greater likelihood that one’s wishes for end-of-life care are followed. This conclusion was based on the reports of the next-of-kin after the death of their relatives and it was never clear on what basis those next-of-kin knew what kind of treatment the person who died had actually wanted. There is very little evidence that advance directives trigger or reflect conversations about the goals of care. There is even less evidence that they influence care before the last few days or weeks of life. Maybe the national media ignored National Healthcare Decisions Day because it was overshadowed by other more exciting events. Or maybe they ignored it because it continues to focus on a form, on completing a legal document, rather than on addressing head on what is most important to us in the last year or two or five of life, on the decisions that will matter most in life’s last stage, however long that lasts.

All or None

My father turned 90 today. He has Parkinson’s disease and dementia, both of which involve his brain, but the rest of his organs are going strong. He lives in a nursing home, where much of his time is occupied by basic activities—washing up, dressing, shaving, eating, and going to the bathroom—all of which he needs help with. My father doesn’t say very much and spends most of the rest of his time staring at the NY Times, which he can no longer understand but which my mother brings him faithfully, and sitting in the dining room doing nothing. My mother visits with him and walks with him almost every day. Very occasionally, he agrees to attend one of the many nursing home activities such as a current events discussion group or a concert. He is in no position to make decisions about anything important, certainly not about what approach to medical care makes most sense for him.

It’s hard to know what kind of medical care my father would choose for himself if he could miraculously emerge from his state of debility and dementia and see himself as my mother and I see him. I think he would be appalled: he was always a very controlling, capable, resourceful person and now he is none of those. I asked my mother whether she thought he would want life-prolonging treatment for himself—again, assuming he could step outside himself and see his current condition—and she wasn’t sure. When he was physically well but suffered from panic attacks, he used to express anxiety about dying. My mother thought perhaps that implied he wanted to keep going, no matter what the circumstances. I didn’t think the anxiety he experienced years ago was relevant to the present situation. I thought that if my father had known then that one day he would have dementia and need so much personal care that he had to live in a nursing home, he would have been horrified. I believed he would only want medical treatment that kept him comfortable, treatment for pain or shortness of breath or nausea or whatever symptoms he might develop. My mother and I agreed that since we couldn’t be sure what my father would want, we would decide on his behalf (my mother is his health care proxy and I am the alternate).

We decided that the focus of my father’s care should be his quality of life. The goal of medical treatment should be to alleviate discomfort and to preserve whatever limited capacities my father has. So in addition to controlling a cough or a fever, treatment should be designed to preserve vision and hearing. We agreed that we would not want to subject my father to burdensome treatment, the sole purpose of which was to prolong life. In particular, we would not authorize attempted cardiopulmonary resuscitation. In general, we did not think he should be hospitalized; rather, he should get whatever simple treatment is available in the nursing home.

Over the years as a practicing geriatric and palliative care doctor, I have spoken with countless families about their wishes for relatives who are in a position similar to my father’s. Most of them share my perspective that as dementia progresses, so too do the goals of medical care, with life-prolongation playing less and less of a role and maximizing comfort becoming more and more important over time. Some families disagree, usually either because they feel so guilty about limiting medical treatment in any way, anxious that they would be depriving the person they love of something important, or on religious grounds, believing that life-prolongation is the only acceptable goal of medical care. But I realized that my perspective—and that of my patient’s families, when they understand the relevant issues—is not shared by most health policy makers.

In this month’s Health Affairs, a leading health policy journal that devoted the entire April issue to articles about dementia, the otherwise perspicacious overview article, “The Long Reach of Alzheimer’s Disease: Patients, Practice, and Policy,” takes as a given that until the very end, patients with dementia should receive conventional medical treatment, and only when they have very advanced disease, with a life expectancy of 6 months or perhaps a year, should a more palliative approach be considered. The challenge, according to this perspective, is how to be sure that patients with dementia get state of the art care, given that they often cannot express what is bothering them (making diagnosis difficult) and that they often resist when doctors try to perform procedures (making treatment difficult).

It seems so obvious to me that if patients (and their caregivers) understood how frightening a hospital often is for someone with cognitive impairment, how anxiety-producing a CAT scan can be, and how often conventional medical treatment causes worsening confusion and other complications in someone with dementia, they wouldn’t want usual care. That doesn’t mean they wouldn’t want any treatment. But surely taking a pill at home would be far more appealing than getting intravenous treatment in a hospital, even if the risk of death with oral treatment were somewhat greater than with hospital care (and it’s far from clear that usual care is really the most effective care in people with multiple medical problems, one of which is dementia, even if the goal is life-prolongation), once people realize what the two approaches would be like for someone who has dementia. Alzheimer’s disease is not like heart disease or cancer, where many people are willing to put up with painful or otherwise unpleasant treatment in exchange for longer life because they understand both the purpose and the limited duration of the treatment. Alzheimer’s disease is not like heart disease or cancer, where many people accept painful treatment because they continue to enjoy life, conditions where only at the very end of the journey do patients sometimes perceive life as no longer worth living.

Maybe it’s time for me to show this is true and not merely to argue that it is. To that end, I am designing an interlinked series of short educational videos for caregivers intended to help them make crucial medical decisions for their loved ones—decisions about the goals of care as the illness progresses and specifically about hospitalization and hospice care. I will test whether those videos, which will show caregivers what treatment is really like, influence the choices they make about medical care for the person with dementia. Then, perhaps, we will be able to provide medical care to patients with dementia that is more consistent with their wishes.

Call the Concierge

My mother got a letter from her doctor this week informing her that he is going into “concierge medicine.” If she wants to keep him as her doctor—and she does (she’s 88 and has known him for years)—she will have to sign up for his new plan, paying a retainer fee of $3500 a year on top of what she pays for Medicare A, Medicare B, and Medicare D. I know what her doctor will get from this arrangement: if he succeeds in limiting his practice to 400 people, he will collect an additional $1.4 million of revenue a year, minus whatever he loses by shrinking the size of his panel. But what would my mother get?

According to the doctor’s letter, she would have more same-day appointments in case of emergency (though major emergencies still require a visit to a hospital emergency room), longer appointments (she's never felt short-changed by an appointment), and the opportunity to exchange emails with her physician (which she already does). She would also have the services of the doctor’s staff to help make appointments with specialists and to coordinate care (which the secretary does routinely)--this is presumably the “concierge” in the title of the practice. In short, the only reason for my mother to shell out a retainer fee of $300 every month is to be able to keep her current physician.

As is probably pretty clear, I don’t much care for this arrangement. Not only do I see little benefit for my mother, but I am also concerned it is creating a two tier medical care system, with one level for affluent people who can afford concierge medicine and one level for everyone else. And while there are ample examples of inequity in medical care today, the tendency is to try to remove disparities, not to create new ones. In addition, fee-for-service medicine in general and solo practice medicine in particular are widely reported to be on their way out—they have proved to be poor strategies for addressing the complex needs of today’s patients. Concierge medicine is a trend in the wrong direction. What’s going on here?

By all accounts, the first “concierge practice” was set up by two internists in Seattle in the mid 1990s. They were interested in having a small panel of patients with whom they could have a strong, personal relationship. Like all the doctors who have joined or established concierge practices since then, they wanted less stress and less burn out and they wanted to practice a more old-fashioned style of medicine, less driven by algorithms and protocols and with fewer forms to fill out. 

The appearance of concierge practices engendered some fairly heady criticism: one physician expressed concern that while such practices increase continuity of care for the minority of patients who stay in the practice, they increase discontinuity of care for those who leave and have to find care elsewhere.  He worried that these practices would undermine the cross-subsidization that is essential for universal coverage—the whole idea of insurance assumes that risks are shared across a large pool. Concierge practices are a form of cherry-picking patients. And shrinking patient panels could further exacerbate the shortage of primary care physicians. 

What we actually know about how many such practices there are and how they operate is based largely on two government-commissioned reports. A GAO report in 2005 found that there are several different types of concierge practice. There are the fee-for-extra-service models, which accept health insurance but offer something that conventional insurance doesn’t pay for (longer visits, more visits, round the clock physician access). There are the fee-for-care models, which do not accept health insurance and where the fee paid to the physician is supposed to cover care. And there are hybrid models, where physicians offer the patient a choice of either arrangement. The GAO found only 146 doctors with concierge practices in its survey. A MedPAC study (Medicare Payment Advisory Commission) in 2010 tried to determine if concierge practices were complying with Medicare rules  and whether they were adversely affecting the access of Medicare beneficiaries to care. This report identified 756 concierge practices, though they obtained estimates of anywhere between 1000 and 2500, nation-wide. It found the average panel size was 250, about a tenth that of the physician’s practice before he or she adopted the concierge model. MedPAC concluded that after 15 years of concierge practices, they remained a small niche and not a major threat to the viability of primary care.

To add to these reports, a recent survey of 13,500 physicians for the Physicians Foundation found that nearly 7% of respondents said they planned to switch to concierge model in the next 3 years. And another survey, this one by the Concierge Medicine Research Collective found that among physicians already practicing concierge medicine, 55% were very satisfied, 27% were somewhat satisfied and 18% were dissatisfied. The majority, but not everyone (62%) reported an increase in income. 

All these surveys, statistics, and ethical analyses pay relatively little attention to the root problem that is leading to the growth of concierge medicine. What does the very existence of this phenomenon say about the state of primary care? And what can be done to make both physicians and patients happier?

I suspect that the fundamental driver behind concierge medicine is the belief by many physicians that to do a good job caring for their patients, especially those with multiple medical problems or advanced illness, common characteristics of elderly patients, they have to spend a lot of time. And if they spend as much time as they feel is appropriate, they won’t be able to see as many patients as they do now and will therefore make far less income than they do now. If this analysis is correct, we need to ask: how much time do doctors really need to spend? Are they spending their time in the best possible way today? Do insurance company mandates to do various tests and fill our certain forms promote or hinder good care? What are the financial constraints under which they are working?

Now it seems to me that while time is one aspect of taking good care of patients, the real problem with primary care is that doctors need to do a better job sharing responsibility with physician extenders, patients, and their caregivers. Someone needs to make sure patients are adhering to the standard of care, say, for diabetes, by assuring they get appropriate foot care and the necessary eye exams--but it doesn't have to be a doctor; it could be a nurse. Someone needs to monitor chronic conditions, say weight in a patient with heart failure--but it could be the patient's spouse. Physicians need to spend more time on advance care planning (talking about what overall approach to care makes sense for a given patient) and less time on assuring tight control of diabetes in 85-year olds, where such control is probably not useful. The way forward is the team model used in geriatrics, palliative care, and HIV care, enhanced to include patients and patients' caregivers.

But primary care doctors are unlikely to pursue this path with current reimbursement rates. Geriatrics, palliative care, and HIV doctors are among the least well paid of any physicians. To help stimulate greater use of this model, compensation for primary care will need to rise and, in this era of cost-cutting, the only way that can happen is if specialty income falls. A 2013 report on physician compensation in the US found that the mean salary for an orthopedist is $405,000 (with many orthopedists earning considerably more than this), the mean salary for a cardiologist is $357,000 and the mean salary for a family physician is $175,000. Internists are paid only slightly more, at $185,000. Moreover, salaries for orthopedists have gone up by 27% and those of cardiologists by 13% since 2011, compared to salaries for family physicians, which have gone up only 5%. My proposal will encounter enormous resistance by the well-organized lobbies for subspecialty medicine. But it is an important first step.

If we want to stem the growth of concierge medicine—and I think we should—we will need to repair primary care.

No News is Bad News

It’s that time of year again, when the Alzheimer’s Association releases its annual report: Alzheimer’s Disease Facts and Figures. The 2014 edition doesn’t tell us much that’s new—which amounts to a good deal of bad news.

The facts boil down to these: Altogether, about 5.2 million people in the US have one form of dementia or another; 5 million of them are 65 or older. The older you are, the more likely you are to have dementia, except among the very old (those over 85), who have to be pretty vigorous to make it to extreme old age. Fewer than 4% of people under 65 have dementia; in the 65-74 year old age range, it’s 15%; among those 75-84, it’s 44%, and in those over 85, it’s 38%. As far as cost goes, Medicare will probably spend $113.8 billion on dementia care this year (much of that for hospital and post-acute care) and Medicaid will spend another $37.8 billion (much of that for nursing home care). Out of pocket expenditures for people with dementia will amount to another $36.8 billion, leaving another $28.8 of miscellaneous expenses (miscellaneous care seems to be a big ticket item). None of these “costs” include the time spent by informal (a euphemism for unpaid) caregivers. In 2013, caregivers devoted an estimated 17.7 billion hours to providing care to people with dementia. When this is valued at the rather low ball figure of $12/hour, it comes to a total of $220 billion/year. It's worth noting that these informal caregivers do all kinds of things, ranging from cooking and shopping to personal care to some pretty sophisticated medical care.

What is new in the current report is the special section on women and dementia. This refers both to women with dementia and to women who care for others with dementia. Women account for nearly 2/3 of all cases of dementia, principally but not entirely because they live longer than men. They are 2.5 times more likely than men to provide round-the-clock care for a family member with dementia. And what is most striking in this report is the results of a survey asking older women about their perceptions and fears regarding health in general and dementia in particular. When women age 60 and older were asked what disease they feared most, Alzheimer’s was at the top of the list, with 35% of respondents indicating it was their number one bugaboo. Cancer, incidentally, was number two, chosen by 23% of respondents. 

The take home message is that no matter what the epidemiologic studies from England and Denmark tell us about declining rates of dementia, and no matter what “brain games” companies such as Lumosity say will keep us sharp, dementia remains an enormous problem. It’s responsible for a great deal of pain and suffering, both among people with dementia and their caregivers. It’s a major contributor to out-of-control health care spending. But the good news—and there is a little bit of good news here—is that many of these problems have solutions that are within our reach. We cannot, at this point, do much about getting dementia in the first place. But once someone has the diagnosis, there’s a lot we already know how to do to help caregivers, on the one hand, and to avoid burdensome, unwanted, and costly medical tests and treatments on the other. We just need the will to create and finance programs that educate and support caregivers, physicians, and health care insurers.

Lotsa Lumps

Geriatrics would be a good deal easier if every older person suffered from just one medical condition. But most elderly people have more than one chronic disease and the older they are, the more chronic conditions they are likely to have. Since “multiple chronic conditions” is a mouthful, researchers coined the term “multimorbidity,” an only slightly less awkward way of expressing what is probably one of the most critical features of geriatric existence. It’s so critical because the best medical treatment, known as “evidence-based medicine,” is founded on studies of patients who don’t have multimorbidity at all. They are generally perfectly healthy except for the single disease being studied. So when we tell a patient that “studies show’ that blood pressure should be below 140 and that the best medication to take if the blood pressure is elevated is a diuretic, we mean that if the only problem is high blood pressure, then taking the diuretic is the best way to lower the risk of bad outcomes such as strokes and heart attacks. But if the patient also has another chronic condition, say Parkinson’s disease, which is being treated with the medication L-dopa (Sinemet), then giving that patient a diuretic to lower blood pressure could backfire—long before any heart attacks or strokes were prevented, the patient might fall down (both L-dopa and diuretics contribute to sudden falls in blood pressure when a person stands up) and break a hip. Simply assuming it makes sense to apply multiple guidelines to a patient with multiple problems can result in medication lists a mile long that cost a fortune and that cause more problems than they solve. So multimorbidity is a big deal in geriatrics. Now, for the first time, multimorbidity is getting the attention it deserves.

A couple of years ago, the American Geriatrics Society set up a task force to develop an approach to multimorbidity for physicians. This group generated a report that lays out the basic principles that should underlie care for a patient with multiple chronic conditions. And a recent symposium brought together physicians and researchers from a variety of backgrounds to come up with strategies for generating a better evidence base, for designing new guidelines, and for carrying out appropriate systematic reviews for patients with multimorbidity. The results of the symposium are published as 3 articles along with an editorial in the April issue of the Journal of General Internal Medicine. 

I wish I could report that these new guidelines-for-producing- guidelines came up with the solution to the problem of multimorbidity. I wish I could summarize the gist of these papers. They constitute a valiant attempt to find an answer to a vexing problem. They go far in enumerating the many obstacles to a solution. They provide an exhaustive list of all the issues that must be addressed in the future. But in the end, they advocate a technical solution to a dilemma that can only be addressed by eliciting patient preferences and by use of clinical judgment.

There are just too many different clusters of chronic diseases for physicians to come up with recommendations that clearly and unambiguously apply to a given patient. Consider just the 10 most common chronic diseases. For any group of 10 diseases, there are 45 pairs of diseases and 120 groups of 3 diseases and 210 groups of 4 diseases—and that doesn’t include clusters of greater than 4, even though plenty of older people have, say, arthritis, diabetes, heart disease, emphysema, and kidney problems. Even if it were practical to design a guideline for all the most common groupings, which it isn’t, that wouldn’t be good enough. What matters is not just how many diseases you have, but how sick you are and how well you can go about your day to day business (what geriatricians call your functional status). Someone who lives independently and shops and cooks for herself is in far better shape—and able to withstand a new medication or a brief hospitalization—than her counterpart with the same underlying chronic conditions who lives in a nursing home and needs help with all her daily activities.

I tend to agree with an article in the British Medical Journal that argues that “multimorbidity introduces clinical uncertainty in a way that is unlikely to be resolved by ever more sophisticated guidelines.” The authors advocate listening to patients to find out what matters to them and using clinical judgment to try to achieve their ends. A similar recommendation came from the American Geriatrics Society task force,  which even created a pocket card summarizing their key conclusions: consider patient preferences, consider prognosis, consider interactions among treatments, review the entire care plan, and communicate and decide on treatment together with patient and family. 

More research that measures quality of life outcomes as well as survival outcomes is of course important. But empirical studies alone will not solve the multimorbidity problem. Talking to patients about their overall health status and their goals of care, and figuring out together what tests and treatments make most sense in light of their status and their goals, just might.