March 29, 2010

This blog has moved

This blog is now located at
You will be automatically redirected in 30 seconds, or you may click here.

For feed subscribers, please update your feed subscriptions to

March 11, 2010

Drugged in the Nursing home

This week’s Boston Globe featured an article blasting Massachusetts nursing homes for having too many residents on antipsychotic medication, “Nursing home drug use puts many at risk.” It portrayed vulnerable grandmothers as sedated, mute, and drooling, transformed by drugs into tragically diminished versions of their earlier vivacious selves. The headline could have been from the 1970s when Mary Mendelson wrote her muckraking book, “Tender Loving Greed,” pillorying the nursing home industry. Has nothing changed in the last 30 years?

A great deal has changed. In the 1980s, Congress responded to the deplorable state of nursing home care with the Nursing Home Reform Amendments, part of the Omnibus Budget Reconciliation Act of 1987 (OBRA-87).This major piece of legislation sought to make nursing homes free of both physical and chemical restraints. And to a large extent it worked. Translated into regulations in 1991, OBRA-97 led to enormous declines in the use of antipsychotic as well as other “psychoactive” drugs such as benzodiazepines.

Then new, “atypical” antipsychotic medications came along. Touted as equally effective but less toxic, risperidone (Risperdal), olanzapine (Zyprexa), and quetiapine (Seroquel) appeared with growing frequency in nursing homes. Physicians were mandated by OBRA-87 to monitor the use of these drugs—to look for side effects such as Parkinsonian symptoms or drops in blood pressure. They were also supposed to limit prescription of these new drugs, along with other old fashioned or “typical” antipsychotics such as haloperidol (Haldol), to well-defined situations. The drugs were to be used for chronic schizophrenia and for dementia with psychotic features, unresponsive to alternative treatment. As the prevalence of dementia in nursing homes increased, so too did the prevalence of behavior problems: kicking, biting, smearing feces, and screaming. Less dramatic but nonetheless challenging were other behaviors that were also on the increase, such as pacing, wandering into other residents’ rooms, and urinating in trash cans. These behavioral disturbances wee extremely difficult for nursing home staff to manage. Convinced that the new “atypical” antipsychotics were the solution, physicians prescribed more and more of these drugs.

Some residents did improve. Most of them were not transformed into zombies by the medication. But the behavioral symptoms of dementia often come and go without pharmacological intervention. While many physicians were confident they were seeing a benefit of the drugs, others were not so sure. A large randomized trial
found that all three of the leading antipsychotics were equally effective in controlling symptoms—and indistinguishable in efficacy from placebo.

Nursing home physicians were skeptical. The study was conducted among outpatients. Maybe long term care residents were different from outpatients—surely they were more demented and had worse symptoms. But at around the same time came other disturbing news: the atypical antipsychotics were associated with a risk of sudden death. The FDA issued a new “black box” warning to physicians prescribing these drugs,followed 3 years later by another black box warning against using the older, “typical” antipsychotics. The drugs were not prohibited, but sober commentators advised prescribing them with great caution, restricting their use to patients with longstanding psychiatric illness and checking an EKG before and after starting the drugs.

In the face of all this bad news, how can it be that a recent study of over 16,000 nursing home admissions found that 29% received at least one antipsychotic over the course of a year? Of these, just about one-third had no identified clinical indication for the drug. Residents admitted to nursing homes with the highest baseline prescribing rates of antipsychotics were 1.37 times more likely than those admitted to nursing homes with the lowest baseline rates to receive a new antipsychotic prescription in the coming year.

Over-use of antipsychotics in the 1970s and 1980s was bad enough. In the intervening years, we have had regulation (OBRA-87), scientific studies of both efficacy and toxicity, and warnings from the FDA. What is going on?

The problem is that agitated behaviors in demented nursing home residents are a major challenge. There is no simple solution. Staff training can help. Increased staff: patient ratios can help. But with the rise of alternative options for care such as assisted living and a decline in the total number of people living in nursing homes, those individuals who do live in a nursing home tend to be more demented and have greater behavioral problems than ever before: in 2007, 69% of US nursing home residents had cognitive impairment, with 42% diagnosed with moderate to severe impairment. Solving the problem of how to care for demented nursing home residents will require far more sweeping changes than adding a few in-service programs for staff or hiring a few more nursing aides. The key is culture change.

To understand the role of culture change, we should look at the use of feeding tubes in nursing home patients with advanced dementia. The feeding tube story is in many ways similar to the antipsychotic medication story: the practice persists despite studies showing that feeding tubes do not prolong life, they do not prevent aspiration pneumonia, and they do not promote healing of pressure ulcers, the reasons given for their use. There is tremendous state-to-state variability in the use of feeding tubes. And while studies have identified a variety of institutional factors associated with feeding tube use (for-profit status of the nursing home, absence of advance care planning discussions, speech therapist on staff), there has been no deep understanding of why the rates vary so dramatically.

In the very same issue of the Archives of Internal Medicine that reported on the national use of antipsychotic drugs in nursing homes, Susan Mitchell and her colleagues presented an ethnographic study of two nursing homes to explore the role of nursing home culture in promoting the use of feeding tubes. In one nursing home, 42% of residents with advanced dementia had feeding tubes while in the other nursing home, only 11% had feeding tubes. The study found startling differences between the cultures of the two institutions. The low use nursing home had a home-like environment, specially trained nursing assistants, and lots of family involvement in decisions of the goals of care. The high-use facility, by contrast, had an institutional atmosphere, inadequately trained nursing assistants, and a focus on regulatory compliance rather than quality of life.

Preliminary evidence suggests that the same nursing homes that foster a culture conducive to hand feeding rather than tube feeding also have a low use of antipsychotic medications. Providence Mount St. Vincent’s, a nursing home in Seattle, Washington that pioneered the culture change model, reported a 100% decrease in the use of antipsychotic medications. If this is confirmed, it will provide compelling evidence that what matters most to nursing home residents, beyond rules and regulations, is designing a community that is patient-centered, where staff are cross-trained to provide multiple tasks, and that focuses on relationships. It is in homes like this that even individuals with advanced dementia may be able to flourish without either feeding tubes or antipsychotic medications.

January 04, 2010

When Lawyers Practice Medicine

I sat with my elderly parents and their estate planning lawyer, going over documents. The lawyer had updated their will, revised their powers of attorney, and done various other lawyerly things. She handed my parents page after page for them to sign. After an hour of this, my eyes began glazing over. But then I heard her say “and here’s your living will,” and I was all ears.

My work as a geriatrician and palliative care specialist revolves around advance care planning, discussions with patients about their preferences for medical care in the event they lose the ability to make decisions. It’s a complicated enterprise that begins with clarifying the person’s overall health, then moves on to determine the goals of care—what is most important at that point in time—and then seeks to translate those goals into a plan of action. What I have learned by doing this with hundred of patients and their families, as well as by studying and writing about it, is that it is a process. Other researchers and practitioners in the field agree that advance care planning is not about completing a form, it’s about discussing prognosis, it’s about explaining what medical interventions can and cannot achieve and the burdens associated with them, and it’s about factoring the patient’s values and preferences into this complex discussion. What was an estate lawyer doing handing out a living will document and saying “sign here?”

Clearly the lawyer thought the document was self-explanatory and that most people would want to sign such a form. Clearly she believed it was her responsibility to offer such a form to my parents and she felt she was providing them the opportunity to retain control over their medical care at the end of life.

The irony is that Massachusetts, where this encounter between my parents and their attorney took place, does not have living will legislation. In Massachusetts, the only legally recognized form of advance care planning is a health care proxy—choosing who is empowered to make decisions if the patient loses the capacity to speak for himself. The Massachusetts legislature deliberately opted not to accord official status to living wills because they are vague, ambiguous, and seldom applicable to real life medical situations.

The living will my parents were given is a classic example of pseudo-precision: “If a situation should arise in which there is no reasonable expectation of my recovery from extreme physical or mental disability, I direct that I be allowed to die and not be kept alive by medications or artificial means or procedures which serve only to prolong the process of my dying,” it begins. What is a “reasonable expectation of recovery?” A fifty-fifty chance? Is a 30% chance good enough? 10%? What does “recovery” mean anyway? Going home and living independently? Living in a nursing home and needing help with bathing and dressing? Going from unconsciousness and total paralysis to wakefulness and the ability to move one finger?

But there is more—the document seems, at first glance, to spell out the answers to these questions. It says “without limitation, I intend these instructions to apply if I am (i) terminally ill, (ii) permanently unconscious, or (iii) conscious, but have irreversible brain damage and there is no reasonable expectation that I will regain the ability to make decisions and express my wishes.” What is meant by “terminally ill?” Does it mean conforming to the Medicare hospice definition of having a life-expectancy of 6 months or less, if the disease follows its usual course? Does it mean death is imminent—in the next few hours or days? Or does it mean having a disease that is uniformly fatal, such as Alzheimer’s disease, which lasts 3-5 years, sometimes longer, from diagnosis until death? What is “extreme physical or mental disability?” Does this mean the most advanced stage of Alzheimer’s, or does moderately severe dementia—in which the individual can walk and talk, but has completely lost his short term memory and needs help with bathing, toileting, and personal care—qualify?

That’s not all. If my parents did seem to fall into any of the 3 categories of terminal illness, permanent unconsciousness, or irreversible brain damage, then it specifies that they would not want “ (a) electrical or mechanical resuscitation of my heart when it has stopped beating; (b) artificial nutrition or hydration when I am unable to take nourishment by mouth; (c) mechanical respiration when I am no longer able to sustain my own breathing; and (d) medications, tests and treatments for any purpose other than comfort.” There is no explanation of what any of these interventions entails and their benefits and burdens, nor is there any explanation of what the alternative might be. In principle, if my parents had a stroke that impaired decision-making capacity and the ability to swallow, but were awake and alert, they would not be given a feeding tube for nutrition, even if they were hungry and indicated they wanted to eat. In principle, if they were diagnosed with a fatal cancer and had a life expectancy of less than six months and were unable to make decisions for themselves, they could not receive radiation treatments to prolong life or antibiotics to cure a pneumonia.

Not only does this kind of living will potentially limit the use of treatment that might well be entirely appropriate and consistent with my parents’ wishes (if they understood what they were signing), but it also fails to protect them from overtreatment in most of the clinical situations which they are likely to encounter.

Whatever is meant by being “terminally ill” or suffering from “irreversible brain damage” with “no reasonable expectation that I will regain the ability to make decisions,” these are not the situations in which older people commonly find themselves. They find themselves in a whole host of situations in which they might well want limitation of treatment. Many 85-year olds, for example, have mild dementia, heart disease, arthritis, and diabetes. If they develop a severe pneumonia, another common scenario, this might well precipitate “delirium,” an acute confusional state in which they were unable to make medical decisions. Would they want to be put in an intensive care unit on a ventilator? The living will, which is supposed to guide proxies and physicians, is silent on this issue.

This seemingly innocuous living will, in short, failed to protect my parents from excessively burdensome treatments in a variety of very possible medical situations, and could prevent them from receiving desirable treatments in other situations. So how could they assure they would get the kind of treatment they wanted if they were unable to speak for themselves?

The answer is that they could discuss with their physician what sorts of medical interventions are consistent with their goals. Frail elders, for instance, if they wish to remain independent as long as possible, even if it means a few months less life, should opt to forgo attempted cardiopulmonary resuscitation. The reason is that CPR, when performed in this setting, is rarely successful, and in those instances where the person survives, he almost always experiences a significant decline in his ability to function on his own. Ideally, the health care proxy is present when the physician talks to the patient about preferences for care, so he will be well-positioned to carry out the patient’s wishes. It is important for the physician to document such discussions in the patient’s medical record. Formal “instructional directives” that spell out just which procedures a patient thinks he would want under what circumstances may be useful as a supplement to advance care planning discussions. An example is the “Five Wishes,” available from Aging With Dignity.

It is entirely appropriate for lawyers to give their clients a health care proxy form to sign—although it should be pointed out that no legal input is required to choose a surrogate. Anyone can print up the official Massachusetts health care proxy form and sign it, with two adult witnesses. There is no need to pay an attorney $100/hour for such a form. It is also reasonable to recommend to clients that they discuss with their doctor their preferences for medical care—in the context of their medical situation and their personal goals. At best, presenting clients with a living will and suggesting they sign it reduces the complex process of advance care planning to filling out a form. At worst, it is practicing medicine without a license. Is the lawyer in any position to explain the benefits and burdens of the medical interventions referred to in the living will, such as a ventilator (“mechanical respiration when I am no longer able to sustain my own breathing”), a feeding tube (“artificial nutrition…when I am unable to take nourishment by mouth”), or cardiopulmonary resuscitation (“electrical or mechanical resuscitation of my heart when it has stopped beating”)?

Physicians have shirked their responsibility by only sporadically engaging in advance care planning with their patients. The result is that lawyers are taking up the slack. Many state legislatures have contributed to the unfortunate conceptualization of advance medical planning as a legal issue through advance directive legislation: unlike Massachusetts, most states have living will laws as well as health care proxy laws and even prescribe what forms are legally recognized in their state. We need to reclaim advance care planning as a medical intervention that can prevent both over-treatment and under-treatment near the end of life.

November 30, 2009

The Heart of the Matter

Just before the annual deluge of holiday visitors began inundating the Magic Kingdom, over 20,000 people gathered in Orlando, Florida to attend the American Heart Association annual meeting. At one late-breaking scientific session, Dr. Joseph Rogers, a Duke University cardiologist, presented the results of a study of a nifty new cardiac device, the HeartMate II (to distinguish it from its predecessor, HeartMate I, also known as HeartMate XVE). This remarkable 390-gram contraption is the next best thing to a heart transplant for patients with severe heart failure. It is a “ventricular assist device,” (VAD) implanted in the abdomen, hooked up to the heart to propel blood through a failing ventricle, and powered by an external generator. The study, published the very same day in the online version of the New England Journal of Medicine, showed that patients with the new device lived longer and better, with fewer complications, than those who got the earlier variety.

Just six years ago, the Centers for Medicare and Medicaid Services (CMS) approved use of the HeartMate I in Medicare patients as “destination” or definitive treatment of severe heart failure. Based on the results of a 2001 study, the REMATCH (Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure) trial, CMS concluded that the device was “reasonable and necessary” for the treatment of advanced heart failure and therefore, as required by its 1965 Congressional mandate, agreed to pay for implantation of the device.

Physicians, however, were not convinced the device was reasonable or necessary. In fact, the number of patients receiving a ventricular assist device has fallen every year since CMS approved its use. Between March, 2006 and June, 2009, a total of only 1664 HeartMate I’s were implanted (just under 500 per year) and 90% of them were used as “bridge to transplant,” that is to tide patients over until a heart became available for transplant, rather than as “destination therapy,” or permanent treatment. Physicians evidently were not impressed by the data that patients who had a ventricular assist device lived longer than those receiving “maximal medical therapy.” Even with the device, one-year mortality remained high at 48% and two-year mortality at 77%. Moreover, HeartMate I recipients were at risk of stroke, infection, and device malfunction.

Will physicians be more enthusiastic about the new HeartMate II? It is smaller—small enough to fit into a woman, unlike the earlier version. It uses a continuous rotary system rather than a pulsatile mechanism, which means it’s quieter and more durable. And the new device is clearly superior to the old one—the randomized study reported in Orlando found that 46% of those with the new device were alive after 2 years without having sustained a debilitating stroke or required replacement of the device, compared to 11% of those who got the original model. Quality of life was also improved, with recipients of the new HeartMate able to walk further and do more for themselves.

Are these results good enough? They will almost surely be good enough for the Food and Drug Administration to approve use of the device as destination therapy and for Medicare to follow suit by approving reimbursement in Medicare patients—currently Medicare pays an average of $177,000 for a hospitalization in which a ventricular assist device is inserted. Whether physicians will regard the new pump as ready for prime time remains to be seen.

The deeper question is whether we should be putting more and more resources into shoring up failing hearts in patients near the end of life or whether, instead, we should concentrate on palliative care for patients dying of heart failure and on prevention to avoid the development of heart failure in the first place.

Palliative care is an effective, well-regarded approach to end stage heart failure. It focuses on comfort and on supporting patients and their families as the end draws near and may involve enrollment in hospice. Individuals whose heart cannot pump properly repeatedly suffer from fluid backing up into their lungs, swelling their legs, and filling their abdomen, causing shortness of breath, difficulty walking, and fatigue. Instead of hospitalizing these patients, often transferring them to the Intensive Care Unit where they may be attached to a ventilator that forces oxygen into their lungs and put on an intravenous drip to try to stimulate the heart to contract more forcefully, palliative care strives to keep them home. At home they use oxygen and medication such as morphine to help them breathe. Interestingly, the studies of the left ventricular assist device compare patients receiving one particular device to those getting another type (the new study) or patients receiving a device to patients receiving “maximal medical therapy” (the earlier study), but never compare patients receiving a device to those getting palliative care.

Prevention makes a lot of sense in the case of heart failure. Heart failure is the leading cause of hospitalization in individuals over age 65. Americans have a 20% life time risk of developing the condition and treatment costs the US upwards of $30 billion per year. But we know what causes the heart to fail: it is high blood pressure, diabetes, or coronary artery disease (which in turn is caused by some combination of high blood pressure, diabetes, high cholesterol and smoking). And every one of these predisposing conditions is treatable or preventable. While there are few good studies of what it would cost to institute a comprehensive set of measures to prevent heart failure, we do know that 82% of the cardiac events in one observational study, the Nurses Health Study, were attributable to diet and lifestyle issues. An Australian analysis suggests that the best way to prevent heart failure is by a combination of medical preventive strategies (proceeding patient by patient to assess risk factors and intervene to treat high blood pressure, diabetes, and high cholesterol) and public health strategies (using legislative and environmental approaches to promote a healthy lifestyle).

America has its heart set on the technological fix—the HeartMate II is the latest in a series of ingenious approaches to treating the chronic diseases afflicting older people. But its time to get to the heart of the matter, trying to prevent the major scourges that we have the knowhow to avoid, and palliating those who, despite our best efforts, are dying.

October 15, 2009

Balking at Baucus?

Democrats and Republicans alike are balking at the Baucus proposal, the latest health care reform proposal to emerge from Congress. Not surprisingly, this enormous bill has something for everyone to hate. It attempts to legislate improvements in access and quality while simultaneously addressing cost.

But what is sometimes lost in the furor is that there are actually two kinds of costs that reform must deal with: the cost of implementing the new programs (principally the cost of requiring that most Americans have health insurance, which means subsidizing premiums for individuals with low incomes); and the rising cost of health care, principally the Medicare program, which is expected to run out of money in 2017.

The Baucus proposal has a price tag of $774 billion over 10 years (though in the last few days, various modifications have added another $28 billion), which refers to the first kind of cost. But how does the Baucus proposal address the second cost issue, containing spending?

The Congressional Budget Office predicts that total spending on health care will reach 36 percent of gross domestic product by 2035, with Medicare costs alone comprising 8 percent of GDP. The vast majority of the increase in Medicare costs will be attributable to increased spending per enrollee, not to the aging of the population. And the single most important factor accounting for the increases is the emergence, adoption, and widespread diffusion of new medical technologies.

Some of these technologies are clearly beneficial; others are beneficial to some of those who receive the new drugs or devices, but are unnecessary or even harmful to numerous others for whom they are also prescribed. And some of the beneficial technologies produce only a very small benefit but at an extremely high cost, as set by the pharmaceutical or device manufacturing industries. Effective Medicare cost control will need to go beyond appealing proposals of dubious financial benefit such as expansion of preventive services and elimination of Medicare fraud (though this could save as much as $17.2 billion per year if fraud were totally eradicated).

What else does the Baucus bill suggest? It makes three significant proposals that actually might work. Collectively, they seek to modify the culture of medicine by altering the way medical care is delivered and how it is paid for.

The beauty and the peril of its strategy is to create an infrastructure for designing a program rather than actually spelling out what the program will be. The beauty is that it does not require making politically unpalatable claims, for example that CMS may need to stop reimbursing for technologies that are not cost effective. The peril is that the various centers and institutes the plan will spawn may not in fact do what they need to in order to control costs. Just what are these centers and institutes?

The most familiar one, because it is also found in the “Affordable Health Choices Act” of both the Senate and the House, is a nonprofit patient-centered outcomes research institute. This would conduct research comparing the clinical effectiveness of medical treatments. To be effective, its work would need to translate into the CMS reimbursement policies: it would have to become an extra-government agency along the lines of the much-maligned but crucially necessary British National Institute for Health and Clinical Excellence (NICE).

The second new institution is an independent Medicare commission charged with submitting proposals for reducing excess Medicare cost growth. Strong wording would effectively require Congress to do what the commission requested or come up with a fiscally equivalent alternative. This is an example of the politically astute approach of leaving open just what the independent commission would recommend.

The third institution is an innovation center within CMS to test, evaluate, and expand different payment structures and methodologies to foster patient-centered care, improve quality, and slow Medicare cost growth. Among other endeavors, the center is mandated to pilot the development and evaluation of a system of bundled payments for a single episode of care – that is, a single payment covering care in the office, the hospital, and the skilled nursing facility. For bundling to work, doctors, nursing facilities, and hospitals will need to band together to form accountable care organizations, and Baucus promotes the establishment of these organizations by offering them a share in the cost savings they generate.

Will the research institute, the commission, and the innovation center, with it assorted pilot programs, do the job? Without knowing just what they will recommend, it is impossible to be sure. But their focus will be on the right areas: figuring out what works and what doesn’t, determining who benefits from new technologies and by how much, and developing systems of care that assure that patients get the interventions that work but not those that don’t. It’s our best shot to date.

This article appeared on the Hastings Center HealthCareCostMonitor on September 24, 2009.

Read more:

End-of-Life Planning is No Conspiracy

Far from a left-wing conspiracy to deprive you of desirable care near the end of life, advance care planning - talking with one’s physician about goals and preferences for medical care in the event of incapacity - is the best way to ensure that your wishes are respected if you lose the capacity to speak for yourself. Because it is widely recognized as the ideal way to prevent both over-treatment and under-treatment, it has become the standard of care over the past 15 years.

Survey after survey indicates that most patients want to have end-of-life discussions but most do not have them. For example, a 2005 AARP survey in Massachusetts of members over age 50 found that 89 percent of those polled rated having honest answers from their doctor about end-of-life conditions as very important but only 17 percent had discussed their preferences for such care with their physicians. Study after study also shows that American patients continue to die in pain, on ventilators, and in the intensive care unit, though many of them did not want such treatment at the end of their lives. For instance, the Robert Wood Johnson study “Means to A Better End: A Report on Dying in America Today’’ found that 42 percent of all nursing home residents were persistently in pain, and the Dartmouth Atlas of Health Care found that 18 percent of deaths nationwide take place in intensive care units and 32 percent in the hospital, although the majority of Americans say they would prefer to die at home.

When physicians are asked why they do not regularly engage in advance care planning with their patients, they report that they do not have the time for such conversations. The legislation under consideration in Congress that would allow physicians to be reimbursed for an advance care planning discussion with their patients is simply an attempt to correct the imbalance between what patients want and clinical reality.

The current proposal is far from the first attempt to facilitate advance care planning. The federal Patient Self-Determination Act of 1990 mandates that all patients admitted to a healthcare facility receiving Medicare or Medicaid funding be asked if they have an advance directive (a written document specifying their wishes in the event of incapacity) and be given information about advance directives if they do not have one. Every state as well as the District of Columbia has advance directive legislation, which provides for the use of a living will (a statement about preferences in particular clinical situations) or designation of a healthcare proxy (a person to make medical decisions in the event the patient loses the capacity to do so) or both.

The only question we should be discussing is whether a federal mandate to allow physicians to bill specifically for discussions about goals and directives for care will actually succeed in stimulating advance care planning. That is why the bill calls for monitoring the use of the planning code - if it is ineffective in promoting end-of-life discussions, then we will need to find other strategies. The issue is not whether advance care planning is desirable; it is how best to encourage patients, families, and clinicians to have such conversations.

The reason it is important to have the conversations is that death is not optional. What is optional is how we will experience life’s last stage: Will we be in pain or will we be comfortable? Will we be in the hospital or at home? Will we be in an intensive care unit or enrolled in hospice? Discussions about end-of-life care have been shown to result in patients experiencing less depression, less pain, and less anxiety in their final days. Advance care planning gives us the opportunity to avoid both over- and under-treatment by allowing us to express our preferences and by discussing them with our personal physician.

This article appeared as an op-ed in the Boston Globe on August, 12, 2009.

Falling Down on the Job

It’s not often that a report appears that identifies a problem and at the same time, another report is released that offers a solution to that problem. But that may be what happened this month with the publication of an article describing the frequency of hip fractures and their complications and another article detailing an intervention that reduced the complication rate after hip fracture.

The first study presents data on hip fractures in the U.S. between 1985 and 2005. Previous data indicated that there are about 350,000 hip fractures every year in the U.S., almost all of them in people over 65 and the vast majority in people over age 75. Half of those who break their hip never return to their baseline level of mobility, 20% die within a year, and 25% of those who lived independently before the fracture require permanent nursing home care afterwards. The new report concludes that hip fracture rates and subsequent mortality among people over 65 are declining. But closer inspection of the data shows a different picture.

What the report really shows, as the authors indicate in the discussion section of their paper, is that there were two distinct time periods worthy of analysis. Between 1986 and 1995, the incidence of fractures increased in both men and women (corrected for age) but between 1995 and 2005, the incidence declined steadily. It declined a lot: among women, the incidence fell 25% and among men it fell 19%. But the mortality story is another matter. Looking at death rates at 1 month, 6 months, and at a year after hip fracture, the study found that between 1986 and 1995, mortality fell, but between 1996 and 2005, there was no improvement in mortality whatsoever.

The decline in the rate of hip fractures is good news, though it’s not entirely clear what caused it. The authors point to lifestyle factors, such as greater intake of vitamin D and calcium, along with increased exercise. They also point out that bisphosphonates, medications used to treat osteoporosis, were introduced at just about the same time that hip fracture rates began to fall, but these drugs can account for at most 40% of the reduction in risk.

The lack of improvement in mortality in the recent 10-year period, however, is troubling. Much of it reflects the fact that older people who fracture a hip are frail—and their frailty predisposes them to other medical problems in the following year. But the second article on hip fractures that appeared in a major medical journal this month offers a ray of hope. Reporting on the impact of co-management of hip fracture patients by orthopedists and geriatricians, investigators from the University of Rochester School of Medicine found enormous improvements in outcomes with their model.

Patients cared for in the “Geriatric Fracture Center” in Rochester were older than their counterparts who received standard orthopedic care (mean age 84.7 vs 81.6), they were more likely to have dementia (53% vs 21.5%), and they were much more likely to come from nursing homes and assisted living facilities (60.6% vs 12.2%). Despite all these differences, the co-managed patients fared far better. They went to the operating room much more quickly (in an average of 24.1 hours vs 37.4 hours), which has been shown to lead to better results. They stayed in the hospital less long (4.6 days on average vs 8.3 days) and they had fewer complications (30.6% vs 46.3%). In particular, they were less likely to develop delirium (acute confusion), infection, and blood clots, all of which predispose to early death.

The finding that “co-managed” patients did better than those cared for exclusively by orthopedists does not necessarily translate into improved survival rates. Some of the long-term mortality is unrelated to the hip fracture and simply reflects the reality that people who break a hip tend to be old and sick and that people who are old and sick die. But the component of the mortality from hip fractures that is amenable to intervention probably involves the post-operative management of patients. It is precisely those nasty complications of surgery that, over time, cause death. Finding a way to take care of individuals with hip fractures that decreases their risk of complications is an excellent strategy for preventing death.

Is the “Comanaged Geriatric Fracture Center” really the answer? It may be. The Rochester study was a retrospective cohort analysis, so it is possible that the findings reflect undetected differences between the patients or other differences in the institutions where care was provided. A randomized trial, in which some patients receive one form of care and others, chosen randomly, get the other type of care, would demonstrate decisively whether and how much co-management helps. But joint care by orthopedists and geriatricians is in widespread use in other countries such as Australia and New Zealand, with good results. And similar but less potent strategies have been used in the U.S., such as the use of geriatric consultation for patients on the orthopedic service. When this multidisciplinary approach was instituted at Oregon Health Sciences University and at the Johns Hopkins Bayview Medical Center, fewer complications, decreased length of stay, and earlier surgery were all observed. These systematic ways to assure improved care for older people with hip fractures can be instituted now. There’s pretty compelling evidence that they work and cost calculations suggest they save money, too. If we don’t start introducing this model into practice, we’re falling down on the job.

August 26, 2009

Messing with Medicare

Senior citizens have been coming out in droves to express concern that health care reform legislation will deprive them of what they now have. Although most of the provisions of the proposed bills deal with ensuring access to health care for the currently uninsured and with regulating private insurance companies, Medicare enrollees are worried that “the government” will “mess with Medicare,” a program which many seem not to realize is the paradigmatic “public option.” Just how would health insurance reform affect Medicare—in fact, not in fantasy?

Both the Senate’s “Affordable Health Choices Act” and the House of Representatives’ “America’s Affordable Health Choices Act of 2009” address access, quality, and cost-containment. The majority of the many sections of these bills deal with such areas as individual mandates, employer requirements, premium subsidies, insurance pooling mechanisms, and benefit design. Only a few have anything to do with Medicare. One provision, expansion of Medicaid, would increase the number of older people who are “dually eligible,” i.e. who would qualify for both Medicaid and Medicare. This would increase the coverage available to older individuals, not decrease it.

Under cost containment, the Senate bill would establish a “Health Care Program Integrity Coordinating Council” to prevent health care fraud, waste, and abuse, in both public and private coverage. Who could be against eliminating fraud and abuse? In fact, a recent article in the New England Journal of Medicine estimated that the annual price tag for fraud and abuse is $60 billion, with fully $36 billion of this related to Medicare and Medicaid, the current public programs--enough to pay for health care reform. The House bill has a few more provisions under the cost containment heading. These include reducing payments for potentially preventable hospital readmissions—that just means giving hospitals an incentive to solve your medical problem the first time you are hospitalized, instead of making you come back again and again to get it taken care of. In a similar vein, the House bill calls for hospitals to report hospital-acquired infections and suggests eliminating Medicaid payments for preventable infections (Medicare already does this), a strategy intended to encourage hospitals to keep you from getting such an infection in the first place.

In the arena of improving quality, both the Senate and the House version of the bills call for financial incentives to health care institutions to promote efficiency, for example by supporting a “medical home” that coordinates complex patient care. They also call for more research to study the effectiveness of health care services and procedures. This has nothing to do with taking away coverage for useful treatments. If physicians don’t know what works, they can’t treat patients in an optimal way. It’s that simple. And right now there is overwhelming evidence that doctors have routinely used ineffective therapy or have started with expensive and burdensome treatment when less expensive and often less invasive treatment would work just as well, simply because they did not know what would work best.

The bills also contain specific recommendations dealing with prevention that would increase coverage of preventive services under Medicare—for those preventive services that have been shown to improve outcomes. That means many beneficial screening tests that are not currently paid for would be covered in entirety and a few that other covered but that do not improve health would not be. So yes, a few things might be taken away from seniors—tests that are not helpful. But is this being deprived? Why would anyone want to undergo a screening test that doesn’t lead to better health? Knowing sooner that you have a disease just means suffering longer if it does not translate into more effective treatment of the disease.

The Senate and House bills actually include provisions for long term care insurance. Until now, no one in the legislature ever talked about long term care although it is critical to the health and well-being of millions of older people who need home care services or nursing home care. Medicare does not pay for nursing homes, so the 1.5 million people living in nursing homes either pay out of pocket or “spend down” and enroll in Medicaid.

Finally, there are a few miscellaneous provisions in both bills that would affect older people. The House version, for example, would eliminate the notorious “doughnut hole” in the Medicare Part D prescription drug plan, which requires older patients to pay in full for prescription drugs after they have spent $2510 and until they qualify for “catastrophic coverage,” once they have reached a whopping $4350 in out-of-pocket expenditures. The Senate version calls for educational reform to increase the supply, education, and training of doctors, nurses, and other health care workers in geriatrics and primary care.

What’s not to love?

June 19, 2009

Waste Not, Want Not

A recent commentary in JAMA begins with the sweeping condemnation: “An epidemic of waste blights the US health care delivery system.”Sad as this claim is, it is also encouraging, for who wouldn’t want to get rid of waste? Any attempt at reforming American health care should start by eliminating waste—but to do so, we need to know exactly where to find the waste in the system and how best to dispose of it.

Waste comes in several varieties. A RAND study identifies 3 principal flavors: administrative waste (such as excess expenditures on running a health insurance plan); operational inefficiency (for example, duplication of diagnostic tests); and clinical waste (for instance spending money on expensive drugs when cheaper ones would be equally effective).

A good starting point in thinking about reform would be to enumerate all the types of each of these forms of waste and to put a dollar amount on each of them. It turns out that the McKinsey Global Institute, the economics research arm of the McKinsey management consulting firm, has done exactly this kind of analysis.
In fact, McKinsey first undertook to identify where the waste is in the US health care system using 2003 data and then repeated the analysis with 2006 data. What’s fascinating about the report is both seeing where the waste is and discovering how much has changed in 3 short years.

The McKinsey report compares US expenditures on health care with those of a group of 13 countries belonging to the Organization for Economic Cooperation and Development (OECD). Reasoning that richer countries are more willing to spend money on health care, the report computes the Estimated Spending According to Wealth (ESAW), a prediction of how much a given country would spend if it were like the OECD average, adjusted for per capita GDP. In 2006, the US spent nearly $2.1 trillion on health care, or $6800 per person, eating up 16% of GDP. This was an increase of $363 billion since 2003—and $643 billion more than the average spent in the 13 comparison countries (Austria, Canada, the Czech Republic, Denmark, Finland, France, Germany, Iceland, Poland, Portugal, South Korea, Spain, and Switzerland) after adjusting for per capita wealth, or the ESAW

The excess spending produced no discernible health benefit for Americans. In fact, comparisons of the US health care system to that of OECD peers typically find that the US performed worse than everyone else. On one report card, where 1 is the best score and 6 is the worst, the US managed to achieve scores of 5 or 6 on 5 measures: quality care, access, efficiency, equity, and healthy lives.

The waste, according to the McKinsey Report, is in 5 areas: outpatient care ($436 billion or 68% of the excess), drugs ($98 billion or 15%), administrative costs ($91 billion or 14%), investments in health ($50 billion or 7%), and inpatient care ($40 billion or 6%).

By outpatient care, the Report means visits to physicians’ offices, same day surgery, dental care, and treatment in ambulatory surgical centers, diagnostic imaging centers, and other outpatient clinics. What’s striking is that this is the fastest growing component of wasteful care, growing at 7.5% each year since 2003. The higher costs in this sector are due principally to two factors: how much physicians are paid in the U.S. and the high profit margins for ambulatory surgery centers and diagnostic imaging centers. Extremely generous physician compensation adds $64 billion of costs to the system each year. This reflects what we pay specialists and our extravagant use of specialist care: while generalists in the US are paid somewhat better than their counterparts in the comparison countries, specialists are paid much more highly than in the rest of the developed world. Both ambulatory surgery centers and imaging centers are proliferating rapidly, attracted by operating margins of as much as 25%. The result is that the US has 4 times as many CT scanners and MRI machines as the average OECD country, and does 4 times as many imaging studies each year, again with no measurable benefit in terms of patient outcomes.

Drugs are the second major area of wasted spending (this includes both drugs spent by outpatients and drugs spend by hospitals). What’s interesting here is that Americans actually take 10% fewer prescription drugs than the average OECD patient each year. The source of the waste is that drug companies charge on average 50% more for brand name drugs in the U.S. than elsewhere in the world and Americans use a more expensive mix of drugs (a large fraction of brand name or newly released drugs).

Health care administrative costs are next on the list. Most of this relates to operating expenses and profits among private health insurance companies. The American multi-payer system, far from driving down costs through competition, adds costs to the system in the form of marketing, sales, and management overhead. Even Medicare, which until recently had far lower administrative costs than the private sector, has experienced a rise in costs since 2003 because of payments it makes to private plans to administer the Medicare Advantage Plans and the Part D drug benefit.

Investments in health are also higher than in peer countries after adjusting for per capita wealth. This means the US invests more in public health and basic research than other countries. This is one area where the excess, relative to other nations, may well be beneficial and not wasteful. Whether NIH and state public health departments are spending their money in the most effective way—whether they are getting maximum value for their investment—should be carefully examined.

At the bottom of the list, but still a major source of waste, is inpatient care. What’s fascinating here is that both the number and length of hospitalizations are shorter in the U.S. than anywhere else. The waste stems from the cost per hospital day, which is roughly twice the OECD average. This in turn reflects more spending on high tech equipment and subspecialty care.

At least as interesting as the breakdown of the types of waste is the way the distribution has changed in the last few years. The identical analysis by McKinsey using 2003 data found that by far the largest source of waste was inpatient care, accounting for slightly under half of all the waste. Why has the contribution of hospitalization gone from first to last? The mechanism of reimbursement for hospital care by diagnosis related groups, which gives hospitals a fixed amount of money depending on the reason the patient was admitted rather than a per diem rate, has been in effect for older patients since 1983. What has changed is the availability of a more lucrative alternative—day surgery and treatment in ambulatory surgical centers—for the treatment of conditions such as hernias and cataracts. The move from one site of care to another dramatically demonstrates the tremendous adaptability of the health care eco-system.

The implications for health care reform of this type of analysis are profound. If we truly want to decrease wasteful spending, both short term and long term interventions will be required. If we want to manage the flow of procedures from the hospital to the outpatient setting wisely, we need to regulate the proliferation of ambulatory surgical centers and diagnostic imaging centers and to control what they charge for their services. If we want to affect the balance of specialty and generalist care, it will not be enough to provide incentives to medical students to go into primary care: we will need to markedly decrease the phenomenal rate of reimbursement for specialists. If we hope to decrease waste in the medication arena, we will need to determine whether new drugs are better than old ones and if so how much better. We will need to institute some sort of price control over the pharmaceutical industry, at least by negotiating prices (not currently an option under Medicare Part D). And if we truly want to get rid of wasteful administrative costs, we need to consider a single payer system, something that is currently not even on the table as Congress debates health care reform. Finally, we need to recognize that some of what other countries deem wasteful, such as high tech care near the end of life in exchange for a minute chance of life-prolongation, Americans seem to value. If we want to get rid of this type of expenditure, we will need to change the culture that supports this approach, not merely the economic incentives that further facilitate it.

June 02, 2009

Is Death Optional?

Just how far attitudes and expectations about aging have changed in the last 60 years hit home on reading an article from the NY Times Magazine from 1950 called “Recharting Life for an Aging America.” The author, a physician, wrote that “To lead a long and happy life falls, for the average citizen, into the same category of irrational wishes as to be a millionaire or a movie star.” The reality, he said, is that most old people are “lonely, poor, ailing, crippled, ugly, [and] mentally and physically deteriorated.” Today, by contrast, older people take the possibility of ever increasing longevity for granted.

The change in perspective is dramatic and it’s very new: while Americans born in 1950could expect to live far longer than their grandfathers did, most of the improvement in life expectancy was due to decreases in infant mortality. It was only in 1970—5 years after the introduction of Medicare—that 65-year-olds could look forward to a longer period of retirement than any previous generation. By 2005, white men could anticipate another 17.2 years of life and white women 20 years.

Is the result really that Americans today fail to accept that death is inevitable? Or do patients appear to believe that death is optional because physicians seldom discuss life’s final stage and continue to offer treatments, even if they are of little or no benefit?

For all the lip service paid to informed consent and joint physician-patient decision making, older patients seldom understand their likely trajectory with and without a particular treatment. I recently saw a dramatic example of this problem in the course of palliative care consultation at a major teaching hospital in Boston.

The patient was a man in his late 70’s who had been hospitalized with a devastating stroke due to massive bleeding in his brain. He was being kept alive in the ICU with a variety of high tech interventions. The attending neurologist told the patient’s wife that the likelihood of any recovery was very small but that the full extent of his improvement might not be known for months. The doctor held out no hope of a full recovery and expected that if the patient did survive, he would require total care and would have little if any language capacity.

The patient’s wife didn’t think her husband would have wanted life-prolonging treatment if he would be left with profound limitations on his functioning, but she wasn’t absolutely sure. She wondered if she should authorize further vigorous treatment to “give him a chance.”

What quickly became clear to me was that the wife’s conception of what it would be like for her husband over the next two months if she opted for attempted rehabilitation and life-prolonging treatment bore little relationship to reality. She imagined that “going to rehab” would be as benign as taking a daily vitamin pill. I explained to her that after transfer to a rehab facility, he would likely suffer multiple complications such as pressure ulcers or pneumonia. He would probably be shuttled back and forth between the rehab facility and the hospital—and after all that, he would either die or be left extremely debilitated. Once she understood both what treatment would entail and how unlikely meaningful recovery was, she had no further hesitation: the right course of action for her husband was to focus exclusively on his comfort.

In today’s medical world, this kind of discussion is rare. If Medicare patients are to get appropriate care, and if costs are to be controlled, physicians must have such conversations. But since the focus is on life-prolongation throughout a physician’s training, with little attention to maximizing quality of life or to deciding when to stop, medical education will need to change.

Medicare and Medicaid pay just under $10 billion per year to hospitals in the form of General Medical Education funds to train residents. But as the Council on Graduate Medical Education observed in a letter to the Secretary of Health and Human Services in May, 2009, hospitals are not held accountable for how they spend the money. Their concern is with their own labor needs, not with training the next generation of physicians to manage chronic disease. It is time to monitor and regulate the way the federal government’s money is spent and require proficiency in end-of-life discussions along with disease management and care coordination.

Some experts believe that telling patients about the trajectory of illness with different treatment options won’t suffice because patients engage in magical thinking: physicians can lay out the various possible scenarios but patients will gamble that they will be the lucky ones who have the best outcomes. My experience suggests that most patients do respond to realistic discussions about their future, but the way to deal with the minority of patients who might want to try a treatment that has a vanishingly small chance of working is simply not to offer such interventions.

Decisions to take certain kinds of treatment in certain situations off the table should be made at the policy level. This will require holding National Institute of Health consensus conferences to determine a new standard of care for patients with a variety of chronic conditions such as dementia or heart failure in the last phase of life. The Centers for Medicare and Medicaid Services (CMS) will then need to give teeth to the practice guidelines that emerge from such conferences by agreeing to reimburse only for treatment that is consistent with those guidelines.

Perhaps the greatest challenge is that policy makers, who will need to endorse the kinds of changes I am suggesting, share the same expectations of ever increasing longevity as other Americans. A good starting point is therefore to limit treatments that are burdensome, unlikely to be effective, and expensive.

The next step will be to dispassionately analyze interventions that offer only a slight chance of benefit and that are expensive but that are not burdensome to patients. Some devices such as pacemakers have become increasingly acceptable as they have become smaller and implanting them has become safer and less invasive. Likewise, some cancers have become chronic illnesses because of the development of relatively non-toxic, targeted therapy. Patients naturally want potentially life-extending treatment if it comes in the form of a pill, without the nausea, vomiting, hair loss, and bone marrow depression associated with conventional chemotherapy.

Ultimately, policy-makers will have to take into consideration cost-effectiveness in deciding whether CMS will cover such treatments. Far less politically charged are the steps that should be taken immediately: regulating spending on graduate medical education and limiting reimbursement for treatment that comes at a high price to both patients and society without conferring any appreciable benefit.

A modified version of this piece appeared on the Health Care Cost Monitor, a blog of the Hastings Center.