At a time when most of the news about health and healthcare is pretty bleak, it’s nice to hear that one program for older patients is doing very well. The PACE program—Program of All-inclusive Care for the Elderly—has been around since the 1970s, when a neighborhood health center in San Francisco called On Lok pioneered a “social HMO,” providing both medical and social services using a capitated model. Since that time, PACE programs appeared in much of the country, each petitioning Medicare and Medicaid for special permission to collect one fee in return for taking care of some of the oldest and frailest patients, those who receive both Medicare and Medicaid and are eligible for nursing home placement. Ultimately, it became so clear that PACE provided high quality care while saving money that in 1997 it lost its experimental status: any integrated program that follows the basic tenets established by On Lok can be designated a PACE site. As of 2013, there were 98 PACE sites in 31 states, according to the National PACE Association. What’s new is that a recent study showed that PACE enrollees go to the hospital far less often than other dually eligible patients (patients receiving both Medicare and Medicaid, a particularly vulnerable and sick population).
The report, published in the most recent issue of the Journal of the American Geriatrics Society, reveals that the hospitalization rate was 79% higher among dually eligible in general than in PACE members (because not all patients are enrolled in PACE for an entire calendar year—they may join at different times or disenroll, usually due to death, before the end of the year—the report actually cites the rate per 1000-person-years, which comes to 539 per 1000 person-years in PACE members and 962 per 1000 person-years in the dually eligible). And when the investigators looked at “potentially avoidable hospitalizations,” admissions related to any of four medical conditions that some authorities believe should be treatable outside the hospital or in which optimal outpatient care should be able to prevent the kind of exacerbations of chronic conditions that trigger hospitalization, they found a similar pattern: the rate of potentially preventable hospitalizations was 2 1/2 times higher in the dually eligibles compared to patients in PACE (100/1000 person-years for PACE patients and 250/1000 person-years for dually eligible patients). Since hospitalization causes all kinds of problems in frail older people, and the cost of hospital care is the biggest piece of the Medicare budget, any program that produces these results deserves our attention.
So how does PACE keep patients out of the hospital? Unfortunately, the new study does not address this crucial question. The authors comment only that there are “many components that may affect hospitalization” and enumerate the unique features of PACE: it is capitated (the program receives a lump sum payment for each member to take care of all his or her medical needs), it provides care via an interdisciplinary team (physician, nurse, social worker, and physical therapists and others as needed), and it centers much of the care it offers in a kind of senior daycare program known as adult day health.
I suspect that the secret to PACE’s success is twofold. First, the members of the interdisciplinary team spend time with the patient and family to review the big picture—the enrollee’s major health problems and their implications—and, based on this understanding, they come up with a plan of care. Second, the clinicians try to bring services to the patient instead of bringing the patient to the services. That means diagnosing and treating problems in the adult day health center or at home.
If this sounds familiar, it is the prescription I wrote about in my recent article, “When Frail Elderly Patients Get Sick: Alternatives to Hospitalization” and have blogged about in these pages. If PACE has the master key to the problem that has bedeviled American healthcare—how best to care for its sickest, most vulnerable patients—then maybe it’s time to fashion other similar keys, creating models that will appeal to a frail elders of many ethnic groups, varying socioeconomic status, and in many geographic regions.
I just read a report describing the state of palliative care around the world that made me appreciate how far we have come in the US. Here, about 45% of people who die were receiving hospice care at the time of death; worldwide, the figure is 14%.
We still have an enormous amount of work to do at home. Many of those who die with hospice services were enrolled in the program for only a few days, giving them little chance to benefit; and palliative care, a more comprehensive approach than hospice, (hospice is focused on the last six months of life and is generally limited to comfort care), is available in only two-thirds of American hospitals. But according to the recently released “Global Atlas of Palliative Care,” the majority of countries in the world have no palliative care services at all.
The report endorses the World Health Organization definition of palliative care as an approach to patients with life-threatening illness and their families that focuses on improving quality of life. It supports the idea that palliative care should be a basic human right for the 20 million people who die each year with medical conditions such as chronic heart disease, cancer, and chronic lung disease. By treating pain and other physical symptoms and by addressing psychosocial and spiritual issues, palliative care improves the final phase of life in a compassionate and cost-effective way. And since nearly 70% of the 20 million deaths are in older individuals, the issue is primarily although not exclusively a geriatric concern.
What struck me about the report is that one of the major barriers to patients’ receiving palliative care is restrictions on opioid prescribing (morphine and its cousins): 80% of the world’s population lack adequate access to opioid medications for pain control. In fact, the US, Canada, Australia, New Zealand and a few other western countries account for 90% of the world’s use of these medications that are so essential to good care near the end of life.
At the same time, we read in the newspapers about the epidemic of deaths from opioid overdose. The death of actor Philip Seymour Hoffmann with a needle in his arm is just the most recent example. In 2011, a landmark study found that opioid overdose is the second most common cause of unintentional death in the US, close behind motor vehicle accidents: 74% of the unintentional deaths involving medications were related to opioids.
How are we to reconcile the need for more palliative care, which will require greater opioid use, and the rising rate of morphine-related deaths in those parts of the world where morphine is widely available? First, it’s important to realize that most, probably almost all, of the overdoses occur in patients who are not receiving palliative care or hospice. In one study of the association between opioid prescribing patterns and opioid-overdose-related deaths, only 2 patients were enrolled in a palliative care or hospice program at the time of the overdose. While the cause of death might have been misclassified in some of the palliative care patients, the evidence so far suggests that overdoses are not common in this population. Second, the best way to address opioid overdose involves a variety of strategies other than limiting use in situations of need. Many states have recently introduced regulation to prevent “diversion” of narcotics (funneling of prescription drugs from the intended recipient to a drug addict). Another promising approach involves providing narcan, an opioid antagonist, to members of the community where people at risk of overdosing live. When such a program was introduced in North Carolina, the result was a drop in the proportion of opioid deaths due to prescription drugs from 82% to 10% in 2 years.
I would regret to see the many older patients with pain or shortness of breath (another symptom that responds to morphine) refuse opioids because they are afraid of their potential for abuse. Yes, we have a problem with the abuse of opioids in the US. But patients in the final stage of life who are receiving palliative care are not the problem.
When I was a medical resident at Boston City Hospital, a large, public, inner city hospital, I began wondering whether hospitals sometimes caused as many problems as they cured. Over and over, I saw older patients admitted with one disease such as pneumonia or a heart attack, who ended up falling and breaking a bone or getting a blood clot from being confined to bed. So I did a study in which I measured how often older people became confused, stopped eating, developed incontinence, or fell while they were in the hospital. I tried to separate out those cases in which the new symptom could be plausibly related to the admitting diagnosis: for example, someone who was hospitalized with a stomach ulcer would normally stop eating during the initial treatment, and someone with a stroke might well be confused. Then I compared the rates at which people over 70 developed these unexpected complications with the rates at which younger people developed them. Finally, I speculated that each of these problems might trigger a cascade of adverse events: a patient who became incontinent might have a catheter placed in his bladder, which in turn might cause a urinary tract infection; a patient who got confused might be physically restrained and his immobility might lead to a blood clot.
What I found was that among the 502 patients I examined, a startling 41% of those over 70 developed 1 or more of the 4 problems I was interested in compared to only 9% of the younger group—and these were all problems that could not clearly be related to the acute illness for which the patient was being treated. It made me question whether hospitals were a safe place for older patients.
Of course the older patients might have developed the same difficulties if they had been cared for at home, but I was skeptical. I thought that the unfamiliar environment of the hospital might be causing the confusion so common in hospitalized older individuals. I believed that being in a strange place might lead people to have trouble finding the bathroom at night and to fall as they groped trying to find their way. And certainly the response of doctors and nurses to these new symptoms—putting in a catheter or using physical restraints or sedating medications—was far less likely to occur in the home setting.
Stimulated by my study and others like it, some physicians tried to design safer hospitals. That’s a noble endeavor and the result, the ACE unit (acute care for elders) has made it somewhat less likely that patients develop the kinds of complications I enumerated and far less likely that doctors and nurses responded to those complications, when they did occur, in the unfortunate ways I wrote about. But desite these improvements, older hospitalized patients have an unacceptably high risk of becoming acutely confused, receiving sedating medications, and falling, among other problems.
At the same time that older patients continue to do poorly in the hospital—particularly the oldest and the frailest of the old—they have very high rates of repeat admissions to the hospital, sometimes within 30 days of discharge. As a consequence, many frail elderly patients have not just one opportunity to develop complications in a short period of time, but several chances. As it turns out, Medicare is very concerned about the high rate of readmissions, not so much because the Medicare program recognizes that hospitals can be hazardous for your health, but because repeated hospitalizations are very expensive. Not only that, but Medicare officials suspect that the reason older patients are being readmitted is that they weren’t properly taken care of the first time round. So the solution, Medicare reasons, is to penalize hospitals for high readmission rates, thus strongly incentivizing them to assure they get things right during the initial hospitalization.
It is a nice idea, and there is quite a bit hospitals can do to prevent readmissions. They can make sure patients are taking the medications they are supposed to take once they get home. They can check that they have a follow up appointment with their primary care physician within a few days of discharge. They can guarantee that the primary care physician receives a good summary of what happened in the hospital. But the assumption underlying the push to decrease readmission rates is that if only doctors do a good job with their frail, old patients, those patients won’t get sick again. The truth is that frail, old patients will inevitably get sick, even if physicians and nurses and hospitals do the best possible job taking care of those patients. The reason for this sad reality is that the essence of frailty is a heightened vulnerability to illness. So frail people will, by definition, repeatedly become sick.
Which brings me full circle to the study I did in the 1980s about the adverse consequences of hospitalization in the elderly. The question, as I posed back when I wrote up my findings many years ago and again in an article I published in the Annals of Internal Medicine this week, is how we should respond when frail elders get sick. And my answer, then as now, is that we should try very hard to keep them out of the hospital. That means developing viable alternatives to hospital treatment. Frail, old people will sometimes want to be hospitalized for their acute medical problems. The hospital may prove to be the best site for achieving certain goals of care. But if they are interested in maximizing their quality of life, as many frail elders are, and if they have the option of home rather than hospital care, many of them would choose that route.
The secret that policymakers do not seem to have recognized is that doing all these good things could save Medicare billions of dollars. Hospital care makes up 25% of Medicare spending and frail elders account for a disproportionate share of hospital days. So if we want to avoid hospital-induced iatrogenesis and assure that treatment conforms to patients’ goals and save money, we need to design new ways to treat patients outside the acute care hospital.
Recently, a friend commented that she was not sure whether or not to agree to a “DNR order” for her 90 year-old mother. Her mother has dementia and lives in a nursing home; she is her mom’s health care proxy. Complicating her decision was the knowledge that her mother had chosen a DNR status when she was cognitively intact, but then reversed her decision at the time of acute illness, believing that “DNR” meant she would not receive vigorous treatment for medical problems such as an infection or congestive heart failure. My friend told me she believed the purpose of a DNR order was to avoid a protracted, painful death and that DNR was synonymous with comfort care.
The truth is that DNR means “do not perform CPR”—the “R” in DNR refers to resuscitation, which is shorthand for “cardiopulmonary resuscitation.” CPR is a procedure intended to counteract a cardiac arrest—the sudden cessation of the heart beat and of breathing. CPR involves administering electric shocks and, more often than not, providing artificial respiration, usually through a breathing machine (ventilator). In principle, DNR or, as it should more properly be called, since there is only a small chance that CPR will actually succeed in restoring cardiac function, DNAR (do not attempt resuscitation) refers only to what will or will not be done if the heart stops beating and breathing ceases. It has no implication for any other form of treatment. But nearly 40 years after the first DNR orders were written in hospitals, confusion about their meaning persists.
It’s not just patients and families who are confused: physicians are confused as well. I just read Sheri Fink’s book, Five Days at Memorial: Life and Death in a Storm-Ravaged Hospital, in which the triage decisions made during Hurricane Katrina included placing patients with a DNR order at the bottom of the evacuation list. Reportedly, physicians at Memorial thought that patients could have a DNR order only if they were terminally ill, which is false: a patient can elect a DNR order whenever the burdens of CPR are felt to outweigh the benefits, or if CPR is thought to be extremely unlikely to succeed--regardless of the underlying medical condition. When it looked as though not all of Memorial Hospital’s patients were going to be evacuated, it was the DNR cases who were to be left behind, sedated so as to ease their allegedly inevitable and imminent deaths.
And it’s not just in the midst of a disaster that physicians mistake “DNR” for “Do Not Treat.” Study after study has shown that physicians say they would not administer a whole variety of treatments to patients who are DNR. One representative study of 241 physicians found that they were far less likely to agree to transfer a patient to the intensive care unit or even to perform simple tests such as drawing blood.
Most recently, physicians and nurses caring for pediatric patients also told interviewers that in practice, DNR means far more than just do not perform CPR. In this survey of 107 pediatricians and 159 pediatric nurses in a hospital setting, 67% believed a DNR order only applies to what to do after a cardiac arrest—but 33% said it implied other limitations. And 52% said that once a DNR order is in place, a whole host of diagnostic and therapeutic interventions should be withdrawn, over and beyond CPR, and a small but disturbing minority, 6%, said that a DNR order means that comfort measures only are to be provided.
Why, so many years after the concept of DNR was introduced, does it remain so problematic? I think both patients and physicians implicitly recognize that asking about CPR is simply the wrong question—that’s why they assume DNR refers to so much more than what doctors should do if the heart stops beating. The right question has to do with figuring out what it is that medical treatment is supposed to achieve. Is it prolonging life at any cost? Is it providing comfort and dignity? Or is it something in between, preserving those basic functions such as walking and talking and seeing and hearing that a person values most? Only after defining the overall goal of care does it make sense to focus on delineating the specific treatments that are most consistent with achieving those goals.
An enormous amount of energy and ink have been squandered on debating about cardiopulmonary resuscitation, one of many possible medical treatments that seeks to achieve one of several possible medical goals. Maybe it’s time to pull the plug on DNR orders, banning them unless there is first a conversation about the overall direction of medical care.
One approach to doing exactly this is POLST, Physician Orders for Life-Sustaining Treatment or, as it is called in some states, Medical Orders for the Scope of Treatment. This is a medical order, just like an order a doctor writes in a hospital for antibiotics or for blood tests; it is not just a wish statement about future care. It includes a list of specific interventions that a prospective patient has agreed are undesirable under any circumstances, and often includes CPR along with other treatments such as dialysis or artificial nutrition. But before a physician and patient agree on such an order, a conversation is supposed to take place about the goals of medical care. The problem is that it's far from clear that patients and doctors are having those discussions. They may simply be jumping to the final step, writing that order. POLST is a promising development, but only if at least as much time is spent on talking about the big picture as on its implementation.
