A Success Story

In a research letter published this week in JAMA, geriatrician Susan Mitchell presents some startling—and encouraging—data. She reports that between 2000 and 2014, the rate of feeding tube insertion in US nursing home residents with advanced dementia fell from 11.7 percent to 5.7 percent.

It’s an achievement to be proud of—and even though the rate of feeding tube use remains high in blacks, the extent of the drop over the last 15 years is if anything more dramatic in the black population: for whites in nursing homes who developed advanced dementia and feeding difficulty, the rate went from 8.6 percent to 3.1 percent; for blacks in went from 37.5 percent to 17.5 percent. How and why did this dramatic change occur? 

Much of the credit goes to the work of Dr. Mitchell herself, who has tirelessly studied the use of feeding tubes in this population since the ate 1990s. Her initial foray into this field was a study with the somewhat inauspicious title, "The Risk Factors and Impact on Survival of Feeding Tube Placement in Nursing Home Residents with severe Cognitive Impairment." Published in what back in 1997 was called the Archives of Internal Medicine, it provided the first compelling evidence that, contrary to common sense expectations, performing a surgical procedure to provide ongoing artificial nutrition to nursing home residents who, in the course of developing advanced cognitive impairment, had stopped eating, did not prolong life. It wasn't a randomized controlled trial and it was confined to nursing home residents in the state of Washington, but it was a carefully conducted analysis of a reliable data base.

When the findings were confirmed in subsequent studies, Dr. Mitchell went on to examine the nursing home factors associated with tube feeding (on-site speech therapy, low aide to patient ratios, and evidence of poor quality care) and the clinical and organizational factors associated with tube feeding (for profit status, large size, and urban location). She looked at the financial incentives for feeding tube use (Medicare pays more) and the cultural factors associated with feeding tubes (facilities with a home like environment, a focus on food, specially trained aides, and an emphasis on advance care planning were much less likely to use them).  And she established that the development of difficulty chewing and swallowing is par for the course as dementia becomes very advanced. Many other investigators contributed to the subject as well: a search on Google Scholar using the phrases “artificial nutrition” and “advanced dementia” today produced 1340 hits.

But evidence does not always change practice, especially in an area as emotionally charged as the use of food and drink. During the same period that tube feeding use was falling in advanced dementia, we witnessed the spectacle of Terri Schiavo, a young woman in a persistent vegetative state, whose feeding tube was repeatedly inserted and withdrawn for seven years as her husband (her official surrogate) and her parents battled in the courts. What else accounts for the change in behavior and what can we learn from this experience about how to influence health care policy?

I suspect three factors played a role in moving the medical profession, hospitals, and the public towards a growing acceptance of palliative care--and away from technological interventions such as feeding tubes-- for individuals with advanced dementia.

First, physicians who cared for individuals with dementia were increasingly disturbed by the suffering that they saw inflicted on their patients from the growing acceptance of technological interventions. It wasn’t just feeding tubes; it was dialysis and ventilators and ICU care. Just as CPR had spread from use in otherwise healthy individuals who suffered an acute myocardial infarction, complicated by a ventricular arrhythmia, to all dying patients, so too did other invasive technologies proliferate due to “indications creep.” But it was bad enough to do something unpleasant to a dying patient who understood the intention behind the procedure or test; it was worse to do something painful or frightening to a dying patient who was by definition unable to understand its purpose (or intended purpose). As patient involvement in decision-making became more common and competent patients had the option of choosing or rejecting Hail Mary treatment, the plight of the demented patient became more poignant. Several of us wrote articles arguing that from an ethical perspective, there was no requirement to provide patients in the final phase of life with artificial nutrition and hydration. In fact, it was arguably this growing malaise on the part of physicians with the widespread use of burdensome technology in patients with advanced dementia that led to the burgeoning research on its effects.

Second, the field of palliative care has taken off during the last fifteen years. While much of palliative care focuses on cancer patients and on patients who are imminently dying, the discipline is in principle concerned will all serious, life-threatening illness. The growing acceptability of palliative care—as evidenced, for example, by Medicare choosing to pay doctors and nurses for advance care planning meetings—has brought a recognition that what we customarily do to and for patients near the end of life is not always what they want or what is most beneficial to them. Responding to eating difficulties in a patient with advanced dementia by inserting a tube rather than by limited hand feeding is just another example of the tendencies that palliative care challenges.

Finally, forgoing a technological solution to a medical problem is attractive because it saves money. A gastrostomy tube in the setting of advanced dementia is an example of an intervention that manages to simultaneously be useless, burdensome, and costly. We need to find more examples of widely used treatments that meet all three criteria, such as renal dialysis in frail older people. Feeding tubes for people who are dying of dementia is just the beginning. 

Showing We Care

Since the 1990s, physicians and patients have been fighting over futility. The doctors look at a patient who is dying and say that further tests and treatment cannot possibly work and shouldn’t be done. The patients, or more commonly their families, look at those same patients and say that they want “everything done” to try to prolong life. 

As often happens in the US, the futility battle ended up in the court room. In the case of Helga Wanglie, an 86-year-old woman in a vegetative state after hip surgery, the doctors went to court over whether the patient's husband had the right to insist that she remain on a ventilator. The court, as also often happens, didn’t address the issue of whether the ventilator was or was not appropriate treatment for Mrs. Wanglie; it simply ruled that her husband, as her surrogate, had the right to make the decision. After that case, many physicians concluded that the fight over futility was itself futile. For the last 15 years, physicians have tried to focus on determining a patient’s goals of care and then suggesting what treatments are most consistent with those goals. When they still cannot agree with family members about the right course of action, they resort to mediation, sometimes provided by a hospital ethics committee. But simmering below the surface, conflicts over perceived futility rage bubble vigorously.

A short article in the New England Journal of Medicine, “The Debt of Life—Thai Lessons on a Process-Oriented Ethical Logic,” offers a refreshing way of looking at futility. Based on his experiences doing ethnographic field work in Thailand while a graduate student in Anthropology, physician Scott Stonington shines a new light on the typical ICU dilemma. The physicians, he reports, are loathe to perform various possible tests and treatments because they think in terms of outcomes. They argue that their interventions won’t work in the sense that they won’t overcome the existing medical problems and that they are burdensome to the patient and, parenthetically, expensive. The patient’s family, he observes, think in terms of the process of care. He comments on one Thai family who said that their father had given them “flesh, blood, and breath” so they had a “debt of life” to pay. The ICU, they reasoned, allowed them to repay their debt: it gave their father flesh (tube feedings for nutrition), blood (intravenous medications and dialysis to cleanse the blood), and breath (a ventilator for breathing). The family was not so much interested in the outcome of treatment as in the treatment itself. In this scenario, the conflict was ultimately resolved when the family came to the conclusion that they had paid their debt and further aggressive care could be discontinued.

I made a very similar argument in my essay, “The Standard of Caring: Why Do We Still Use Feeding Tubes in Patients with Advanced Dementia?” I noted that it had been over 10 years since a series of studies in the medical literature reported that feeding tubes (a tube inserted into the stomach to provide nutrition) did not prolong life in patients with advanced dementia who had eating difficulties. These patients are nearing the end of their lives and no matter what procedures they have, their prognosis remains pretty much the same. Not only don’t the tubes prolong life, but they don’t accomplish a variety of other goals that doctors had hoped they might: preventing pressure ulcers (skin breakdown that is often related to malnutrition) or preventing pneumonia (caused by food going into the lungs instead of the stomach). As a result of these studies, the rate of tube feeding people with advanced dementia has declined, but it is still far from zero. I suggest that the reason some families want a feeding tube is to show that they care. It’s not that they expect to improve some quantifiable outcome—living longer or avoiding pneumonia. It’s that they want to have a way to demonstrate caring. For the same reason, we keep people with advanced dementia clean and dressed. We don’t require a study that shows that they will be less likely to develop an infection if they are kept clean. We don’t demand proof that they will live longer if they are clothed. We assume that being clean and clothed contribute to well-being because they are among the only ways we as caregivers have of showing respect for the human being who happens to have dementia. Tube feeding, from this perspective, is a means of proving that we care. 

There’s an important conclusion to draw from the tube feeding example, a conclusion that applies to the ICU situation as well. If we want to dissuade families from advocating feeding tubes for their relatives with advanced dementia, we need to offer a viable alternative way to demonstrate caring. I suggested using special popsicles made by freezing high protein liquid supplements for patients who have trouble chewing and swallowing but can still suck. Hand feeding, laboriously spoon feeding someone who has trouble feeding himself, is an alternative for people who can still process food in this way. But simply telling families that we won’t feed their relative at all and trying to assure them that the person with advanced dementia will not experience hunger or thirst fails to offer any means of caring. In the ICU setting, perhaps what we need to do is not continue burdensome treatment until families feel their “debt is paid.” Perhaps instead what we need to do is to find genuine alternatives to painful or uncomfortable or undignified treatment. But unless we offer something rather than what families perceive as nothing, we will be stuck with providing what physicians regard as futile treatment. Hospice care is intended to serve this role, but may not offer enough active interventions to satisfy family members. Our challenge is to identify ways to truly show we care.

The Truth About Tube-Feeding

For the last 15 years, physicians have raised questions about the efficacy of artificial nutrition—tube-feeding—in individuals with advanced dementia. Study after study of the outcomes of feeding tubes in patients with advanced dementia has failed to show any measurable benefit. But despite the
mounting evidence, feeding tubes continue to be used. For the first time, a study has been published concluding that not only do feeding tubes fail to benefit people with advanced dementia, they actually harm them. In fact, they seem to cause one of the conditions they are intended to prevent, pressure ulcers. But is it true?

Advanced dementia is a progressive neurological condition in which there is gradual loss of all the most basic capacities, including the ability to speak, to use the bathroom independently, to walk, and to eat. Fully 86% of nursing home residents who develop advanced dementia also develop difficulty eating. Once the technology for artificial feeding became simple and widespread—a feeding tube is inserted into the stomach through a minor surgical procedure and allows liquid nutritional supplements to enter the body directly, bypassing the need for the person to chew or swallow food—the temptation to use it in anyone who couldn’t eat was irresistible, regardless of whether the cause of the problem eating was a temporary, post-surgical condition, a congenital abnormal of the gastrointestinal system or, as with dementia, an end-of-life condition. In particular, tube feeding was used in patients with advanced dementia even though it does not prevent aspiration pneumonia, a common cause of death in demented individuals in which food goes into the lungs instead of the stomach; it does not prevent pressure ulcers, another common condition in which skin breaks down due to immobility; and it does not prolong life. Tube feeding nonetheless continues to be used because families find nutrition of symbolic importance as a means of demonstrating caring,  because many physicians and families are skeptical of the clinical data, none of which are based on randomized, controlled studies, and in part because it is financially advantageous for nursing homes to tube feed demented residents than to hand feed them.

The trouble is that without a randomized trial, in which roughly half the patients are arbitrarily chosen to receive a feeding tube and the other half do not, it’s very difficult to be sure that whatever outcome is measured—whether it is pressure ulcer development or survival or anything else—is directly
related to the feeding tube. And physicians and families have been extremely reluctant to randomize patients with advanced dementia in this way to definitively answer the question. After all, the patients are profoundly cognitively impaired and unable to make decisions about whether to accept tube feeding. Their caregivers, who must make decisions on their behalf, tend to have strong feelings about tube feeding, as do their physicians. The current study attempts to circumvent the problem of bias due to non-randomization through several ingenious strategies: the patients with feeding tubes are “propensity-matched” to patients who did not get feeding tubes (the two groups are similar in terms of the major characteristics associated with the development of pressure ulcers); and patients must have recently developed advanced dementia to be enrolled in the study (the two groups are therefore similar in the stage of their disease based on standardized, federally mandated assessments performed every few months in all nursing homes).

Using a large database of just over 18,000 nursing home residents with advanced dementia, the authors of the new study were able to identify 1124 patients in whom a feeding tube was inserted and who had no pressure ulcers initially and match them with 2082 other nursing home residents who did not
receive a feeding tube and who also had no pressure ulcers. The two groups were comparable in their other medical conditions and in their risk factors for developing skin breakdown. Their six-month mortality rate was likewise similar, though it was slightly higher in the residents without feeding tubes (24.0%) than in those with feeding tubes (20.6%). The striking finding was that the risk of developing a new pressure ulcer (stage 2 or higher: ulcers are graded from stage 1, which is minimal skin erosion, to stage 4, in which the ulcer extends deep into the muscle or bone) was a shocking 35.6% in residents with gastrostomy tubes and only 19.8% in those without.

The new study also looked at whether feeding tubes could help heal existing pressure ulcers. Using an identical procedure, the authors identified 461 nursing home residents who had a pressure ulcer (stage 2 or higher) at the time the feeding tube was inserted and compared them to another similar group of 461 residents with a pressure ulcer but no feeding tube. Surprisingly, pressure ulcers improved in 34.6% of the residents without a gastrostomy tube, but in only 27.1% of those with the tube.

Prior studies have demonstrated that feeding tubes don’t help. But could it be true that feeding tubes actually make nursing home residents with advanced dementia worse off? The study’s authors try to explain their counterintuitive findings by suggesting that individuals with dementia who get a feeding tube may be physically restrained to prevent them from pulling out the tube and that the restraints impair mobility, predisposing to pressure ulcers. But this group had very limited mobility even before they had a feeding tube. The authors argue that tube feedings (the liquid dripped into the stomach through the gastrostomy tube) can cause diarrhea, which may accelerate pressure ulcer formation, but they provide no evidence that the patients in their study actually had diarrhea.

What I think is more likely, although I cannot prove this hypothesis, is that feeding tubes are just a marker for an aggressive approach to care and that it was that aggressive approach that led to the higher rate of pressure ulcers. After all, the patients in the Archives study were not randomized to receive a
gastrostomy tube. They got a tube because their families and their physicians believed that medical interventions were appropriate care for patients with advanced dementia. These medical interventions no doubt included a variety of strategies, of which feeding tubes were only one example. The current
study says nothing about what other interventions, aside from the feeding tube, the nursing home residents received after enrollment. But in a landmark study of the natural history of advanced dementia carried out by some of the same authors,fully 41% of nursing home residents with advanced dementia underwent at least one burdensome intervention in the last months of life: either a hospitalization, a visit to the emergency room, intravenous therapy, or a feeding tube. Undoubtedly, patients who were treated with one of those modalities were often treated with several of them. It seems plausible that it was the various aggressive medical treatments given to the nursing home residents with advanced dementia that caused them to develop new or non-healing pressure ulcers, not necessarily the tube feeding in isolation. It is even possible that tube feeding alone could have been beneficial, but it was typically administered in concert with a variety of other measures that produced harm.

So what can we conclude about tube feeding in advanced dementia? Without a randomized trial, we still do not know for sure what clinical outcomes, if any, tube feeding alters. Perhaps the time has come to focus on the indisputable fact that advanced dementia is a terminal condition and that only those measures that are inextricably associated with preserving the patient’s dignity and comfort, such as keeping the person warm and clean and clothed, are appropriate.