The provision of nutrition and hydra tion through intravenous (IV) catheters or gastric intubation devices (e.g., nasogastric [NG], gastrostomy, jejunostomy tubes) has become commonplace. Such interventions are routinely used not only in acute care settings, but in long-term care and home care. Howard, Ament, Fleming, Shike, and Steiger (1995) examined Medicare data and reported that by 1 992, there were approximately 192,000 people who received artificial nutrition and hydration (either parenteral or enterai) in the home setting. The estimated costs for all nutritional support in the non-hospital setting (i.e., home, nursing home) were $1.6 billion (not including such costs as required nursing care and laboratory tests). According to data from the North American Home Parenteral and Enterai Nutrition (HPEN) Patient Registry, usage in the older than 65 age group tripled from 1978 to 1988. The United States has a 4 to 10 fold greater use of HPEN than other countries of comparable medical sophistication (Howard et al, 1995).
As Western society has become more comfortable with the moral acceptability of withholding aggressive medical interventions that are either unwanted by patients or serve only to prolong the dying process, such as cardiopulmonary resuscitation or mechanical ventilation, questions about the permissibility of withholding food and fluids have also arisen. While any treatment limitation decision can prompt doubt, ambivalence, and disagreement among patients, their families, and the professionals caring for them, questions related to nutrition and hydration have been particularly problematic.
Several factors have been recognized as contributing to increased difficulty in making decisions related to feeding and hydrating patients. Callahan (1983) pointed out the social significance attached to provision of nourishment for vulnerable and dependent others and argued that allowing withdrawal of food and fluid might weaken this important social norm. Although the traditional distinction between "ordinary" and "extraordinary" treatments is now understood to have no moral significance, some still argue that meeting nutritional needs of patients is basic, ordinary care and should never be foregone (Lo & Dornbrand, 1989). Last, provision of food and fluid is sometimes viewed as so different from technologically complex interventions such as dialysis or mechanical ventilation that it should not be considered a medical intervention at all (Veatch, 1993).
The increasing frequency of use of artificial nutrition and hydration, coupled with the societal ambivalence about these treatments, has led to considerable discussion and debate. Courts have repeatedly judged the provision of nutrition and hydration through artificial means to be a medical treatment and thus governed by the same considerations directing other forms of treatment limitation (Meisel, 1992). Professional associations, including the American Nurses Association (1988), have also issued position statements outlining the justification for withholding or withdrawing artificial nutrition and hydration in some circumstances. Nevertheless, decisions about whether or not to institute, continue, discontinue, or withhold artificial nutrition and hydration continue to present difficult challenges to patients, families, and clinicians.
The purpose of this article is to review the decision-making process, with particular attention to the factors that make decisions about feeding so difficult. Ethical principles and empirical data about the outcomes of artificial nutrition and hydration are discussed and recommendations for decision-making are offered.
THE ETHICAL PROBLEM
There is strong consensus about the general ethical principles that should direct decisions about treatment withholding and withdrawal. The principle of autonomy requires respect for the informed choices of competent individuals regarding any proposed intervention. In addition, professionals are guided by the principle of beneficence, imposing a duty to help, benefit, or protect all under their care. While these two foundational principles provide clear and solid direction in many situations, they are not sufficient by themselves. As noted by others (Meyers & Grodin, 1991), feeding decisions are usually problematic precisely because individuals for whom these questions arise most often are no longer able to express wishes and because it is not clear what constitutes "good" or best interest, particularly in states of severe dementia. When individuals do not currently appear to be physically or emotionally suffering, yet are in a state of profound cognitive impairment and family members are certain they would not have wanted to exist like this, questions about the appropriate standard for judging "best interest" arise. Therefore, to understand and address this complex issue, it may be helpful to look at the elements of sound ethical decision-making in more detail.
Questions about instituting artificial feeding can arise in many situations. When patients are able to make their wishes known, when the need for supplemental nutrition and hydration is related to an acute event and expected to be only temporary, or when death is expected to occur imminently (within days), decision- making is much easier. Situations in which patient wishes are unknown, when death may not occur for weeks, months, or even longer, and when patient capacity for pleasure or pain are uncertain challenge health care workers to be most careful in their reasoning and their approach to families.
The usual approach in these situations is to analyze benefits and burdens of either providing artificial feeding or withholding it. In general, analyses of this sort have two components - the medical and the moral - each of which informs the other. That is, some potential treatments, such as dialysis or mechanical ventilation, are clearly effective in addressing physiologic needs or problems. Even in these cases, decisions about using the intervention is informed by ethical considerations, particularly the right of competent individuals to accept or refuse virtually any treatment. In other cases, the ethical factors are clear (e.g., the patient wants to survive and requests treatment), but the decision of whether to use the treatment is importantly directed by the medical factors. For example, the decision of whether to perform surgery on the patient with coronary artery disease rests on the particulars of the biomédical context, such as whether his lesion is amenable to surgery.
This decision process is an iterative one in which one asks several questions:
* What are the problems?
* What are the options?
* Which options are permissible?
* Of the permissible options, which are desired?
* Of the desired options, which are possible?
* What are the benefits and burdens of each?
In conducting this evaluative process, one considers both the biomedical facts (which should be understood as probabilities) and personal, subjective elements (e.g., preferences, quality of life, family interests). Decisions regarding artificial nutrition and hydration are often particularly problematic because of the uncertainty regarding the biomédical factors or the individual subjective elements, or, in some cases, because there is misunderstanding of these factors by some participants. Because of this, it is useful to begin with what is known about the usual benefits and burdens associated with artificial nutrition and hydration.
BENEFITS AND BURDENS
The most obvious expected benefit of providing artificial nutrition and hydration is survival. Clearly patients who cannot voluntarily ingest any food or fluids orally will not survive unless these nutrients are provided either parenterally or enterally. In temporary situations in which voluntary intake is nil, such as post-trauma or in the initial period following stroke, provision of artificial nutrition and hydration thus will reliably lead to continued life.
POSSIBLE MECHANISMS EXPLAINING THE ABSENCE OF SUFFERING IN THE WITHHOLDING OF NUTRITION AND HYDRATION'
In situations in which the individual is capable of some oral intake, but in inadequate amounts, in the context of protracted terminal illness, and in advanced dementia, the benefits are less certain. Mitchell, Kiely, and Lipsitz (1997) analyzed data on 1,386 nursing home residents with severe cognitive impairment in the state of Washington. These data were abstracted from the federal Minimum Data Set. Of these individuals, 135 (9.7%) underwent placement of a feeding tube. The authors performed a survival analysis comparing residents with and without feeding tubes during a 582 day period. Feeding tube placement was not associated with survival, even after adjusting the model for independent risk factors of feeding tube placement such as aspiration, stroke, pressure ulcers, and absence of do not resuscitate (DNR) orders. Thus the data did not confirm the assumption that death without feeding would be more probable or certain than if artificial feeding and hydration were administered. The most important predictors of poor survival were the residents' medical conditions, advanced age, and DNR status.
In addition to survival, artificial nutrition is often provided to prevent physical deterioration, particularly muscle and skin breakdown. While inadequate nutritional intake can lead to muscle wasting and skin disorders in individuals who are otherwise healthy, the effects of supplemental nutrition in elderly individuals and in terminal illness are less clear. Finucane (1995) reviewed studies conducted between 1985 and 1994 examining the relationship between malnutrition, tube feedings, and pressure sores and found conflicting results among the studies. There was no consistent evidence that standard measures of nutritional status (e.g., serum albumin, total lymphocyte count, weight) correlated with pressure sore outcomes or that tube feedings reliably led to improved pressure sore outcomes. He concluded that routine use of tube feedings to prevent or treat pressure sores is not supported by data. Specifically, it is not known if poor outcomes among tube-fed patients reflects the effects of underlying disease condition, overall lack of efficacy of tube feedings, or simply disadvantages that outweigh advantages. As frequently noted in discussion of this topic (Howard et al., 1995; Mitchell et aí., 1997), he identified the limitation of previous reports that did not control for confounding variables, such as underlying diagnosis or route of administration of feedings.
A third intended benefit of artificial nutrition and hydration, particularly through use of percutaneous enterostomal gastrostomy (PEG) tube or jejunostomy tube, is lessening of the risk of aspiration. As with other aspects of this issue, there are no controlled trials providing conclusive evidence about the efficacy of either NG or gastrostomy feedings compared to oral feedings in patients with demonstrated risk of aspiration. However, there is considerable data that show that NG, gastrostomy, and jejunostomy tubes do not eliminate the risk of aspiration and resultant pneumonia. Cicone, Silverstone, Graver, and Foley (1988) studied 70 tube-fed patients in a skilled nursing facility. All patients were older than age 65. In addition to the problem of self-extubation, aspiration pneumonia was common among both patients with NG tubes (44%) and gastrostomy tubes (56%). Cogen and Weinryb (1989) and Cogen, Weinryb, Pomerantz, and Fenstemacher (1991) found a similar incidence with both gastrostomy and jejunostomy tubes in their studies of nursing home patients. In addition to finding that patients who had previously aspirated continued to do so after jejunostomy, Cogen, Weinrub, Pomerantz, and Fenstemacher (1991) also reported that 36 of the 44 patients (82%) studied experienced some complication associated with jejunostomy, including two deaths related to the tube feedings.
While prolonged survival, preservation of muscle and skin integrity, and protection from aspiration are all expected and intended benefits of tube feedings, data suggest that these benefits among individuals who are terminally ill or in the end stages of dementia are possible, but by no means likely. It therefore is particularly important to carefully consider the probable burdens of either providing or withholding artificial feeding.
The strongest concern about withholding artificial nutrition and hydration is worry that this will cause suffering. In any analysis of the possible burdens of a proposed treatment plan, certainly the aim of avoiding or minimizing discomfort is fundamental.
Both families and clinicians assume going without food and fluid will result in significant suffering. This seems to be a reasonable belief based on personal experiences of hunger and thirst. However, as with the assumption that provision of nutrition will increase survival and promote skin integrity, there is considerable data that suggest the opposite - that foregoing nutrition and hydration in terminal illness does not cause suffering and may even promote comfort (Table 1).
For many years, clinicians caring for dying patients have observed and reported anecdotally that terminally ill patients rarely seem to experience hunger, often are anorexic, and that patients who do receive either tube feedings or IV fluids seem to have more problems with nausea, vomiting, and managing secretions (Koshuta, Schmitz, & Lynn, 1991; Sullivan, 1993). Printz (1992), in a discussion of the appropriateness of allowing terminal dehydration in some situations, offered several explanations for why dehydration progressing to death is not uncomfortable. Food and water deprivation has been shown in laboratory animals to be associated with an increased production of endorphins, referred to as endogenous opioids, which exert analgesic effects. Ketones, produced in starvation states, and their metabolic products, also may yield analgesia. Billings (1985) noted that the dehydration occurring in terminally ill patients who are not rehydrated is physiologically different, with different symptoms and sequelae, than either hyponatremic or hypernatremic dehydration occurring when gastrointestinal losses are inadequately replaced, when water loss exceeds sodium loss, or sodium loss exceeds water loss as in pituitary dysfunction. In contrast to these primary electrolyte disturbances, dehydration associated with the decision to withhold artificial hydration in the terminally ill is less likely to be accompanied by any uncomfortable or distressing symptoms other than dry mouth.
McCann, Hall, and GrothJuncker (1994) confirmed this in their prospective study of 32 competent, terminally ill patients who chose not to receive artificial hydration and nutrition during their stay in a "comfort care unit." All of the patients except one wanted only occasional small amounts of liquids or solids from the time of admission until their deaths, which occurred, on average, 40 days after admission. Of the 32 patients, 20 (63%) never experienced hunger, while 1 1 (34%) reported hunger only initially. Eleven (34%) experienced no thirst, 9 (28%) felt thirsty or complained of dry mouth initially, and 12 (38%) experienced thirst or dry mouth until the time of their death. In all patients, mouth care, ice chips, or sips of water or food relieved their symptoms. Most patients (94%) were receiving narcotics for relief of pain or shortness of breath. The authors concluded that iack of food and fluids sufficient to meet basal metabolic needs does not cause suffering as long as mouth care and sips of water as requested were provided. In fact, 9 patients who ate to please their families reported abdominal discomfort and nausea after eating. They suggest, similar to Printz (1992), that conversion to fat metabolism, rather than glucose, for energy in starvation states leads to increased water production and ketosis, and thus spares muscle breakdown. Anorexia, partie - ularly in patients with cancer, may be the result of changes in taste threshold or reduced hypothalamic stimulation associated with changes in gastric contraction and secretion. Providing small amounts of carbohydrates, which prevents ketosis and stimulates gastric secretion, may actually increase discomfort. In the absence of food, the urea load presented to the kidney is reduced, with resultant reduction in water excretion. The symptoms of dehydration may therefore be less than expected as fluid requirements are decreased. It should be noted that most of the empirical data substantiating the claim that foregoing nutrition and hydration in end-stage disease does not cause suffering were not gathered from studies of patients whose only diagnosis was severe dementia. However, most of the physiologic mechanisms posited are not connected to pathology associated with particular diagnoses, such as cancer, and thus can be reasonably applied to patients with end-stage dementia.
One of the most serious indirect disadvantages of providing artificial nutrition and hydration to patients who are cognitively impaired is the associated use of restraints. Patients who are confused and disoriented often pull at tubes and may attempt to dislodge IV lines, NG tubes, and even PEG tubes. To prevent this, patients are either physically or chemically restrained. In Quill's (1989) study of 55 elderly patients in a community hospital who received NG feeding, 53% were restrained. The use of restraints can significantly diminish the quality of life in any situation, and is a particularly egregious offense in the last days of life.
Given these facts, it is evident that the burdens of artificial nutrition and hydration can and do outweigh the benefits in many cases. However, the real assessment of benefits and burdens can only meaningfully be performed in the context of individual patient goals. Many conventional medical interventions have a low probability of physiologic benefit and are associated with enormous costs and burdens, such as chemotherapy for advanced cancer or multiple-organ transplants, yet patients elect these therapies because the potential benefits are subjectively evaluated as outweighing any burden. Often, in acute care, the potential benefit is survival, but sometimes it is just the meaning attributed to aggressively fighting disease. Decisions about tube feedings for elderly individuals are subject to the same influence of very personal values and interpretation of the meaning of accepting or refusing intervention.
Ample evidence exists that patients and families, when given the choice, express widely varying opinions regarding the desirability of tube feedings in states of cognitive impairment. Mezey, Kluger, Maislin, and Mittelman ( 1 996) interviewed 50 spouse caregivers of patients with Alzheimer's disease and reported 48% of the spouses would consent to a feeding tube if their spouses became critically ill, although only 14% would consent in the situation of irreversible coma. O'Brien et al. (1995), in a multi-site study of 421 cognitivelyintact nursing home residents, found that 33% would wish to receive tube feedings if they were unable to eat because of brain damage. Ouslander, Tymchuk, and Krynski (1993) found the same variation in attitudes in their study of 68 elderly individuals from a day care center and a residential care unit, with 50% of respondents indicating they would accept enterai feedings if they were severely ill and unable to eat.
CRITERIA OF A GOOD DECISION PROCESS
The decision to feed is complex because the biomédical facts are not conclusive and because the standard or test of "best interest" is dependent on individual values and the particulars of each situation. There is no one decision, treatment plan, or even outcome identified as optimal in all situations. Process thus becomes critically important.
It may be helpful, before discussing specific approaches to this problem, to consider the framework for a decision. There is no one right answer to the question of whether terminally ill elderly patients, including those with end-stage dementia, should receive artificial nutrition and hydration. The challenge, then, is to devise approaches to maximize the probability of achieving a satisfactory outcome in each individual case. That is, health care providers seek a procedure that will provide reasonable assurance that the outcome of the procedure will meet the criteria of a good, moral, or fair decision. Focusing on the procedure, rather than the outcome, is especially helpful when there are deep disagreements about what the content of the decision (i.e., to administer or to withhold feedings) should be.
What would be the criteria of best process? There are some standards on which there is clear agreement and which should serve as guiding principles. These are outlined in Table 2 and discussed below.
First, bad information leads to bad decisions. For example, placing a PEG tube only to prevent aspiration would be a poor decision in light of the data that demonstrates a continuing high risk of aspiration after PEG placement. Withholding NG feedings from a severely demented individual because of an assumption that no one would want life maintained in this condition would also be a bad decision, not because the decision itself is categorically wrong, but because the justification for the decision is inadequate.
In particular, given the demonstrated plurality of values evident among individuals, one must make every effort to determine what the patient's values have been and search for any clue related to what the patient's preferences would be in the given situation. Arras (1988) refers to this as "probing the patient's subjectivity" (p. 938). Both previous wishes and evidence of current sensation, such as evidence of ability to experience pleasure and pain, must be assessed.
Second, the benefits and burdens of artificial nutrition and hydration can only be evaluated in light of both short- and long-term goals. Patients who are terminally ill and unable to eat because of physical problems such as swallowing dysfunction or esophageal obstruction may have a short-term goal of surviving until an important event, such as the birth of a grandchild. Providing needed food and fluids to reach this goal is both justified and necessary. In contrast, if the goal is to promote a peaceful death, which is expected to occur quite soon, insisting on the use of IVs or tube feeding is not only unjustified, but probably harmful.
Meeting this criterion requires explicit discussion of the goals of care. While seemingly obvious, this is an often neglected aspect of care planning. Both clinicians and family members may be uncomfortable expressing the view that promoting a peaceful death can be a reasonable goal. It may be helpful to begin a conversation with a clear, straightforward discussion introducing the subject by reviewing the patient's current condition. Clinicians may explain that some patients prefer nothing be done to forestall death and instead focus on comfort and dignity, while others wish to receive any measure offering a chance of supporting survival.
Third, the interests of all should be represented. There are several reasons for this. As mentioned earlier, withholding food and fluids is a topic likely to evoke strong feelings and beliefs, both on the part of patients and families and on the part of professional and non-professional staff. Family members obviously have a stake in these decisions, but so do the nurse aides who provide care to patients in hospitals and skilled nursing facilities. All parties should have an opportunity to discuss the proposed plan of care and voice an opinion.
In making the actual decision about feeding, the patient's best interest is the touchstone. In situations in which patients themselves cannot explain what would constitute "best interest," it is generally best to solicit as much information and as many opinions as possible from those who have known the patients. Including everyone in decisions does not guarantee the interests of vulnerable patients will be well represented, but it makes it much less likely that possibly idiosyncratic values of any one individual will dominate. Therefore, even when there is an identified nextof-kin who has legal authority to make decisions, that individual should be encouraged to include other family members in discussion. Nurses may offer to attend family meetings or talk with others on the phone. The voice of those closest to the patient should, in most situations, be given greatest weight, but even in the clearest cases, the reassurance of having a consensus can be of enormous value to the designated decision-maker.
The last criterion of a good decision is that it does not violate the most strongly held beliefs of anyone, including society at large. While it may not always be possible to reach full agreement among all on this issue, particularly with large families, the presence of strong objections to a plan is reason for delaying implementation. Fortunately, the decision to either institute artificial feedings or to withhold them rarely needs to be made emergently. Continued discussion, education, and allowing time for reflection can lead to acceptance, if not actual consensus.
Decision-making regarding the issue of administration of artificial nutrition and hydration is complex in the best of circumstances. Impartial analysis of the justification for either providing or withholding the provision of artificial nutrition and hydration establishes neither that this intervention is categorically wrong nor right. Even in conditions of terminal illness, there may be good reasons for providing food and fluids and there may be good reasons for withholding food and fluids.
Situations in which individuals without cognitive impairment make fully informed decisions to forego food and fluids are relatively straightforward in terms of ethical mandates. Although the determination of decisional capacity may be complex, and designing effective support mechanisms for patients and families facing these decisions is always challenging, the principle directing us to respect competent treatment refusal is clear. Unfortunately, it is more common to face situations in which patients are no longer able to make treatment choices and have not left any advance directives to inform health care workers of their wishes.
Review of available data suggests several poults. Administration of food and fluid via NG, gastrostomy, or jejunostomy tubes in chronic or terminal illness in elderly individuals has not been shown to be reliably effective in increasing survival nor in reducing the incidence of aspiration or skin breakdown. Yet, there is consistent evidence that a sizable proportion of older adults and their spouses often feel obligated to provide such support when oral intake is inadequate. It is unknown whether this attitude is related to misconceptions or impervious to the evidence that withholding artificial nutrition and hydration does not cause suffering and may even promote comfort.
Given these facts, rather than assuming there are a priori "right" decisions in regard to the desirability of artificial feeding, clinicians should focus on the process of reaching consensus. This includes:
* Appropriate review of what is known about tube feedings.
* The medicai facts of each individual situation.
* Investigation of any available evidence that could yield understanding of what would constitute the best interest of the individual.
* Solicitation of the views of all stake-holders.
* Delay of action until some semblance of consensus and acceptance of the plan is reached.
Artificial nutrition and hydration, similar to most medical therapies, has some potential benefits but much, if not more, potential for causing harm. It should not be provided without a clear understanding of the goals of therapy and reasonable expectation that its provision will contribute to these goals.
- American Nurses Association. (1988). Guidelines on withdrawing or withholding food and fluid. In Ethics in nursing: Position statements and guidelines. Washington, DC: Author.
- Arras, J.D. (1988). The severely demented, minimally functional patient: An ethical analysis. Journal of the American Geriatrics Society, 36, 938-944.
- Billings, J.A. (1985). Comfort measures for the terminally ill. Journal of the American Geriatrics Society, 33, 808-811.
- Callahan, D. (1983). On feeding the dying. Hastings Center Report, 13(5), 22-27.
- Cicone, J.O., Silverstone, P.A., Graver, L.M., & Foley, CJ. (1988). Tube feedings in elderly patients: Indications, benefits, and complications. Archives of Internal Medicine, 148(2), 429-433.
- Cogen, R., & Weinryb, J. (1989). Aspiration pneumonia in nursing home patients fed via gastrostomy tubes. American Journal of Gastroenterology, 84(12), 1509-1512.
- Cogen, R, Weinryb, J., Pomerantz, 8e Fenstemacher, P. (1991). Complications of jejunostomy tube feeding m nursing facility patients. American Journal of Gastroenterology, 86(\\), 1610-1613.
- Council on Scientific Affairs and Council on Ethical and Judicial Affairs. (1990). Persistent vegetative state and the decision to withdraw and withhold life support. Journal of the American Medical Association, 263, 426-430.
- Finucane, TE. (1995). Malnutrition, tube feeding and pressure sores: Data are incomplete. Journal of American Geriatrics Society, 43, 447-451.
- Howard, L., Ameni, M., Fleming, C.R., Shike, M-, & Steiger, E. (1995). Current use and clinical outcome of home parenteral and enterai nutrition therapies in the United States. Gastroenterology, /09(3), 355-365.
- Kayser-Jones, J, Schell, E.S., Porter, C., Barbaccia, J.C., Steinbach, C., Bird, W.F., Redford, M., & Pengilly, K. (1998). A prospective study of the use of liquid oral dietary supplements in nursing homes. Journal of the American Geriatrics Society, 46, 1378-1386.
- Koshuta, M.A., Schmitz, P.J., & Lynn, J. (1991). Development of an institutional policy on artificial hydration and nutrition. Kennedy Institute of Ethics Journal, 1, 133-139.
- Lo, B., & Dornbrand, L. (1989). Understanding the benefits and burdens of tube feedings. Archives of Internal Medicine, /49(9), 19251926.
- McCann, R.M., Hall, WJ., & Groth-Juncker, A. (1994). Comfort care for terminally ill patients. Journal of the American Medical Association, 272, 1263-1266.
- Meisel, A. (1992). The legal consensus about foregoing life-sustaining treatment: Its status and its prospects. Kennedy Institute of Ethics Journal, 2, 309-346.
- Meyers, R.M., & Grodin, M.A. (1991). Decisionmaking regarding the initiation of tube feedings in the severely demented elderly: A review. Journal of the American Geriatrics Society, 39, 526-531.
- Mezey, M., Kluger, M., Maislin, G., Mittelman, M. (1996). Life sustaining treatment decisions by spouses of patients with Alzheimer's disease. Journal of the American Geriatric Society, 44, 144-150.
- Mitchell, S.L., Kiely, O.K., & Lipsitz, L.A. (1997). The risk factors and impact on survival of feeding tube placement in nursing home residents with severe cognitive impairment. Archives of Internal Medicine, /57(3), 327-332.
- O'Brien, L.A., Grisso, J.A., Maislin, G-, LaPann, K., Krotki, K.P., Greco, P., Siegert, E.A., Evans, L.K. (1995). Nursing home residents' preferences for life-sustaining treatments. JAMA, 274, 1775-1779
- Ouslander, J.G., Tymchuk, A.J., & Krynski, M.D. (1993). Decisions about eternal tube feeding among the elderly. Journal of the American Geriatrics Society, 41, 70-77.
- Printz, L. A. (1992). Terminal dehydration, a compassionate treatment. Archives of Internal Medicine, /52(4), 697-700.
- Quill, T.E. (1989). Utilization of nasogastric feeding tubes in a group of chronically ill, elderly patients in a community hospital. Archives of Internal Medicine, 149(9), 19371944.
- Sullivan, R.J., Jr. (1993). Accepting death without artificial nutrition or hydration. Journal of General Internal Medicine, 8(4), 220-224.
- Veatch, R.M. (1993). Foregoing life-sustaining treatment: Limits to the consensus. Kennedy Institute of Ethics Journal, 3, 1-20.
POSSIBLE MECHANISMS EXPLAINING THE ABSENCE OF SUFFERING IN THE WITHHOLDING OF NUTRITION AND HYDRATION'
CRITERIA OF A GOOD DECISION PROCESS