The Moral Implications of Medical Uncertainty: Tube Feeding Demented Patients

For more than a decade, a moral and legal debate has raged regarding the conditions under which it is permissible to withhold or withdraw tube feeding from patients unable to eat by mouth. One focus of this dispute is nondying, older patients with dementia. Some commentators have insisted that preservation of their lives is always a good to be protected. Tube feeding is viewed as basic, humane care whose routine use acknowledges the goodness inherent in the life of these patients, however cognitively disabled they may be. Other analysts have countered that the diminution of quality of life in progressive dementia is relevant to decisions to withhold tube feeding. When distinctive human capacities are lost — for self-direction and self-care, for experiencing pleasure, and for engaging in interpersonal relationships — life is no longer a benefit for the demented person. Under these circumstances, the use of tube feeding may be forgone. In constructing sophisticated moral and legal arguments, parties to the debate have made rather facile assumptions about the medical consequences of tube feeding. Tube feeding is represented as an unobtrusive, safe, and effective medical intervention for patients unable to eat by mouth. Although these assumptions permit a head-on philosophical clash between advocates for sanctity of life and quality of life views, they fail to reflect the clinical reality of tube feeding. The significance of the concise review by Sheiman and Pomerantz of recent studies on the medical consequences of tube feeding older demented patients is that it forces us to confront the possibility that tube feeding may impose on patients an excessive burden without offsetting benefits. Tube feeding may be initiated for various “medical indications”: to prevent aspiration and aspiration pneumonia, to forestall skin breakdown and pressure ulcers, to avoid dehydration and metabolic derangements, and to ameliorate the discomfort of hunger and thirst. The goals of therapy for these indications include prolongation of life, maintenance or improvement of life quality, or enhancement of patient comfort. However, several critical elements of medical uncertainty are suggested by the review. One problem is that the probability and magnitude of the benefits of tube feeding demented patients have not been clarified. For example, tube feeding is thought to lessen the frequency and severity of pressure ulcers. Yet two studies found that pressure ulcers developed in as many tube-fed patients as in those able to eat by mouth. Although these studies did not utilize suitably matched control subjects, they do raise doubts about whether tube feeding for this medical indication effectively improves patient comfort and quality of life. In addition, available information on the risks of tube feeding raises the concern that complications may be excessive in relation to the intended but uncertain benefits. For example, several studies cited by Sheiman and Pomerantz documented a high incidence of aspiration pneumonia in demented patients who are fed enterally. While the intervention might prevent a decline in the nutritional status of these patients, frequent episodes of aspiration pneumonia can severely compromise quality of life and cause a substantial burden of suffering. Perhaps most importantly, current information is inadequate for determining whether the risk/benefit ratio of tube feeding is more favorable than conservative management. Consider a basic evaluative parameter: mortality. One study cited found considerable mortality 18 months after PEG tube placement, reaching 75% in demented patients more than 80 years of age. This result suggests that mean duration of survival might not differ significantly whether patients are tube-fed or managed conservatively. While most of the studies reviewed lack suitable control groups, they underscore the meager evidentiary grounds for the practice of tube feeding older demented patients. See also p 1268 There are several important moral implications of this medical uncertainty. First, the focus of the moral and legal debate about tube feeding has been misplaced. The interlocutors have debated vigorously the relative priority of respect for life and quality of life in tube feeding decisions. The more basic moral questions concern the treatment itself: whether its burdens are justified by its intended benefits and whether its risk/benefit ratio is more favorable than the option of conservative management. Unless tube feeding can be recommended as tolerable, safe, and effective for its intended purposes, the issue of whether sanctity of life or quality of life should determine its use is moot. Our moral energies will be better spent if focused on the risk/benefit assessment and its implications for current practice. Second, the common practice of routinely initiating and indefinitely continuing tube feeding in the nondying, demented patient who is unable to eat must be critically reconsidered. If we are uncertain that tube feeding offers a net benefit for these patients, a more cautious clinical approach is indicated. It might be countered that, when in doubt, we should err on the side of life. However, available information does not clarify whether it is tube feeding or conservative management that enables us to err on the side of life. Some serious morbidities, such as aspiration pneumonia or bowel obstruction, may be more common in tube-fed patients, whereas mean duration of survival may be no better than with conservative management. Thus, initiation of tube feeding may not possess a relative advantage in prolonging life but may increase the burden of suffering. A cautious approach that reflects current medical uncertainty should include several components. The specific “medical indications” for initiating tube feeding should be identified and the goal of therapy should be clarified. For example, tube feeding might be initiated in order to enhance the comfort of a spoon-fed patient who has experienced increasing episodes of aspiration. In addition, periodic reassessments of the tube feeding should be specified in advance as part of the care plan. Finally, the care plan should include a commitment to discontinue the use of tube feeding if the intervention fails to have its anticipated physiological effects or if complications undermine its ability to achieve intended therapeutic goals. Third, the informed consent process should be altered to reflect the current state of medical uncertainty. Although older patients with advanced dementia lack decision-making capacity, there are two contexts in which revisions of the consent process are appropriate. One involves the execution of advance directives by older persons before being incapacitated by dementia. Older people are increasingly encouraged to execute advance directives, especially at the time they enter a hospital or nursing home. The forms or interviews in which instruction directives are executed often give special attention to the question of whether they wish to receive artificial nutrition and hydration in the event of a serious, irreversible illness. The difficulty is that the potential benefits and burdens of tube feeding are rarely discussed in any detail. Although the standard for informed consent disclosure involves those items of information that a reasonable person in the patient's position would need to know in making a decision, older adults executing advance directives are often provided very little information about the benefits and burdens of tube feeding and the relative merits of more conservative management. They are frequently left to make a decision based on vague general attitudes about “tubes” or their experiences with the similar treatment of friends or loved ones. Unless advance decision making is better informed, the correlation between older persons' directives regarding tube feeding and their own values and goals will be a matter of chance. The other informed consent context in which revisions are appropriate concerns surrogate decision-making for demented older patients whose wishes are not known. Although decisions about tube feeding can be wrenchingly difficult for family members, they must be adequately apprised of the crucial uncertainties regarding its benefits, risks, and burdens. They must also be provided with a careful assessment of the relative merits and liabilities of more conservative management. Moreover, in advising families, physicians must be willing to accept the current limitations in our knowledge about tube feeding. Given these inadequacies, it is not possible to make determinate recommendations based on “medical indications.” Rather, the decision about tube feeding will depend on the relative weight assigned to the potential benefits and burdens of the treatment options. For example, some families may assign sufficient weight to prolongation of the patient's life that the risks of complications are acceptable, whereas other families may assign priority to avoiding any interventions that might increase the patient's burden of suffering. Contrasting priorities of conscientious surrogates may lead to different decisions about similar patients. In addition, all decisions will possess a tentativeness dictated by the current limitations in our knowledge. In this regard, the use of time-limited trials of tube feeding and planned reassessments may ease the burden of uncertainty for both physicians and families. Fourth, current medical uncertainty should be a catalyst for change in restrictive policies of nursing homes and state regulatory agencies regarding tube feeding of demented patients. Many nursing homes refuse admission to patients if they are unable to maintain adequate oral intake and do not have a feeding tube in place. Similarly, state regulatory agencies conducting quality assurance audits of nursing homes may use criteria such as weight maintenance to evaluate nutritional services for residents. These evaluative criteria generate pressure to utilize tube feedings, even if they do not explicitly require them in patients unable to maintain adequate oral intake. If tube feeding of demented patients had been validated in well designed clinical trials as safe, effective, and unobtrusive, then institutional and regulatory requirements for its use would be a defensible standard of care. The problem is that tube feeding has become a widely accepted medical practice in the care of demented older patients without adequate validation in clinical research. Because we cannot be confident that tube feeding is the optimal therapeutic approach, requiring its use may contravene the interests of elderly patients with dementia. It also infringes on the freedom of physicians, surrogates, and patients executing advance directives to weigh the uncertain benefits, risks, and discomforts of tube feeding in a manner that leads to rejection of its use. In addition, it makes more difficult the use of time-limited trials and periodic reassessments inasmuch as these strategies assume the prerogative of withdrawing tube feeding once it is started. Because tube feeding is not a validated standard of care, nursing homes that continue to insist on its use in patients unable to eat might be accused of placing greater priority on the convenience of staff members than on the interests of patients. Similarly, regulatory agencies whose quality assurance rules encourage the use of feeding tubes might be charged with employing operational criteria that are easy to measure in superficial audits but that inhibit defensible variations in the care of older demented patients. Fifth, the current medical uncertainty creates a moral obligation to conduct additional clinical research. As represented in the traditional dictum, “do no harm,” physicians have an obligation to avoid causing harm to patients. Previous research has failed to clarify whether or to what extent tube feeding of older demented patients who are not dying might cause a net balance of harm or possess a risk/benefit ratio that is less favorable than more conservative management. Further research will be necessary to determine which strategy best satisfies the duty to avoid causing harm. Sheiman and Pomerantz identify several crucial areas for research: determining nutritional parameters for older demented patients, testing the safety and efficacy of tube feeding, and evaluating a hospice-type approach to the care of nondying, demented patients. The latter two research objectives can perhaps be addressed in prospective, randomized clinical trials comparing tube feeding of older demented patients with a hospice-style approach. Randomized trials may be difficult to conduct for a couple of reasons. Many physicians may be reluctant to recommend randomization of their patients, based on their philosophical objections to withholding tube feeding or their experience-based assessment that it constitutes optimal therapy. Similarly, surrogate decision makers might be disinclined to allow tube feeding decisions to be made by chance if they assign a clear priority either to doing something to prolong the life of their loved ones or to avoiding any interventions that may increase their burden of suffering. Nor can we be sure that physicians and surrogates will allow the results of randomized clinical trials to override their philosophical and experiential biases. However, the extent of these various obstacles can be explored only if we attempt to organize suitably designed randomized trials. These offer the best hope for clarifying the relative merits of tube feeding and conservative management. Finally, preliminary evidence regarding the limited efficacy and frequent complications of tube feeding should soften a persistent impediment to exploring the relative merits of conservative management. This obstacle is the concern that not feeding older patients is to “starve” them and thereby to create an intolerable burden of suffering related to hunger and thirst. As Sheiman and Pomerantz point out, available research also challenges the correctness of this assumption. Some studies have suggested that discontinuation of all but tolerated feeding may not be uncomfortable or painful. In addition, any burden of suffering associated with this option must be compared properly with the frequency and magnitude of the burdens secondary to enteral feeding. Thus, the current medical uncertainty about tube feeding should stimulate more extensive research assessment of well designed hospice-style programs for demented older people who are no longer able to maintain adequate oral intake. Older patients with progressive dementia constitute a growing population of highly vulnerable persons. This vulnerability creates a compelling obligation to assiduously protect their welfare. The current state of medical knowledge leaves unanswered the question of whether routine tube feeding of demented patients maximizes their interests. Given this limitation, we must refocus the moral debate on the merits and liabilities of the treatment itself. We must also revise our informed consent protocols and practice patterns to reflect our medical uncertainty. Finally, we must encourage the development of research programs designed to address the vexing questions posed by Sheiman and Pomerantz.