Physicians' Attitudes about the Care of Patients in the Persistent Vegetative State: A National Survey

Methods

Questionnaires were mailed to systematically selected samples of 250 persons from a random sample of practicing members of the American Academy of Neurology in 1992 and of 250 physicians on the 1992 membership list of the American Medical Directors Association, a national organization of nursing home medical directors. The first mailing was done in January 1993 and was followed by two additional mailings.

In our questionnaire, the PVS was defined as “an irreversible condition characterized by an apparent total lack of awareness with preserved sleep-wake cycles.” We chose to define the PVS as an irreversible, permanent condition because this conformed to the definition provided in the 1989 position paper of the American Academy of Neurology [2]. However, we recognize that the position papers published by the American Medical Association [3] and the American Neurological Association [6], as well as a paper on the management of the patient in the PVS published by the American Academy of Neurology in 1995 [7], differ from the views of the American Academy of Neurology in distinguishing between establishing the diagnosis of the PVS as a persistent condition and determining that the condition is permanent. The report of The Multi-Society Task Force on PVS [4], published after our survey was completed, also emphasizes the distinction between diagnosis of the PVS and the determination that the condition is permanent. However, it was our judgment at the time we mailed our survey that physicians most often used the term “PVS” to refer to a permanent condition, as we defined it in our survey, and that using such a definition would minimize the potential for confusion.

Questions were answered on the basis of a four-point scale (1 equals definitely yes, 2 equals probably yes, 3 equals probably no, and 4 equals definitely no). For the purposes of reporting data, responses 1 and 2 were combined to count as “yes” and responses 3 and 4 were combined to count as “no.” Copies of the questionnaire are available from Dr. Sachs.

Statistical analysis was done using SAS software (SAS Institute, Inc., Cary, North Carolina). Group means were compared using the t-test. The chi-square was used to test associations between categorical variables. We calculated CIs using the formula Equation 1 where p₁ is the medical directors' response rate, p (₂) is the neurologists' response rate, η₁ is the population of medical directors, and η₂ is the population of neurologists.

Participation by physicians was voluntary, and the confidentiality of responses was assured. The Institutional Review Board of the University of Chicago approved the research protocol.

Results

One hundred sixty-nine American Academy of Neurology members (neurologists) and 150 American Medical Directors Association members (medical directors) returned completed surveys, yielding response rates of 68% and 60%, respectively. The demographic characteristics of both groups and the percentages of physicians who reported that they have an advance directive are listed in Table 1. The two groups were similar except that more neurologists reported practicing in an academic setting and more medical directors reported having an advance directive for their own medical care. Perhaps more striking than the difference between the groups is that so many physicians in both groups had executed advance directives.

Table 2 lists responses to questions about diagnosis of the PVS, beliefs about adult patients in the PVS, and the minimum amount of time from initial brain injury during which an adult patient must be continuously unconscious before the PVS can be diagnosed. Although patients in the PVS are, by definition, unconscious, substantial minorities of both physician groups believe that these patients experience pain, thirst, and hunger; are aware of self and environment; and are made more comfortable by intravenous fluids and tube feedings. These beliefs were equally prevalent among academic and nonacademic physicians. About 94% of both physician groups thought that patients in the PVS would be better off dead, and almost half of all responders thought that patients in the PVS should be considered dead. The duration of unconsciousness believed to be necessary before the PVS could be diagnosed was relatively short in both physician groups (Table 2).

Although it is common to discuss whether tube feeding, mechanical ventilation, or cardiopulmonary resuscitation should be provided around the time the PVS is diagnosed, we were interested in the treatments that physicians believe should be given as part of ongoing care for patients who remain alive in the PVS. Table 3 lists the responders' answers to the following questions: 1) Should the following health maintenance procedures be provided routinely to adult patients in the PVS? and 2) Do you believe that the following illnesses should be treated aggressively in patients in the PVS? Few responders thought that mammography and prostate examinations should be done, more than one third of both groups would monitor blood pressure, and almost one third of medical directors would vaccinate to protect against pneumococcal and influenza infections. Similar proportions of both groups would aggressively treat infections and such illnesses as hypertension and diabetes. Few responders would aggressively treat the more catastrophic or serious illnesses listed in Table 3.

Additional questions focused on the use of specific treatments instead of illnesses. Table 4 lists responses to the three questions: Should the following therapies be generally provided to adult patients in the PVS? Do you believe that it is ethical to withhold or withdraw the following therapies from adult patients in the PVS when requested to do so in a living will or by an appropriate surrogate decision maker? If you were in the PVS, which of the following medical interventions would you want provided? Ten percent or less of both groups thought that cardiopulmonary resuscitation, ventilators, dialysis, and transfusions should generally be provided, and about one third of both groups believed that antibiotic agents should be used. Neurologists and medical directors substantially differed in beliefs about artificial hydration and nutrition; almost half of the neurologists believed that food and fluids should be provided. Most responders from both groups believed that it would be ethical to withhold or withdraw all of these therapies according to the dictates of a living will or a patient's surrogate. Few responders said that they would want any of these therapies for themselves if they were in the PVS.

In the context of a hypothetical universal health plan, more than 93% of responders believed that it would be ethical to (and that they would) discontinue payment for cardiopulmonary resuscitation, mechanical ventilation, dialysis, and transfusions for adult patients in the PVS to provide adequate funds for other pressing health care needs. Under the same circumstances, 83% would agree to discontinue payment for antibiotic therapy and 76% would agree to discontinue payment for artificial hydration and nutrition 1 year after diagnosis of the PVS.

After a decision has been made to discontinue all therapy (including fluid and nutrition) for a patient in the PVS so that death becomes inevitable, 65% of surveyed physicians believed that it would be ethical to use that patient's vital organs for transplantation; 20% believed that it would be ethical to hasten the patient's death by lethal injection. The responses of the neurologists and the medical directors did not differ for these two questions.

Discussion

Our survey was designed to study physicians' attitudes about issues related to caring for adult patients in the PVS. Because neurologists often serve as consultants who provide diagnostic and prognostic guidance whereas medical directors provide longitudinal primary care, we initially speculated that these two groups might have different attitudes about the care of patients in the PVS. However, the similarities between the attitudes of these two groups were more striking than the differences.

On the basis of an extensive literature review, The Multi-Society Task Force on PVS [4] concluded that the recovery of consciousness is “exceedingly rare” after 3 months in nontraumatic (hypoxic-ischemic) PVS and “unlikely” after 12 months in traumatic PVS. Both groups of responders believed that permanence can be determined more quickly after hypoxic-ischemic brain injury than after traumatic brain injury, a finding consistent with the empiric data. However, most physicians in both groups also believed that the PVS can be diagnosed more quickly than the results of empiric studies indicate [4]. This discrepancy could reflect either a lack of knowledge of the literature or a willingness of surveyed physicians to accept a lesser degree of certainty in making the diagnosis. It is also possible that the responders thought that they were being asked to state the earliest time at which one can ever determine that any patient is permanently unconscious rather than the length of time that is generally required for patients with a particular kind of brain injury. For example, Levy and colleagues [10, 11] have published empiric data suggesting that, in some cases of coma resulting from nontraumatic brain injury, prognosis can be determined within 2 weeks of injury.

An unpublished survey by the American Neurological Association (Daroff RB. The American Neurological Association survey results on PVS [Presented paper]. 115th Annual Meeting; 1990; Atlanta) also found that a significant minority of its members were “uncertain” about whether patients in the PVS could experience pain (31%) and suffering (26%). Tresch and colleagues [12] found that 22% of the relatives of patients in the PVS believed that their relative in the PVS could experience pain and suffering. Although patients in the PVS are unconscious and unaware by definition, studies have attempted to prove this assertion [13, 14]. Physiologic and anatomical studies of patients in the PVS and their brains suggest that such patients are not capable of subjective experience. This scientific evidence is apparently inadequate to convince a significant minority of physicians and other persons that patients in the PVS are completely unaware; our survey did not determine whether these physicians are unfamiliar with the evidence or whether they are familiar with it but remain unconvinced. It is also possible that these physicians are concerned that patients who actually are aware and who can therefore suffer may be incorrectly classified as being in the PVS.

In contrast with the disagreement about the capacity of patients in the PVS for subjective experience, most responders agreed that patients in the PVS would be better off dead. Thus, most physicians who believed that patients in the PVS are aware of self or environment also believed that these patients would be better off dead, especially if they can suffer.

Veatch [15] and Gervais [16], among others, have argued that the definition of death should be changed to a “higher-brain formulation” instead of the current “whole-brain formulation.” Using this criterion, patients in the PVS would be classified as dead. Bernat [17] and others have opposed this change for both philosophical and practical reasons. However, almost half of our responders agreed that patients in the PVS should be considered dead. Redefining death to include patients in the PVS would have profound consequences: All ethical obligations to such patients, other than the obligation to treat the corpse with respect, would cease. This would, of course, eliminate all ambiguity about the appropriate treatment for patients in the PVS. The recent Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments [18] suggests that many physicians continue to have difficulty addressing the complex moral dilemmas common to the management of critically ill patients. Defining patients in the PVS as dead would eliminate the physicians' responsibilities to these patients and thus may, in some respects, appeal to some physicians. Debates about appropriate care, rationing, and futility with regard to patients in the PVS would become irrelevant if these patients were defined as dead.

Because current U.S. law allows organs to be harvested only after death and because patients in the PVS are not currently classified by law as dead, harvesting their organs (on the basis of an advance directive or a surrogate's decision) is prohibited. Defining patients in the PVS as dead would allow their vital organs to be harvested for transplantation. At least one third of the responders believed that it was ethical to harvest vital organs from patients in the PVS, despite their belief that such patients should not be considered dead. Indeed, our responders' views prefigured the recent opinion of the American Medical Association Council on Ethical and Judicial Affairs [19] that supported the use of anencephalic infants as a source for vital organs (this was subsequently suspended pending further review). Anencephalic infants are neurologically similar to adults in the PVS. The American Medical Association did not recommend changing the definition of death to include anencephalic infants; rather, it recommended permitting the harvesting of vital organs from anencephalic infants even though these infants are not dead. Thus, for at least some physicians both in our survey and in the American Medical Association, the benefit of increasing the supply of organs appears to take priority over maintaining death as a prerequisite for organ harvest.

It is confusing that only one fifth of responders believe that it is ethical to hasten the death of a patient in the PVS with a lethal injection, given that nearly half of these physicians believe that such patients should be considered dead. Why would one object to hastening the death of a patient whom they believe should be considered dead (the semantic inconsistencies inherent in such an act notwithstanding)? Moreover, this minority stands in contrast with the majority, who believe that the vital organs of patients in the PVS may be harvested—an act that would also cause their deaths. A possible rationale for accepting the harvesting of vital organs from these patients as ethical but rejecting lethal injection is the beneficial effect of organ procurement on persons other than the patient in the PVS. For the patient in the PVS, the outcome of lethal injection and of the harvesting of vital organs is the same: immediate death. However, the effects on persons other than the patient are different. Harvesting vital organs could save or improve the life of the transplant recipient. The benefits to the recipient may be considered more important than the negative effects on the patient in the PVS from whom the organs are harvested. In contrast, lethal injection would not directly benefit other persons.

Although a sizable minority of physicians in both groups would routinely monitor blood pressure and treat simple conditions, such as hypertension, diabetes, and infections, in patients in the PVS, few would screen for cancer or routinely treat more severe illnesses. Because we did not explore the reasoning behind these responses, we can only speculate about the role of treatment burden, cost, and invasiveness in the responders' choices. Artificial nutrition and hydration was the treatment that more physicians would provide routinely or request for themselves, as might be expected from the controversy surrounding the topic. We suspect that the large difference in beliefs about vaccinations can be attributed to the greater appreciation medical directors have of the concept of herd immunity as it applies to residents of long-term care facilities.

In the context of a hypothetical universal health plan, most responders endorsed a reimbursement policy explicitly based on the concept of the just allocation of resources. In this context, it appears that for most responders, the principle of distributive justice takes precedence over the principle of respect for autonomy. Thus, when society is paying the bill, the needs of society as a whole apparently outweigh the previously expressed wishes of the patient or the current wishes of the patient's surrogate.

In our survey, the percentage of responders that reported the completion of some form of advance directive was greater than the 20% reported in previous studies of the U.S. population [20, 21]. Because we did not ask for reasons, we can only speculate about why the percentage was so high. Perhaps the experience of caring for patients in the PVS and for other neurologically impaired patients motivated physicians to execute advance directives. Most physicians reported that they would desire minimal therapeutic interventions for themselves if they were in the PVS; this may have prompted some physicians to formally document this preference in an advance directive. More medical directors than neurologists reported having advance directives. This difference may reflect the experience medical directors have in caring for severely debilitated patients over an extended period; neurologists presumably have less-extended contact with such patients. Also, the responding medical directors had an average of 6 more years of clinical experience and were an average of 5 years older than responding neurologists.

The following limitations of our study should be recognized: the modest response rate, the ambiguity of some questions (as discussed previously), the difficulty of extrapolating from hypothetical scenarios to actual practice, and potential response biases (persons with the strongest opinions are those most likely to fill out questionnaires). We did not survey a sample of physicians who did not return questionnaires, and we have no information about this group of physicians.

Overall, our results suggest that neurologists and medical directors differ little in the treatment that each group considers appropriate for patients in the PVS. There was overwhelming agreement that patients in the PVS would be better off dead; that more aggressive forms of medical treatment should not generally be provided; and that all medical therapy, including artificial hydration and nutrition, can be withheld in specific circumstances. Moreover, in the context of a hypothetical universal health care system, almost all surveyed physicians would agree to the explicit rationing of most treatments for patients in the PVS.

The responders disagreed about whether patients in the PVS retain some degree of awareness, whether they should be considered dead, and whether lethal injection or the harvesting of vital organs from these patients is ethically permissible. Thus, controversies that raise profound philosophical and moral questions remain.

Health care reform forces our society to make open and honest decisions about rationing and raises questions about what will we pay for and how long we will pay. Our survey findings suggest that neurologists and medical directors share remarkably similar views about ethically appropriate care for patients in the PVS. This consensus raises several interesting questions. Is there an ethical standard that most physicians believe should be followed in managing these patients? To what extent should the beliefs shared by this majority of physicians be considered when practice guidelines and health care policies are established? What options should be allowed for physicians who disagree with such guidelines? These questions must be addressed not only by physicians but, more important, by society. Perhaps our findings of what experienced physicians believe to be ethically appropriate care can provide a useful perspective for further discussion of these complex issues.

Dr. Taylor: Department of Neurology, Ohio State University, 1654 Upham Drive, Columbus, OH 43210-1228.

Drs. Sachs and Stocking: University of Chicago Medical Center, Department of Medicine (MC6098), 5841 South Maryland Avenue, Chicago, IL 60637-1470.