The President's Council on Bioethics click here to skip navigation


This commissioned working paper was discussed at the Council's December 2004 meeting. The views expressed here do not represent the official views of the Council or of the United States Government.

Treatment Decisions for Dementia Patients: The Search for Normative Boundaries

Issues Summary for The President's Council on Bioethics

Rebecca Dresser1


Because scholars and officials have focused on advance directives to resolve future treatment questions, they have neglected the majority of actual treatment dilemmas.  Relatively few patients have directives.  Most of the directives that do exist simply name a preferred proxy decision-maker.  Directives with instructions are usually quite general, supplying little guidance to those at the bedside.  Yet the standards that should govern care of patients without clear directives remain relatively undeveloped.

When advance directives are absent or imprecise, two standards guide patient care.  One is called the substituted judgment standard.2  This standard instructs physicians and families to make the treatment decision that the patient would make if “miraculously lucid for an interval (not altering the existing prognosis of the condition to which she would soon return) and perceptive of her irreversible condition.”3  Because there is no clear advance directive, however, the patient’s general values, religious beliefs, and attitudes toward medical care become the evidentiary basis for the treatment decision. 

Critics argue that the substituted judgment standard is malleable enough to allow relatives and others to advance their own values and concerns in the guise of a decision that purports to be what the patient would want.  To protect the patient, they say, the scope of decisions based on this standard should be limited: “the inherently speculative nature of substituted judgment, along with the vulnerable position of the incompetent, require that substituted judgment may not be used to justify a course of action that serves the interests of others at the expense of the ward’s basic interests.”4  According to this view, incompetent patients retain a core set of interests that should not be compromised.  To protect these interests, there must be an inquiry into the patient’s current situation and the benefits and burdens that would accompany various treatment decisions.  This requires observers to apply the so-called objective approach to treatment decisions for incompetent patients.

The objective approach is usually referred to as the best interests standard.  It is also known as the benefit-burden or reasonable person standard.  Objective standards are applied when evidence of the patient’s past values and preferences is ambiguous or nonexistent.  Instead of attempting to determine the decision most consistent with the patient’s subjective beliefs and values, the objective approach relies on community norms.  It rests on “a societal consensus, or the perspective of ‘a reasonable person,’ choosing as most people would choose for themselves.”5 

Applying objective standards involves weighing and balancing the benefits and burdens that an incompetent patient would experience if treatment were administered or forgone.  Examples of benefits and burdens are pain, distress, pleasure, and enjoyment.6  Objective judgments require decision-makers to evaluate whether treatment interventions would confer on an incompetent patient sufficient benefit to justify the risks and burdens the interventions would impose. 

Objective standards are controversial because they require healthy individuals to make quality-of-life judgments for vulnerable, impaired patients.  Critics fear that objective standards will incorporate social worth and economic considerations.  These fears lead most scholars and legal authorities to regard the objective approach as the least preferred treatment standard, the absolute last resort when no other standard will work.

A desire to escape the dangers accompanying objective standards contributes to the general enthusiasm for advance directives.  If treatment decisions appear to come from patients themselves, there is less cause to worry about inappropriate quality-of-life evaluations.  At the same time, the scarcity of directives and the deficiencies of the directives that exist leave patients exposed to the same risks.  And because decisions based on advance directives and substituted judgment are portrayed as choices coming from the patients themselves, social worth and economic considerations affecting those decisions may go undetected.7   

In sum, the dearth of advance directives and the shortcomings of the substituted judgment standard make the objective approach central to most real cases.  Because they have been preoccupied with advance directives, however, scholars and policy officials have neglected this approach.  As a result, it lacks precision and rigor.

The task is to clarify the scope of permissible treatment alternatives for relatives and other surrogates deciding on behalf of incompetent patients.  The growing population of dementia patients increases the need for scholars and legal authorities to develop an objective standard that provides adequate protection to incompetent patients, but gives due regard to family concerns and values, as well as competing demands for limited social resources.

All objective treatment standards share the same central difficulty.  To apply objective standards, there must be agreement about what should count as benefits and burdens to patients and what combination of benefits and burdens permits a specific treatment choice.  But people have different views on the appropriate content and application of the objective approach.  Should the approach permit withholding or withdrawing interventions that appear to impose discomfort on dementia patients?  Should it permit forgoing treatment in advanced dementia patients who appear comfortable but hardly aware of their surroundings?  In a diverse society, complete consensus on these and similar issues seems out of reach.

Nevertheless, courts and legislatures do take normative positions on the above questions.  For example, in In re Conroy, the New Jersey Supreme Court set forth an objective standard for cases in which there is no trustworthy evidence of the patient's former treatment preferences.  In such cases, treatment may be forgone if the "net burdens of the patient's life clearly and markedly outweigh the benefits that the patient derives from life [and] the recurring, unavoidable and severe pain of the patient's life [is] such that the effect of administering life-sustaining treatment would be inhumane."8  Burdens to be considered include pain and suffering, while benefits are "physical pleasure, emotional enjoyment, or intellectual satisfaction." 

Yet Conroy's formulation of the objective test has been criticized as too narrow to encompass all of the situations in which nontreatment might be defensible.  Justice Handler, in his partial dissent in Conroy, criticized the majority opinion’s focus on pain.  Although the opinion refers to other factors, such as the patient’s level of cognitive function, humiliation, dependence, and loss of dignity, in the end, Handler wrote, "[t]he presence of significant pain becomes the sole measure of such a person's best interests."9  Another New Jersey judge complained that the narrowness of Conroy’s objective test means that "large numbers of people will be thoughtlessly and automatically compelled to continue lives of intolerable bleakness."10  These judges think that the objective approach should permit families and clinicians to withhold or withdraw interventions in cases that fail to qualify as “inhumane treatment.” 

One problem that the two judges and other critics have with the Conroy test is that it rules out decisions to forgo treatment for permanently unconscious patients, because such patients are incapable of experiencing pain or suffering.  Moreover, advanced dementia patients, many of whom are "barely conscious" but not in severe pain, must receive feeding tubes and other life-extending interventions according to this test. 

Concepts such as interests, benefits, and burdens are closely tied to a patient's conscious awareness.  How should objective standards apply to patients whose awareness is permanently diminished or absent?  Courts and scholars take different positions on this question.  In the Cruzan case, the Missouri Supreme Court took the position that prolonged life in an unconscious state constituted a benefit to the patient that outweighed any possible treatment burdens.  In contrast, the Wisconsin Supreme Court held that forgoing treatment in permanently unconscious patients is permissible under an objective standard.  The rationale for the latter position is stated below:

Patients who are permanently unconscious are unaware of benefits and burdens.  The only possible benefit to them of life-sustaining treatment is the possibility that the diagnosis of irreversible unconsciousness is wrong and they will regain consciousness. Accordingly, the major considerations are whether a reasonable person in the patient's circumstance would find that this benefit, as well as the benefits to the patient's family and concerned friends (such as satisfaction in caring for the patient and the meaningfulness of the patient's continued survival) are outweighed by the burdens on those loved ones (such as financial cost or emotional suffering).11

Another conflict concerns what may count as burdens to the patient.  Some writers defend what is referred to as the "sanctity of life" view, which holds that extended life should never be regarded as a burden to the patient.  According to this view, "burdensomeness should be assessed by focusing on the pain or invasiveness caused by the treatment itself, not by evaluating the quality of life that such medical intervention may sustain."12

An alternative view considers it appropriate to include as possible burdens and benefits the kind of life the patient will gain with treatment.  Thus, "discontinuing treatment, even if it leads to the patient's death, is consistent with his or her best interests when the treatment is hopeless and serves only to sustain biological existence that is painful or of no benefit to the patient."13  Supporters of this view say that as long as the examination remains focused on the value life holds for the individual patient, as opposed to the value the patient's life holds for others, the lives of disabled persons will be adequately protected:

The question is not whether the patient's quality of life is below average, or worse than it used to be, or anything of the sort.  Instead, the proper quality of life judgment is only whether the quality of the patient's life with the life-sustaining treatment will be so poor as to be not worth living or worse than no further life at all.14

Related to these debates is an argument about whether objective standards should consider only burdens and benefits the patient actually experiences, or should consider "reasonable person" values and preferences, too.  One view is that decision-makers should consider factors such as pain, suffering, distress, pleasure, dignity, privacy, and humiliation only if the patients themselves are consciously experiencing these states.  This version of the objective standard is highly protective of individual patients and allows nontreatment solely when the burdens of continued life with treatment outweigh the benefits from the patient's own perspective.  Yet this version is criticized as overly restrictive, since it fails to permit nontreatment when many people would choose this for themselves or their loved ones.

The second and more expansive version of the objective standard permits decision-makers to consider factors that a "reasonable person" would consider relevant to treatment decision-making.  This view is described in the following statement:

In applying the "reasonable person" standard to the patient with an illness or disabling condition that is severe and irreversible and who lacks decision-making capacity, the major consideration is the following:  Would a reasonable person in the patient's circumstances probably prefer the termination of treatment because the patient's life is largely devoid of opportunities to achieve satisfaction, or full of pain or suffering with no corresponding benefits?15

This version of the objective approach allows nontreatment of permanently unconscious patients, on grounds that a reasonable person would find unconscious existence insufficiently worthwhile to insist on its prolongation.  Even though patients in this condition are unaware of their circumstances, normative values such as preservation of dignity and the importance of consciousness and relationships to human life suggest that forgoing treatment is permissible.     

But those supporting the expansive interpretation of the objective approach must decide how expansive the approach should be.  There is concern that if the standard allows decisions to rest on concerns such as preserving dignity, which many incompetent patients are themselves unable to appreciate, vulnerable patients will go untreated.  The expansive approach could sanction, for example, nontreatment of patients with mild or moderate dementia who appear comfortable and content, but lead lives that many people see as humiliating and undignified.16  

One way to restrict the scope of reasonable person judgments is to require evidence that a substantial percentage of the public agrees with them.  For example, in several surveys of the general public, at least 80 percent said that if they became irreversibly unconscious they would want life support withheld at their families’ request.17  Decision-makers asked to apply reasonable person values to cases involving patients in other circumstances, such as moderate or severe dementia, might insist on evidence of a similar social consensus regarding the permissibility of nontreatment.

Yet even if its scope is limited by public opinion, the reasonable person approach permits relatively healthy people to evaluate the lives of incompetent patients with various disabilities and illnesses.  Because healthy people have never personally experienced these disabilities and illnesses, they may assign inappropriately low value to life with such conditions.  Two authors have attempted to ameliorate this problem by shifting the locus of evaluation to a different population.  In their proposal, judgments on when nontreatment would be permissible would come from a population of competent patients who have personally experienced illness or injury.  This approach seeks to incorporate an element of subjectivity into the objective standard, by adopting the viewpoint of a reasonable person with a degree of personal insight into the patient's situation.18 

Still another controversial issue is whether objective standards should incorporate burdens experienced by others.  According to the President's Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research, "[t]he impact of a decision on an incapacitated patient's loved ones may be taken into account in determining someone's best interests, for most people do have an important interest in the well-being of their families or close associates."19  Another policy group criticized this position, however, questioning whether such “imputed altruism” should be part of the reasonable patient standard.  According to this group, the interests of others should influence treatment decisions only if there is strong evidence that the average person would take such interests into account in making medical choices.20

A related issue is whether a lack of adequate services and facilities for the care of incompetent patients should influence treatment decision-making according to an objective standard.  For example, "some would justify withholding life-sustaining treatment from a severely demented elderly individual ... on the grounds that it would not be in that individual's best interests to live in an inadequately funded state institution ...."  Buchanan and Brock reject this view, arguing instead that the best interests standard calls for remedying deficiencies in the care available to such patients.  Yet they concede that "neither the interests of others nor the defects of existing institutions can be ignored in determining what is in the incompetent's best interest."21

Factors such as family burdens and available care options are likely to influence decisions at the bedside.  The moral challenge is to keep patient welfare at the forefront.  Although most scholars and officials agree that families should ordinarily be empowered to make treatment decisions for their incompetent relatives, they also support some limits on family decision-making.           

These limits seek to guard against both overtreatment and undertreatment.  Overtreatment is possible because family decision-makers could request invasive and burdensome interventions out of guilt, a desire to avoid responsibility, or a desire to avoid the appearance of suspect motivations.  Undertreatment is also a concern, for empirical evidence indicates that family members underestimate their elderly relatives' quality of life and functional status.  Because financial and other burdens often accompany a decision to continue an incompetent patient's life-sustaining treatment, some families could be too eager to stop treatment to end their own ordeal.22

The above concerns point to the need to restrict the family's discretion to continue or terminate an incompetent patient's life-sustaining treatment.  Some statutes and court decisions have implicitly adopted such limits by permitting family surrogates to decide only on behalf of permanently unconscious, terminally ill, advanced dementia, or other seriously compromised patients whose interests in continued life are regarded as comparatively small.23

Commentators have attempted to formulate more general limits.  Law professor Nancy Rhoden, for example, argued for a presumption in favor of the family's choice to forgo treatment, unless that choice is "unreasonable."  She proposed that a decision to forgo treatment should be classified as unreasonable if the incompetent patient "can experience some of life's pleasures in a manner not outweighed by the pain or limitations of her condition or medical treatment or the patient's past preferences and statements show she would want treatment continued under the circumstances."24  Other commentators speak in terms of permitting families to decide "within the range of medically sound alternatives, as determined by appropriate medical standards" or "a range of plans of care [recognized] as generally acceptable by professional and community standards."25

Additional guidance comes from the literature on parental decision-making for children.  In a recent article, for example, Michael and David Benatar argue that parents should be allowed to decide whether boy babies should be circumcised.  These authors recognize that in light of the available medical data, the procedure fails to advance the child’s clear best interests.26  At the same time, they say, the procedure is not so detrimental to the child’s welfare that medical ethics and law should deem it unacceptable. 

The Benatars propose that even though circumcision cannot be justified under a strict interpretation of the best interests standard, the intervention falls within the scope of permissible parental choice.  They contend that parental discretion should encompass decisions that express religious and cultural values.  At the same time, they argue for limits on this discretion.  For example, they say, parents should not be permitted to demand circumcision without the application of pain-relieving measures.  In that situation, the parents’ values should not be permitted to trump the child’s interests in being kept comfortable during the procedure.

A similar analysis could support classifying certain dementia treatment decisions as outside the limits of acceptable family choice.  One example is family refusal of oral antibiotics for a patient with mild dementia who retains the ability to enjoy numerous activities and relationships.  A much closer case is family refusal of major surgery or chemotherapy for a patient with moderate dementia.  In this situation, the patient may be unable to understand or remember the reasons for the restrictions and physical discomfort the interventions would impose.  Some people would see this treatment question as one that the family should be free to resolve.  Others would disagree, arguing that nontreatment would be too detrimental to the patient.  A third case is family refusal of tube feeding for “barely conscious” advanced dementia patients.  Although this option may not be covered by a narrow concept of the patient’s best interests, it is often defended as a reasonable choice in light of the questionable benefit that tube feeding confers on such patients.27 

Developing a robust objective approach to treatment decisions for dementia patients will require public and professional debate over cases such as these.  It will require an understanding of how different treatment interventions can affect dementia patients.  It will also require thoughtful deliberation about the complex moral dimensions of these cases.  People with dementia have a progressive terminal illness, but many years can elapse before death becomes inevitable.  During much of this time, patients are aware and able to interact with the world.  They can be harmed or helped by the medical decisions others make for them.  The Council’s inquiry could contribute to the search for defensible moral boundaries governing these decisions.  

1. This review consists of slightly edited material from three prior publications: Dresser, Precommitment: A Misguided Strategy for Securing Death With Dignity, 81 Tex. L. Rev. 1823 (2003); Dresser, Standards for Family Decisions: Replacing Best Interests With Harm Prevention, 3 Am. J. Bioethics 54 (2003); Shapiro, Spece, Dresser & Clayton, Bioethics and Law: Cases, Materials and Problems (St. Paul: West Group, 2003).

2. This standard is also known as the limited objective standard.  In re Conroy, 486 A.2d 1209, 1232 (N.J. 1985).

3. In re Quinlan, 355 A.2d 647, 663 (N.J. 1976).

4. Buchanan & Brock, Deciding for Others: The Ethics of Surrogate Decision Making 117-18 (1989) .

5. New York State Task Force on Life and the Law, When Others Must Choose 33 (1992).

6. See In re Conroy, 486 A.2d 1209, 1232 (N.J. 1985).

7. For example, empirical studies indicate that nursing homes allow family members to sign “advance directives” for older adults who lack decisional capacity.  Molloy, Systematic Implementation of an Advance Directive Program in Nursing Homes, 283 JAMA 1437 (2000); Elizabeth Bradley et al., Institutional Efforts to Promote Advance Care Planning in Nursing Homes: Challenges and Opportunities, 25 J. L. Med. & Ethics 150 (1997).  Although family decision-making is often appropriate, it is disingenuous to categorize it as an advance directive.  This is a clear demonstration of how advance directives can be manipulated to justify decisions that rest on considerations other than the patient’s former choice.  In such cases, the directives “may not represent the informed [nursing home] resident’s wishes but rather what the surrogate deems to be in the resident’s best interests, or, potentially, in the surrogate’s best interests.”  Bradley et al., supra, at 155. 

8. Conroy, 486 A.2d at 1232.

9. Conroy, 486 A.2d at 1247, 1249.

10. In re Visbeck, 510 A.2d 125, 133 (N.J. Super. Ct. Ch. Div. 1986). 

11. Hastings Center, Guidelines on the Termination of Life-Sustaining Treatment and the Care of the  Dying 29 (1987).  See also President's Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research, Deciding to Forego Life-Sustaining Treatment 181-86 (1983) (remote possibility that permanently unconscious patient was incorrectly diagnosed is insufficient to give patient overriding interest in receiving life-sustaining treatment); Quinn, The Best Interests of Incompetent Patients: The Capacity for Interpersonal Relationships as a Standard for Decisionmaking, 76 Calif. L. Rev. 897 (1988) (arguing that nontreatment should be permissible for patients who have irreversibly lost minimal capacity for social interaction).

12. New York State Task Force, When Others Must Choose, at 57. 

13. Id. at 58.

14. Buchanan & Brock, Deciding for Others, at 124 (1989).

15 Hastings Center, Guidelines, at 28-29.

16. See Brock, Justice and the Severely Demented Elderly, 13 J. Med. Phil. 73, 93 (1988) ("significant numbers of persons, though by no means all persons, find the continued life of the moderately demented that is sometimes supported by various kinds of life-sustaining treatment on balance no longer a benefit or desirable in comparison with the alternative of an earlier death").

17. Davis, Shifting the Burden of Proof, 31 Second Op. 31 (1993).

18. See Emanuel & Emanuel, Decisions at the End of Life: Guided by Communities of Patients, Hastings Center Rep., Sept.-Oct., 1993, at 6.

19. President's Commission, Deciding to Forego Life-Sustaining Treatment, at 135 (footnote omitted).  See also In re Torres, 357 N.W.2d 332 (Minn. 1984) (patient's presumed desire to avoid burdens on survivors is one reason treatment withdrawal is in patient's best interests).  

20. New York State Task Force, When Others Must Choose, at 62-65, 109.

21. Buchanan & Brock, Deciding for Others, at 134.

22. Emanuel & Emanuel, Proxy Decisionmaking for Incompetent Patients, 267 JAMA 2067 (1992).

23. Oregon Rev. Stat. § 127.635 (1997). 

24. Rhoden, Litigating Life and Death, 102 Harv. L. Rev. 375, 442 (1988).

25. Buchanan & Brock, Deciding for Others, at 143; Lynn, Procedures for Making Medical Decisions for Incompetent Adults, 267 JAMA 2082 (1992).

26. Benatar & Benatar, Between Prophylaxis and Child Abuse: The Ethics of Neonatal Male Circumcision, 3(2) Am. J. Bioethics 35 (2003).

27. Gillick, Rethinking the Role of Tube Feeding in Patients with Advanced Dementia, New Eng. J. Med. 1755 (2000); Finucane, Christmas & Travis, Tube Feeding in Patients with Advanced Dementia, 282 JAMA 1365 (1999).

  - The President's Council on Bioethics -  
Home Site Map Disclaimers Privacy Notice Accessibility NBAC HHS