The place of autonomy in medical ethics

‘Autonomy’ is a key concept in our argument, so it is essential to be very clear about its meaning, and its relevance for the Substituted Judgment Standard. We outline possible moral justifications for the Substituted Judgment Standard, describe our chosen justification’s place on this normative map, and describe its normative appeal. By ‘autonomy’, we here mean ‘second order autonomy’: the ability to make reflective choices between one’s available courses of action and the ability to reflect over one’s preferences.5 This conception emphasises the ‘authenticity’ of one’s choices. Authenticity in this sense refers to the person’s ‘second order identification with first order desires’.6 That is, she endorses only desires that survive her own critical scrutiny. Such ability occurs in degrees, and may also vary depending on the context and what types of choices people face. Given the myriad of interesting conceptions of autonomy and their potential relevance for the Substituted Judgment Standard, we choose to focus our attention on second order autonomy. We focus entirely on so-called personal autonomy, which can be exercised regarding all areas of one’s life and that does not only concern the source of moral obligation in the Kantian sense.7 In virtue of having this capacity for second order autonomy, patients are commonly regarded as possessors of certain rights to self-governance against healthcare professionals. Such a right could be understood as ‘a barrier to unchecked paternalism’.6 8 These rights are, according to Feinberg, ‘a set of rights expressive of one’s sovereignty over oneself’.9 These rights protect ‘the independence of one’s deliberation and choice from manipulation by others’.6 Christman suggests: ‘Autonomy is the ability to decide, so for the autonomous subject of such interventions paternalism involves a lack of respect for autonomy.’6 Seana Shiffrin suggests that respect for autonomy in this sense is appealing because she takes the value of having ‘basic control’ and ‘self-direction’ over one’s experiences to be important.10 This principle of respect for autonomy is relevant for the question of the moral basis of the Substituted Judgment Standard. If healthcare professionals are required to respect patients’ (sufficiently) autonomous, informed decisions, it might be defensible to suggest that they should aim at showing implicit respect for incapacitated patients’ autonomy by trying to reconstruct the choices patients with, for example, dementia, severe mental disorders or neonatal conditions would have made under the circumstances at hand had they been competent. Beauchamp and Childress, in their influential Principles of Biomedical Ethics, emphasise that the value of autonomy does not override all other moral considerations involved in substituted decision making.11 They recommend the use of reflective equilibrium in order to adequately consider the moral values at stake in clinical decisions. We do not attempt to establish that ethical principles always outplay pragmatic clinical judgment in difficult decisions. We neither attempt to say that the value of autonomy is the only potential moral basis for the Substituted Judgment Standard or that the value of autonomy always trumps all other moral considerations. Rather, we suggest that the value of autonomy is one interesting candidate for a moral basis. Norman Cantor, for instance, emphasises clearly that he considers substituted decision making to be a tool for respecting the previously autonomous patient’s exercise of her capacity for self-determination:

Every jurisdiction that has spoken to end-of-life surrogate decision-making has upheld a formerly competent patient’s prerogative to shape post-competence care either by advance medical directive or other prior expressions. (…) All this reliance on prior expressions and previously formed values seeks to honor the previously competent patient’s autonomy or self-determination.2

Agnieszka Jaworska makes similar points. However, she suggests that autonomy is the moral basis of the Substituted Judgment Standard.1 She suggests that the value of autonomy should be considered as trumping considerations of beneficence in the circumstances at hand.1

Alternatively, clinicians’ obligation to collect informed consent could be justified by referring to some moral theory about a social contract. Use of the Substituted Judgment Standard could be justified on the basis of some moral theory about hypothetical contract in situations where the patient is incompetent.12 According to this idea, using substituted Judgment is justified because the patient would have consented to certain treatment under idealised conditions, that is, had she been competent.

In what sense do unconscious adults have interests?

Medical ethicists commonly endorse two claims: (1) Unconscious patients who have never been conscious but will become conscious can have interests while they are unconscious. (2) Unconscious patients who have been conscious and will become conscious can have interests while they are unconscious.

Some answer to the question ‘what does the claim that an unconscious patient can have interests of any kind while she is unconscious really mean?’ is required in order to answer the question ‘what type of treatment serves the unconscious patient’s ‘best interests?’ We will first review suggested answers to the latter question, which has received significant attention. Then we will focus on the first question.

There is a very extensive literature assessing how different jurisdictions interpret and apply the concept ‘best interests’, and how they distinguish between the Best Interests Standard and the Substituted Judgment Standard. We cannot here embark on a comprehensive survey of all the interpretations of The Substituted Judgment Standard and The Best Interests Standard and their applications in different jurisdictions. Instead, we will briefly outline influential interpretations and applications, and then carefully specify what interpretations this article concerns. We will outline how these issues have been treated in a British context. This is particularly interesting because the British interpretation of ‘best interests’ incorporates interpretations of the Substituted Judgment Standard. Hence, we will see that the Best Interests Standard regulates substituted decisions in the British context but interpreted very inclusively.

The British Mental Capacity Act of 2005 (MCA) holds that ‘an act done, or decision made, under this Act for or on behalf of a person who lacks capacity must be done, or made, in his best interests’.14 The MCA provides a non-exhaustive checklist of factors to be considered when determining what is in the best interests of an incapacitated individual. The MCA suggests that considerations of what the patient would have wanted had she been capacitated should be integrated in the process of determining the patient’s best interests.14 Antal Szerletics’ report Best interests decision-making under the Mental Capacity Act 15 includes discussion of the application of the MCA to financial decisions on behalf on an incapacitated individual. Such applications have influenced the application of the MCA to clinical decisions and are therefore relevant to our understanding of the evolution of the best interests standard. Szerletics15 notes that ‘it might be possible that the best interests of the individual actually corresponds to what the person would have wanted if he or she had capacity to make the decision’ (compared with ref 16). Szerletics15 also discusses MCA’s application of the Best Interests Standard to medical decisions regarding a permanently vegetative patient17 and notes that ‘The Court also makes it clear that the English notion of objective best interests cannot be equated with the substituted judgment approach as adopted in the United States but the views and the personality of P will necessarily form part of the best interests assessment’ (section 3). Szerletics15 also notes that ‘This view has been further elaborated in a subsequent Consultation Paper in which it is argued that the ‘best interests’ and ‘substituted judgment’ standards are not mutually exclusive and it favours ‘a compromise whereby a best interests test is modified by a requirement that the substitute decision-maker first goes through an exercise in substituted judgment’ (section 3:3). Szerletics15 notes that there has been an ‘integration of the substituted judgment standard into the best interests scheme’ (4:1, compared with ref 18).

The use of something akin to the Best Interests Standard might be traced back to the 14th century legislation Parens Patreae, which set legal standards for the custody of mental incompetents.18–21 Subsequently, in 1959, the Mental Health Act occurred. Szerletics holds that, before 1989, the Best Interests Standard was not commonly applied to cases involving mentally disabled individuals. The Scottish Adults With Incapacity Act22 adopted the concept of benefit instead of best interests (15: section 5.3). Significant legal cases include an early formulation of Substituted Judgment Standard.23 24 The so-called ‘Bolam Test’ has been proposed as a criterion of ‘best interests’.25 According to this criterion, a patient’s best interests are protected by ‘a practice accepted as proper by a responsible body of medical men skilled in that particular form of treatment’.25 Szerletics describes how the Bolam Test was replaced by a ‘broader, welfare based assessment of best interests’ (section 3:1)14. Szerletics points out that although the Best Interests Standard interpreted in MCA’s inclusive sense is the standard test for substituted decision making in the UK, the best interests standard is rather a ‘fall back principle’ in Canadian legislation, to be used when the previously expressed wishes of the patient are unknown (14: section 5:4) The debate regarding what treatment is in the best interests of unconscious patients assumes, however, that unconscious patients can have interests in some senses while unconscious. However, what do we mean when we claim that an unconscious patient has interests, in any sense, while she is unconscious? We will explain why we find several proposed answers to this question puzzling, although one answer is somewhat less puzzling than the others.

It might simply be highly intuitive to hold that an individual’s interests can be affected by treatment of her while she is unconscious. Cantor and Feinberg are representative advocates of such a view.2 26 Ascribing desires to unconscious adults might appear attractive simply because of the apparent absence of alternative explanations of the offensiveness of certain interventions. Michael Tooley argued that a temporarily unconscious adult has a ‘conceptual capability’ for having desires if she had desires immediately before becoming unconscious.27 He introduces this concept in order to explain how individuals who are incapacitated in certain ways can remain rights bearers while they are incapacitated. He does not explain or justify that claim except from suggesting that conceptual capacities are needed in order to preserve his theory’s intuitive plausibility.

One of the most interesting suggestions of how unconscious humans could be described as having interests while they are unconscious has been developed by Ronald Dworkin in his 1993 book Life’s Dominion. Dworkin distinguishes between ‘experiential’ and ‘critical’ interests. Having an experience or engaging in an activity is in one’s experiential interest if the individual enjoys the experience or activity. Listening to poetry or music might give the individual a pleasant experience, while hearing someone sing out of tune might give the individual an unpleasant experience. Satisfaction of experiential interests does not make the individual’s life go genuinely better. Something is in one’s critical interest if it contributes to making one’s life go genuinely better; it contributes to ‘what makes a life good’.28 According to Dworkin, one should want things that contribute to the genuine goodness of one’s life, and people can be mistaken regarding what things make their lives go genuinely better. Having close personal relationships, achieving valuable accomplishments and live one’s life with integrity make one’s life genuinely better. He holds that if a person’s critical interests are unsatisfied, her life goes worse, even if she does not understand what her critical interests are. Dworkin rejects the so-called ‘experience requirement’, which roughly says that in order for something to be good or bad for an individual at time T1, she must actually experience it at T1. Those who reject the experience requirement claim that the satisfaction of a desire can be good for an individual even if the individual does not even know that her desire has been satisfied.28 29 The expression ‘experience requirement’ originates with Griffin.29 If the experience requirement is inaccurate, rejection of it would contribute to explaining how unconscious humans can have interests while they are unconscious.

Imagine a fully competent individual: ‘P1’. She lapses into temporary unconsciousness at some stage in her life. Call her ‘P2’ during her unconscious period. Applying Dworkin’s account, we could ascribe P2 critical interests but not merely qua being P2. Dworkin’s account suggests that P2 has critical interests only through her connection to her previous competent self P1. As temporarily unconscious, P2 will relapse back into her competent self P1. P1 before and after she became unconscious has critical interests, and P2 has critical interests only because of P2’s ‘connection’ to P1. Dworkin’s discussion of permanently unconscious, ‘vegetative’ patients illuminates how his views could apply to temporarily unconscious individuals as well: ‘When we ask what would be best for him [the vegetative patient], we are not judging only his future and ignoring his past. We worry about the effect of his life’s last stage on the character of his life as a whole, as we might worry about the effect of a play’s last scene or a poem’s last stanza on the entire creative work’.28 The critical interests that the individual expressed while she was competent should guide our treatment of her while she is incompetent, because the preferences expressed by a competent person indicates what holistic structure, or overall character, she wishes for her life as a whole to have. Knowing the individual’s previously expressed desires helps us understand the holistic structure she wishes for her life to have and to understand what treatment of the unconscious patient would preserve it.

Dworkin’s and Tooley’s accounts seem equally puzzling. A mere claim that an unconscious patient ‘has’ interests while she is unconscious, in virtue of her previous interests, does not explain why she has interests while she is unconscious.

Hawkins recently30 defended the claim that some incapacitated individuals, qua incapacitated individuals, can have critical interests. Hawkins focuses on demented humans and does not apply her account to foetuses and premature babies; we will suggest that her account is relevant for these beings, but not take a stand on whether Hawkins would want to apply her account to these beings. We will argue that her account is the least puzzling explanation of how unconscious patients can have interests, and, if we accept her account, then we should accept that the Substituted Judgment Standard coherently applies to unconscious patients, even if they have never been conscious. She aims at establishing the ‘Nonalienness Principle (NA)’:

It is a necessary condition of X’s being intrinsically good for A at T1 that either (1) A respond positively to X at T1 if she is aware of X at T1 or, (2) A be such that she would respond positively to X at T1 if she were aware of X at T1. (p. 527, italics added)30

Interpretion and defence of NA

Hawkins holds that NA could be read as indicating either that X is intrinsically good for A now because she would respond positively if she were aware of X, or that X is instrumentally good for A now because she will respond positively to X once she becomes aware of X. However, Hawkins holds that in order for X to be good for A in either of these senses, X must actually be experienced by A at some point (p. 530).30 We believe that the assertion ‘X is good for A if she would respond positively to X if she were aware of X, and will respond positively to X at some point’ provides a less puzzling explanation of how A can have interests while she is unconscious than Dworkin’s suggestion that her interests while unconscious can be determined by the interests of her previous competent self, or Tooley’s stipulation of conceptual capacities for having interests. We cannot argue conclusively for this claim and merely suggest that NA is attractive because the claims that A would respond to X in some way were she conscious, and that she will respond to X in some way once she regains consciousness, are not metaphysically curious and are fully comprehensible. NA also avoids all of the challenges involved with explaining how P2, while unconscious, can have critical interests because her competent self P1 had critical interests. It would be fully comprehensible to claim merely that the critical interests of P1 should be honoured once P1 lapses into her unconscious self P2. However, the claim that P1’s critical interests should be honoured by treating P2 in certain ways does not imply that P2 has any critical interests. The critical interests of P1 cannot be assumed to be the critical interests of P2 simply because P1 and P2 are the same organism. We will suggest that examples advanced by Hawkins could be considered as providing intuitive support for NA, but note that opponents of NA could advance these, or other, examples in favour of their view. We conclude that both positions have intuitive support that cannot be conclusively quantified and compared.

Hawkins describes an individual who has asymptomatic pancreatic cancer and who is unaware of his cancer. Applying NA, we could say that something intrinsically bad has happened to him already, because he would be devastated if he knew about the tumour, and he will learn about it and be devastated at some point. Alternatively, NA would allow us to say that something instrumentally bad has happened to him already as the tumour will cause something intrinsically bad and that something intrinsically bad will happen to him once he responds negatively. Also, being loved by one’s relatives although one is incapable of ever knowing of it is not, according to Hawkins, a ‘good’ in any sense. However, proponents and opponents of unsensed goods could use the same examples to support their view, which suggests a methodological challenge.

Any of the following claims is coherent: if a patient is being mistreated while unconscious: (1) something instrumentally bad happens to the individual while she is unconscious and something intrinsically bad will happen to her once she learns about the incident; (2) something intrinsically bad happens to the individual while she is unconscious because she would have responded negatively had she known about the incident and she will respond negatively once she learns about it; and (3) something intrinsically bad would have happened to her had she been conscious when the molestation took place and that something intrinsically bad will happen to her once she learns about the molestation. An interest in physical integrity could arguably be classified as a critical interest.

Accepting NA implies that the Substituted Judgment Standard coherently applies to unconscious patients who have never been conscious

A capacity for having critical interests implies a capacity for reflecting over one’s life as a whole. Hence, an unconscious person would also uncontroversially be capable of making the choice not to be intrusively intervened with while she is unconscious, and once awake, she may hold that she would have made this choice had she been conscious. Everything NA claims about under what circumstances an unconscious adult can have critical interests while she is unconscious also establishes that the individual can have a dispositional capacity for exercising choices under the circumstances NA mentions. If the individual would be capable of having some critical interests were she conscious, she would also be capable of exercising some choices were she conscious. If this is accurate, we may claim that the Substituted Judgment Standard coherently applies to unconscious adults who have never been conscious.

Dispositional choices

Never-competent patients such as premature infants and formerly competent patients whose preferences are not easily identified are relevantly similar to temporarily unconscious adults and, therefore, the Substituted Judgment Standard applies to them as well. The claim that an individual would have certain interests or a capacity to exercise choices if she were conscious, and that she will have certain interests and capacities once she becomes conscious, does not presuppose previous consciousness. Even an individual who has not previously been conscious would have some interests and be capable of making choices if she were conscious. A premature infant would be capable of making the choice not to expose herself to certain kinds of degrading touching if she were capable of autonomous agency and will, once she becomes aware of it, hold that she would have chosen not to be exposed to such touching. If this claim is adequate, the never-competent patients and hard cases of former competent patients can also have capacity for exercising autonomous choices in a dispositional sense. Previous interests and choices have merely epistemological relevance: they can indicate, as a validity check, what interests or choices should be ascribed to the temporarily unconscious adult in a dispositional sense.

Does this argument imply the implausible claim that we can coherently talk about what a mouse or a fork would autonomously choose were it to become conscious? And if what any entity E would choose were it to become conscious determines how we may treat E, then we are committed to moral restrictions on how we may treat any entity E whatsoever. One might claim that premature human babies have, as a biological matter, the potential for autonomous choice while mice and forks do not. So restrictions on treatment would not apply to mice and forks. However, (1) not every premature baby has this biological potential and (2) the conceptual weight would then be put on the moral relevance of the potential for autonomous choice. The moral relevance of such potential remains controversial. It cannot simply be assumed. Our response to this prominently discussed challenge is the following. We focus solely on temporarily unconscious humans. We defend what we consider to be the least puzzling explanation of how such humans can have interest while they are unconscious. We saw that the least puzzling explanation of how such beings can have interests is that they have interests in NA’s sense. We argued that if they can have such interests, it follows that they can also have a capacity for making choices in this sense: a capacity for having critical interests implies a capacity for exercising choices. Now, never-competent patients who will become conscious if they survive and remain healthy are also temporarily unconscious. Since such never-competent patients are relevantly similar to previously competent patients in the sense of being temporarily unconscious, never competent patients also have dispositional interests and dispositional capacities for exercising agency. Beings who used to be conscious but are now permanently unconscious might be similar to spoons; those who will regain consciousness are not.

The importance of expressing a respectful attitude

Our defence for the practical relevance of our account comes in four related parts. First, our account can provide a new type of justification for choosing a treatment option. Referring to a patient’s dispositional choices as justification for a substituted decision expresses at least implicit respect for patient autonomy. People who share the defensible view that it is morally desirable to show implicit respect for the patient’s autonomy should welcome this justification. Implicit respect for autonomy is not the only, or necessarily trumping, moral value at stake in substituted decisions. However, the concept of dispositional choices allows us to incorporate expression of implicit respect for autonomy into substituted decisions, without loss with regard to other morally relevant considerations. Providing conceptual space for justifying one’s decision by reference to respect for the patient’s dispositional capacity to make choices does not prevent advocates of the Best Interests Standard to interpret the decision by reference to the patient’s best interests.

Suppose that a substituted decision maker recommends the healthcare team to withdraw life-prolonging treatment from a premature infant with Down syndrome and justifies this decision by referring to the infant’s dispositional interests: if the infant would be capable of responding positively or negatively in NA’s sense, she would not want to live her life with Down syndrome and also in chronic pain and a complex set of severe disabilities. Alternatively, she might justify her recommendation by referring to the infant’s dispositional choice not to live such a life.

We should acknowledge the distinction between being unable to know the content of the patient’s dispositional choices and holding that the claim that the patient can have a capacity for dispositional choices is incoherent. The former is an epistemological difficulty. The latter claim points out an alleged logical difficulty. By making this distinction, we have begun understanding the benefits of giving clinicians the conceptual opportunity of expressing the attitude of implicit and indirect respect for patients’ autonomy.

Second, our proposal serves as a structured mental checklist that might be a practical tool. The question ‘what would this particular patient have reason to choose if she were capacitated’ (let us label this question Q1) views the patient as a possessor of dispositional choices. Suppose a previously competent patient is conscious but has not had the relevant decisional competence for a long time, and has not, while competent, expressed any explicit preferences regarding her treatment should she lose decisional capacity. One proposed strategy for substituted decision making for such patients is attempted reconstruction of the patient’s ‘authentic’ preferences by identifying the patient’s life as a coherent ‘narrative’. This disputed concept has been interpreted in numerous ways. Here, we merely assume that the substituted decision makers consider previous choices and statements by the patient that might guide a reconstruction of what treatment she ‘would want’ in the current situation, if she were capacitated. Scholars have pointed out that this method is respectful because it recognises the patient’s individuality, and also recognises the profound difficulties of knowing what the patient ‘would have wanted’. Scholars have also noticed the risk of substituted decision makers’ imposition of their own views regarding what characterises ‘a good life’ in this reconstruction process. We suggest that viewing the patient as a dispositional ‘chooser’ throughout this reconstruction process helps us preserve a patient-centred perspective when creating such a narrative. This is because viewing her as a dispositional chooser reminds us to view her as a dispositional initiator of actions: as the source of whatever actions she might have engaged in had she been capacitated. Viewing her in this way reminds us to attempt to assume her individual perspective at all times. Viewing her as a holder of interests might fill the same function as long as we focus on her as a ‘holder’, without slipping into subjective evaluations of what interests such ‘holders’ should have. We may view her either as a dispositional initiator of actions or as a dispositional holder of interests. Both descriptions are fine as long as they serve the purpose of helping us view her as an ‘initiator’ or as a ‘holder’ and remind us not to impose our own views regarding what interests are weightier or what choices would be preferable according to the substituted decision makers. These descriptions could help us focus on what treatment option is most consistent with the patient’s previous preferences and choices. If substituted decision makers may choose how to view the patient (as a ‘holder’ or an ‘initiator’), this might help them remind themselves to focus on the patient’s own perspective rather than imposing her own preferences on the patient.

Third, if substituted decision makers may choose their preferred ‘mental tool’ for assessing their own attitudes, the ‘tool’ might be more efficient. A substituted decision maker who endorses the view that it is imperative to show implicit respect for autonomy might be more motivated to remain patient centred in her decision if she pictures the patient as a dispositional chooser. A substituted decision maker who endorses the view that protecting interests is imperative might be more motivated to remain patient centred in her decision if she pictures the patient as a holder of interests in NA’s sense. The substituted decision maker is then given some discretion regarding what description of the patient she considers morally relevant. However, allowing her to be subjective in this sense is consistent with a respectful approach because she may only choose among descriptions of the patient that are patient centred. Scholars have suggested that applying the Substituted Judgment Standard to these patients might take the focus from the substituted decision makers’ own preferences and help them focus on the perspective of the patient. We need empirical evidence to see how these two standards come into play in clinical decision making.

Robin Dillon suggests what all understandings of ‘respect’ have in common: ‘respect is a particular mode of apprehending the object: the person who respects something pays attention to it and perceives it differently from someone who does not and responds to it in light of that perception […] The idea of paying heed or giving proper attention to the object which is central to respect often means trying to see the object clearly, as it really is in its own right, and not seeing it solely through the filter of one’s own desires and fears or likes and dislikes’.32 Integrating (Q1) with clinical decision making may increase awareness of how one’s own personal and professional values come into play in the interaction with patients in shared decision making. Pragmatic aspects of the decision-making process, such as the need to make decisions quickly and the technical intricacy involved in many of these clinical decisions, might add further complexity to the suggested normative framework. Decisions in the neonatal intensive care unit, especially clinical decision making for the smallest babies with serious conditions, where the prognosis of survival and/or future life quality is very uncertain is one example where and technical complexity are involved. Pragmatic aspects of the decision-making process challenges any theory on substituted decision making, not just ours.

Fourth, if we are sceptical to the usefulness of the dispositional choice account, then we should be equally sceptical to Hawkins’ explanation of how unconscious people can have interests in NA’s sense. The two accounts are equally coherent and both have intuitive appeal. However, many people do find Hawkins’ type of explanation useful. If we are reluctant to jettison Hawkins’ proposal, then we should be equally reluctant to jettison the dispositional choice account.