Varied and Principled Understandings of Autonomy in English Law: Justifiable Inconsistency or Blinkered Moralism?

Abstract

Autonomy is a concept that holds much appeal to social and legal philosophers. Within a medical context, it is often argued that it should be afforded supremacy over other concepts and interests. When respect for autonomy merely requires non-intervention, an adult’s right to refuse treatment is held at law to be absolute. This apparently simple statement of principle does not hold true in practice. This is in part because an individual must be found to be competent to make a valid refusal of consent to medical treatment, and capacity to decide is not an absolute concept. But further to this, I argue that there are three relevant understandings of autonomy within our society, and each can demand in differing cases that different courses of action be followed. Judges, perhaps inadvertently, have been able to take advantage of the equivocal nature of the concept to come tacitly to decisions that reflect their own moral judgments of patients or decisions made in particular cases. The result is the inconsistent application of principle. I ask whether this is an unforeseen outcome or if it reflects a wilful disregard for equal treatment in favour of silent moral judgments in legal cases. Whatever the cause, I suggest that once this practice is seen to occur, acceptable justification of it in some cases is difficult to find.

Introduction

Autonomy is a principle that is generally agreed to have considerable importance in good medical practice,¹ and this is accepted at law. There is not, however, one generally accepted definition of the concept or the limits of its validity. I argue that it has been possible for judges to use the equivocal nature of the concept to try to achieve outcomes that they consider to be morally desirable in cases, rather than simply to apply a single concept to comparable legal questions. When various conceptions of autonomy are explicitly considered, it is possible to observe in the case law decision-making that reflects attitudes to the status of patients. Understood in this way, the variation in the judgments does not seem justified. This suggests a need to modify what the law allows or demands, or else that a limit should be placed on the scope of judicial interpretation of it.

The English Law Position

If we are to criticise the courts’ application of the law, we first need an account of the legal principles that are supposed to guide or bind them. It will be particularly instructive when looking at the approach taken to autonomy also to consider capacity, rationality, life-shortening decisions, advance directives, and the Mental Capacity Act 2005. There is some overlap between these, so it is not suggested that each can be perfectly separated from the others. What is important is that together they allow us to gain a fuller understanding of what the law says about autonomy. It should be noted at the outset that the courts do not claim that any one principle is always supreme. We can see, however, from an analysis of the case law that certain principles can be found to be supreme in different definable situations.

It is rare for a judge to provide an explicit, philosophical investigation of autonomy. This is both unsurprising and understandable. Perhaps it is even desirable. Many bioethicists seem to have made something of a religion of autonomy, and were judges to follow too assiduously a specific doctrinal approach, they may well be open to the sort of criticism that they are employed to judge the law and not questions of ethics. On the other hand, when judges make claims regarding the moral neutrality of their judgments, it is not hard in some circumstances to be rather cynical of their alleged capacity to divorce matters of ethical theory from matters of legal theory.² With regard to autonomy, we often find references to (at least what are thought to be) synonyms: ‘self-determination’ or the right to govern bodily integrity, or the like. There are occasional references to philosophical works—for example, Dame Elizabeth Butler-Sloss cites a paper by Kim Atkins in the case of Ms B [38]³—but we may doubt whether such works are intended to be brought into law, or even into an area of the law, with universal effect or application.

What is the legal effect of judges not treating autonomy with full theoretical rigour? We live in a time where increasingly concepts are boiled down to aspects of some right or other, and then are declared to be a right in their own right. Autonomy itself, for example, is now generally accepted to be contained in Article 8 of the European Convention on Human Rights, and may be argued for elsewhere in that convention [41]. It may still be suggested, however, that autonomy is not something that of itself we have a right to [47, Part III]. On this view, what we in fact have rights to are things that serve or enhance autonomy, autonomy being a good, but one that is defended by things that strengthen it rather than by having a right to it as a thing in itself. Whether a judge would be swayed by this analysis is really only of academic interest. We may assume that as time goes by we will read more and more dicta referring to the patient’s right to autonomy both at common law and under the European Convention. The effect of this is potentially that the courts will find themselves defending a concept of unspecific scope or meaning.⁴

The present law regarding autonomy may be summarised as follows. On the whole, judges talk of autonomy as being equivalent to self-determination.⁵ In English medical law, this allows patients to refuse any medical treatment or intervention, and to choose one of any treatments offered by a medical professional. The right is applied as one with a positive and negative side. The positive right is limited to one of choice between treatments offered: a patient may not demand that a treatment be provided if his doctors are of the opinion that it is not medically indicated [43, para. 31]. This is the compromise between a patient’s right to define his own interests and the doctor’s duty to act within his professional and ethical obligations. So, just as a doctor may not inflict a treatment on a patient—a patient’s interests involve broader issues than those that may be defined as medical and thus falling within a doctor’s remit—a patient may not demand that he be given any which medication or operation that he has found on Google—the defining of medical interests falls entirely within a doctor’s responsibility. Autonomy, then, is subordinated as soon as positive interventions outside of a doctor’s remit are required. For example, a patient may not demand life-prolonging or life-shortening treatment which is not medically indicated [41, 43, 45]. With regard to the provision of treatments, the patient has a right that his request for treatment be considered, but not that it necessarily be respected [46].

Some commentators observe a move to stronger respect for autonomy over recent years, with patients being entitled to higher levels of information to allow them to act autonomously, and being entitled to damages for not having received such information [11]. It should be stated that supporters of increased autonomy must not overestimate the relevance of such advances in the case law. In the case of Chester v. Afshar [10] for example, the damages reflected the physical harm done to the patient rather than the metaphysical harm done. In a case where a similar failure to inform occurred, but in which no physical harm resulted, it seems hard to believe that a court would allow damages for the harm done to the patient’s autonomy.⁶ Nevertheless, we do see judges increasing autonomy’s importance. Failure to respect a patient’s autonomy at the consent stage can lead to a patient receiving damages for harm caused as a result of an operation even if the operation is not itself negligently performed.

What of capacity and rationality? Given that most of us ascribe special moral status to beings with autonomy—either in the sense that we think such beings deserve special protection over other conscious beings, or in the sense that we think such beings are responsible or accountable for their contemplated actions—it is no surprise that we require a patient to have capacity if we are to believe his decision to be autonomous. The English courts have taken an approach to assessing capacity that is said to originate in the case of Re T [31]. There, Lord Donaldson held that whilst the requisite mental capacity to be able to make a decision varies depending on the gravity of the treatment and the patient’s medical condition, a patient has the right to choose:

whether the reasons for making that choice are rational, irrational, unknown or even non-existent. That his choice is contrary to what is to be expected of the vast majority of adults is only relevant if there are other reasons for doubting his capacity to decide [31, pp. 116–117].

This appears fairly unequivocal, and has been subsequently cited with approval many times. We can see from Re T that the law does not ask for a consequentialist approach to be taken in its assessment of decision-making. Given Lord Donaldson’s statement that a decision need not even be rational, so long as there are no other reasons to doubt the patient’s capacity, a patient is free to make any decision in line with the law as spelled out above. In the case of Re C [49], Thorpe J gave what has been seen to be the appropriate test for assessing capacity.⁷ He held that capacity could be found where the patient meets three conditions: first, he comprehends and retains treatment information; second, he believes it; third, he weighs it in the balance to arrive at a choice. This test will allow a finding of capacity in a patient who decides that no treatment is worth accepting, even if medically it seems only rational to consent. Combined with Lord Donaldson’s statement of the law, we see that so long as a patient’s thought processes are not in doubt, his underlying reasons are not open to question. They are irrelevant. If a patient thinks it is a bad idea to accept treatment on a Tuesday or when the sun is out or when the Welsh Rugby Team is barely able to win a game, so long as he meets Thorpe J’s three criteria, it is his right to refuse.⁸

The law holds, therefore, that a competent patient may refuse treatment even if this will result in his death. The law was applied to this effect in cases such as Re C ⁹ and the case of Ms B [38], and has been held to be so in many other cases. Whilst with incompetent patients judges generally apply values that err on the side of life [28, para. 23], competent patients are unencumbered by any external value system at law, and may choose death as well as life when making treatment decisions. As Munby J has well illustrated, what is right for one person is not necessarily right for another, and each individual is best placed to assess what is right for him [42, para. 97].¹⁰

English law’s respect of the right to choose extends to advance decision-making. Currently this is governed by rules developed at common law. They are clearly expressed in the case of HE [28]. Here, Munby J held that, so long as a patient was competent at the time of making an advance directive, it would be binding on medical practitioners. Of course, an advance decision may be revoked, and absent formal requirements, medical practitioners may be reluctant to accept that an advance refusal of life-sustaining treatment is valid unless it is absolutely clear that the patient envisaged circumstances such as those that have materialised. As far as the law goes, though, a patient may at present make an advance refusal of any treatment in any way he chooses, and that will be binding on medical practitioners [29].

We should briefly mention the Mental Capacity Act 2005, although this is not in force at the time of writing. This Act will codify the common law position on advance statements, with a few modifications. Most important here is that an advance refusal of life-sustaining treatment must be made in writing, signed, and witnessed. It must state that it is intended to refer to treatment that is needed for continued life, and must relate to the circumstances that eventuate.¹¹ Whilst this is, in a formal sense, a tightening of the law, we may note that Parliament is perpetuating the liberal position that currently exists in society; namely, that an adult may choose what treatment he will and will not receive for any reason or no reason whatsoever. In their assessment of capacity, the Act prohibits medical practitioners from taking an unusual decision to imply incapacity. In English medical law, adult patients are not obliged to accord to any doctrine they do not accept, so long as they can do so without the positive intervention of a third party. Where a third party intervention is required, of course, the exercise of autonomy may be hampered by criminal, civil, and public law factors.

Three Philosophical Understandings of Autonomy

Whilst there are various possible definitions, [12, p. 6; 40, pp. 21–27] there are three important understandings of autonomy that may be held within a liberal society.¹² For a full philosophical taxonomy, these may be further dissected and subcategorised. For the purposes of this paper, however, the three are sufficient to gain a good understanding of the theoretical application at law. The three provide a framework with which to assess the law that is not achieved when we merely adhere to one conception of autonomy for whatever reason. Theory must meet practice, and one practice that we must not forget is the inconsistent application of theory. The three definitions of autonomy that I use here are important because it is their coexistence that allows confused and inconsistent resolutions of cases to arise in an ostensibly respected system of precedent. None of the understandings is controversial as a theory, although there is clearly great disagreement amongst commentators regarding which conception best reflects the way the law ought to be applied. For an action to be considered autonomous, then, I propose these three heads in accordance with which it may be judged:

• 1. Ideal desire autonomy—Leads to an action decided upon because it reflects what a person should want, measured by reference to some purportedly universal or objective standard of values.

• 2. Best desire autonomy—Leads to an action decided upon because it reflects a person’s overall desire given his own values, even if this runs contrary to his immediate desire.

• 3. Current desire autonomy—Leads to an action decided upon because it reflects a person’s immediate inclinations, i.e. what he thinks he wants in a given moment without further reflection.

Let us consider each of these in a little detail.

Ideal desire autonomy is compatible with a Kantian or neo-Kantian conception of autonomy [33]. As such, it requires an agent’s decision-making to accord with some objective set of ideals. A good example of a contemporary account of autonomy that would fall under this head is Onora O’Neill’s ‘principled autonomy’ [40]. This theory holds that whilst it is important that we be in control of our decision-making, we must not try to imagine ourselves in the untenable vacuum that is sometimes implied by individualism. Ideal desire autonomy requires agents to consider their reason for acting, and only to pursue a course of action if it could be made a universal law. That is, if it could be a successful maxim for all agents to follow. Therefore, if a person chooses to act in a way that is incompatible with a universalisable theory, that person is not acting autonomously. Also, ideal desire autonomy might include simple reference to a purportedly objective system of ideals that for some other reason would lead to an agent’s being considered ‘irrational’ were he to ignore it. On this account, obedience of the will does not necessarily equate with acting autonomously: autonomous action requires ‘responsible decision-making’.

Best desire autonomy is akin to the conception of autonomy advanced by Harry Frankfurt [19] and Gerald Dworkin [12, chapter 10]. These commentators famously distinguish ‘first-’ and ‘second-order desires’, and hold that being able to act in accordance with second-order desires is what makes a being autonomous. On this account, we find a person to be autonomous if he acts in accordance with his own value system. This will sometimes require him to act against his immediate inclination, but differs from ideal desire autonomy because the values that command an action inhere in the individual. They may be selfish, self-destructive, or subject to some other condition that would make them impossible to hold as a universal. They are, nonetheless, settled—although not necessarily permanent—and an agent recognises them as his values, and seeks to act in accordance with them.

Current desire autonomy looks close to an agent’s ‘first-order desires’. It may refer either to a person’s impulsive desire, or to a person’s desire that is settled and lasting but on which he has not reflected.¹³ When we say that someone is acting in accordance with his current desire autonomy, we suggest a level of conscious choice, but one that is not very (if at all) reflective. If it is reflective, it nonetheless succumbs to the call of the moment, even if that may be a matter of contemporary regret for the agent—i.e., even if the agent would not wish to be subject to the desire.

We need more than one conception of autonomy for our analysis precisely because this helps demonstrate how certain judgments can claim to be in accord with the law of a liberal society at the same time as achieving an end that is at odds with that in other judgments. We have already seen that the law as stated by Lord Donaldson in Re T allows a patient to make a decision for any reason, rational or irrational, or no reason at all. This appears very much to be encompassed by current desire autonomy. This might be surprising, although we may remember that in our analysis we are predominantly considering negative rights regarding medical treatment. We are looking at a minimum requirement of the law. When it comes to demands for positive interventions, we have seen that the courts seek to be more restrictive over the exercise of autonomy. Lord Donaldson’s test also allows for the exercise of best desire autonomy. When a patient wants to act on a settled desire, that too is his right, provided the treatment is otherwise lawful. We will see below that judges can be sympathetic to a patient who—for example because of a needle phobia—wants to have his current desire autonomy disregarded in some circumstances, in order to achieve his preferred end. Finally, there are circumstances in reality where necessarily an ideal desire autonomy decision must be made on behalf of a patient. For example, because he is permanently without capacity, or because he has temporarily lost capacity but we have no way of knowing what he would want.

It is no surprise that all three should persist as tenable explanations of autonomy, and not just because different patients present different needs. Consider very briefly a couple of objections amongst theorists. Some might argue that current desire autonomy would never be consonant with autonomy: it would make us slaves to whim and emotion. Yet Petr Skrabanek was clear that to his mind, an agent could not be acting autonomously by ‘tying himself to the mast’ [60, pp. 185–190].¹⁴ Dworkin, on the other hand, would hold that sometimes a person may tie himself up if he is to act autonomously [12, p. 42].¹⁵ Many liberal philosophers would doubt the validity of ideal desire autonomy: for one thing, they may doubt the authority or true objectivity of the norms that may be dictated by the ‘ideal’ desires [64]. Perhaps more fundamentally, they may object that a person is by definition not acting autonomously if he must go against his will, or has no will: arguably there is very little philosophical tenability in ‘constructive autonomy’. Even if (as seems most plausible) autonomy and freedom are not synonymous [12, pp. 105–107], the restriction of a Kantian conception might seem too much. There may be something to this, but we must remember that there are also strong arguments against best and current desire autonomy truly creating autonomous action, or action that can always safely be judged as autonomous [1, 36, 55].

So, we see why we need the three conceptions. We will now consider what respect for autonomy may require, what justifies a system that values autonomy, and what justifies overruling autonomy. We will then be able to analyse the English case law, and obtain an understanding of the application of the law, looking at possible motivations or trends that should not really feature in the types of judgments that we are considering.

Why Respect Autonomy and How?

The purpose of this paper is certainly not to argue tacitly that we should offer autonomy unrestricted respect. If we consider valuable autonomy (under whichever conception) to involve a practical application and not be a mere exercise of the will then clearly it is not attainable in human beings in any absolute sense. This does not prevent us from attempting to maximise autonomous action, but we must remember that within a society maximum overall autonomy will be best obtained by placing restrictions on each person’s autonomy. Even in an anarchist state, autonomy will be limited by others’ exercise of their autonomy. For organised society, it is considered best to have a legal system that restricts many of our actions. In a way, then, human functioning within a liberal society can be considered to involve a combination of all three conceptions of autonomy, with compulsion also added at times: for example, a prison sentence may involve actions with no exercise of autonomy. But why should we respect autonomy at all? A full answer to this question is well beyond the scope and the purpose of this paper. A brief answer should be given, however.

First of all, of course, there would be practical difficulties with formalising and regulating every aspect of life. Even a religion or state system that is considered to be strict allows in fact for much actual free choice—albeit vastly insufficient for many observers. A further reason is usefully caught by John Harris when he says “it is autonomy that enables the individual to ‘make her life her own’” [27, p. 10]. This reason is more interesting and important to a theorist, and it captures a sentiment that may be ascribed to all who think there is moral relevance in human action, not just those who share Harris’s views on autonomy. Autonomous choice is what enables us to be held to account for what we do, and what enables us to take, for example, credit for or pride in our actions. This is central to the importance of exercising and acknowledging free will. Autonomy, though, while a good, must be limited by just authority in an organised society. Mill’s ‘harm principle’ may be invoked as the simplest expression of the justification for limiting autonomy. We may limit the exercise of one person’s autonomy if it is necessary to protect the well-being of others [37, p. 52].¹⁶

Within medical law we are considering some, not all, means of exercising autonomy. That which is least easy to justify breaching is the right of an individual to refuse any medical intervention. This may be contrasted with its correlative (and largely non-existent) right to demand or receive positively administered harm; even if willingly administered. Respect of refusals, we may hold, accords with the harm principle. We may also assume that a patient’s right to choose any of a selection of treatments offered to him accords with the harm principle. Of course, neither of these assumptions is water tight. With regard, for example, to a suicidal refusal of treatment, this may well harm others: for example, dependants or those with close emotional links. On a more extreme view, possibly society at large is harmed by suicide. And with regard to treatment choice, a patient seemingly unnecessarily taking a more expensive option out of a list of treatments offered may be depriving another patient or patients of valuable resources. These sorts of considerations are important, and worthy of their own separate discussion.¹⁷ I suggest, however, that they do not establish sufficient justification to restrict this exercise of autonomous choice.

We see that generally to exercise autonomy is good. Further, we see that a limit to that exercise inflicted by society needs justification. Thirdly, we see that where we can respect it, we do so by allowing people to make lawful decisions. Does this require confusing law and morality? It does if applied uncritically—consider the ill-ease invoked by a defence of ‘I was just following orders’. However, if we assess the law and question it and push for reform where it is bad, the law provides a bare minimum that we may expect of each other (we may, of course, choose further to limit our autonomy through additional codes, be they religious dogma or etiquette or whatever) [9]. Autonomy may then be exercised through choice; the choice of anything that may lawfully be done.¹⁸

Which Conception Should We Prefer?

An action measured against any of the three understandings of autonomy listed above—ideal desire, best desire, and current desire—could often provide the same answer. For example, my decision not to murder someone will normally accord with all three. But there may be occasions when the decision does not accord with the third. Are we to suggest that a suppressed homicidal urge is an example of a failure to exercise autonomy? Perhaps. It depends how we define autonomy. Under the current desire understanding it could be.¹⁹ Under the ideal and best desire conceptions autonomy would still be exercised. In a secular, liberal democracy a person should be able to live by his own values, so long as this does not breach the harm principle. Yet some might argue that current desire autonomy is sometimes too loose an understanding of the concept, holding that it is right that the law should allow each individual to base decisions on his own value system, but in some situations on a best desire autonomy understanding.

In fact, many commentators have argued in favour of one conception of autonomy over another.²⁰ The arguments vary in strength, and sometimes relevance [23]. I suggest that it is a simple reality that all three understandings given here have a place in our lives, and more specifically in the functioning of good medical practice. For a person who has always had a severely limited mental capacity, ideal desire autonomy certainly has a place, although choosing the ideals remains a controversial matter. Consider the contrasting ideals of two eminently rational beings—John Finnis and John Harris—and you see the potential for (philosophically) irresolvable conflict in the search for ‘right’ answers [16–18, 24–26]. Furthermore, there are examples in society of instances where people are questionably deemed not to be capable to make decisions, and are consequently subjected to the rule of a third party’s conception of what they should want. An example, which we cannot consider here, but should note, is the validity of the courts’ treatment of children whose refusals would almost certainly result in death [50–53]. The decisions in some way accord with more general societal norms and practices that lead many to believe that we are right to maintain in the UK an arbitrary distinction between people who have lived outside the womb for 6,573 days and those who have done so for 6,574.

If we limit our analysis to cases concerning adults, is one conception of autonomy to be preferred? We know that ‘competent’ people are capable of simultaneously wanting different and inconsistent things. A common example is the heroin addict who wants to stop taking the drug, but also is suffering a craving. In such a case, we see that it is possible to help him in a way that can be said to respect his autonomy either by giving him the drug or by refusing to give him the drug. Following Harry Frankfurt’s analysis involving ‘second-order desires’, it is probably right to say that the preferable understanding of autonomy in this situation is best desire autonomy, and thus respect for autonomy would be achieved by refusing to give the addict a fix. We therefore find a legitimate means of overruling an individual’s expressed will that is compatible with a liberal understanding of autonomy. This is relevant to everyday life. We all encounter conflicts between our best desires and current desires: examples might include enrolling in a pension scheme versus spending more now; deciding whether or not to buy cigarettes or alcohol; choosing whether or not to eat fatty foods.²¹ Faced with such conflicts of our own interests, we may well think autonomy is best exercised if we soberly place limits on the exercise of our current desire autonomy. If this practice were to be applied honestly, it would allow a patient to set aside in advance a future refusal that would be based, for example, on an irrational fear of needles, but would allow another patient to insist that a fear of needles be allowed to form the basis of a refusal.²²

It may be argued that a decision can only be autonomous if it is made in full knowledge. This is a really just a means of limiting action: sometimes an underhand means. No one has perfect knowledge, and a demand for it permits those with power to close in when a person is trying to make an unusual decision. A Jehovah’s Witness, ‘for his own good’, may be subjected to a description of the nasty death that awaits him as a result of his refusal. It is not the case, however, that an atheist who is in medical need of a blood transfusion is generally considered to need to hear and reject the views of a Jehovah’s Witness before he gives a valid consent. Advocates of autonomy accept that it is a graded concept [34, p. 69]. Arguments that (perhaps conceitedly) claim to defeat autonomy as a principle because they show it to be imperfect in practice miss the mark. They do not justify every case of inflicting another’s moral code on an individual, and they do not fatally undermine the arguments in favour of allowing people to make their own choices based on their own reasoning.

From what we have seen above, we can conclude that no conception can be universally preferred. We see that ideal desire autonomy has only a limited legitimacy in this area, and that best desire autonomy is preferable for serious decisions. We are left with some problems. First, a conflict can still emerge with regard to the choice between best and current desire autonomy. Second, there are instances where it may be considered unconscionable to respect a refusal of treatment because it seems palpably in breach of some obvious kind of ideal desire autonomy.²³ In short, the bounds of the legitimate use of the conceptions are unclear. This does not provide an excuse for any wrongful application that can be seen to occur, of course. In practice, as I show below, judges are inclined to assess cases differently depending on the quality of the person whose autonomy is in question. This is why, I argue, we need to guard against inappropriate tacit selection. Not to defend one conception of autonomy to the hilt, but because appropriate choice between the different conceptions in different circumstances is what marks the difference between legitimate and illegitimate application of the law.

The Application of the Law in the English Courts

It may be suggested that the law limits which refusals receive respect by reference to a system that equates capacity with choosing the best outcome, where ‘best’ is defined by a judge or medical practitioner attempting to be objective [4, p. 76] (thus stamping a huge question mark beside the right to refuse for an irrational reason or no reason at all, let alone any reason). This suggestion is fair, but it fails to tell the whole story. I will consider it briefly, before looking at two other factors that can affect a judge’s inclination to respect or reject a patient’s desire to refuse treatment. Despite some overlap, these are more easily understood when considered separately. They are the quality of the decision-maker and the quality of the motivation for the decision. These should not affect the value of a patient’s refusal (according to the existing legal position), but case law suggests that each in fact can and does.

Rationality of the Decision

Largely, we see the rationality of a decision is judged by reference to its consequences. This may be severely offensive to those who attach importance to specific acts or decisions because they are of themselves right or wrong, without regard to what follows from them. Furthermore, a consequence-based assessment requires an understanding of better and worse outcomes, and in some cases the quality of the outcome is judged as if in accord with some objective scale. The law states that the right to refuse is absolute,²⁴ but is it? Even if we ignore the fact that a person is supposed to have a right to refuse a treatment irrationally, it would be a profound insult to personal autonomy if a decision that is based on a settled value system were overruled. Nevertheless, the courts can be seen to condone great pressure being exerted upon people to reject their values.

The ‘absolute’ right to refuse treatment is somewhat corroded by the fact that ‘capacity’ to decide is a graded concept [31, p. 113]. There are compelling arguments in favour of requiring a person to be able to decide, but there should be a heavy presumption that the assessment of capacity is not made by bringing in a value-system that is different to that of the individual concerned. Clearly, once the metric used for evaluation is changed, the weighting of the outcome will change too. The law does not treat life as a supreme value [2]. In Re T, it was felt that the values of a Jehovah’s Witness might not truly have inhered in the patient. This is not problematic of itself, but it is clear that the values of that religion were felt to be bad values. We risk here making ourselves subject to the inaccuracy of judges’ ‘moral compasses’. This is not a fatal problem—besides, there may be no better alternative—but we must be alert to it.

Quality of the Decision-maker

Here I suggest that a patient whose refusal has raised questions has the greatest chance of exercising best desire autonomy where he enjoys the judge’s empathy or sympathy. Furthermore, where a judge is unsympathetic with the patient, it is possible that the patient faces what can amount to an indifferent or even punitive understanding of autonomy. This is because of the various conceptions of autonomy listed above, which are deployed as if each amounts to the same thing.

Here we see the insidious utility of the three understandings of autonomy. Their coexistence is what allows the different treatment of different individuals to retain a sheen of consistency. Let us look at some cases to support this view. First, ideal desire autonomy being used because of a sense of care. The decision to reject the refusal in Re T seems to fall into this category: it was not clear that T truly held the values of a Jehovah’s Witness, so a blood transfusion was considered likely to be a benefit to her. Ideal desire autonomy also appears in the case of B v. Croydon HA [5], where Thorpe J allowed a patient to be fed against her wishes because this was in her best interests. This finding was permissible in spite of a finding that she had capacity, because the feeding was found to constitute “medical treatment given... for the mental disorder” that B was afflicted by, and could thus be provided in accordance with section 63 of the Mental Health Act 1983.²⁵

The 2004 case of Ms T [65] is a more recent example. In this case, Charles J was asked to decide whether the claimant’s advance directive refusing blood transfusions or iron, which was drawn up with a solicitor, should be respected. Ms T’s reason for refusing was that she considered her blood to be evil, and that any transfused blood given to her would become evil. She had no psychosis, but was found to be suffering from an untreatable borderline personality disorder. The judge held that her beliefs could be taken on their own to demonstrate a lack of capacity, and therefore her will was not respected.

If, however, a judge can sympathise with a patient, a refusal may be upheld even if the judge would find the decision regrettable. Ms B [38] is a good example of this. And many cases that raise similar principles but in less stark factual circumstances will not need to go to court. There is judicial statement that supports the requirement to respect refusals of treatment in some cases and not in others, even if medically the cases are identical [59, pp. 885–886; 42, para. 97]. This is based on the freedom of the patient to evaluate his position. If the patient’s decision were overruled in cases where judges (and other people) could sympathise with his desire that treatment cease, it would now be considered a cruel despotism [13, p. 127].

So, we can see the courts being careful to select the ‘right’ form of autonomy to reach the decision that care and sympathy would require. In Re MB [30] the Court of Appeal ruled in favour of what amounts to best desire autonomy. This was necessary in order to prevent a pregnant woman’s current desire autonomy (which was dictated by her fear of needles) to preclude her chance of receiving a caesarean section. It should be mentioned, though, that it was held that if a fear of needles accorded with a patient’s best desire autonomy then forcing an operation would be unlawful.²⁶ There is room for some overlap here between what care would demand and what sympathy would allow. It is apparent, however, that distinct approaches to grounds of decision-making are taken in order ‘best’ to ‘help’ the patient.²⁷

The practice described so far in this section can be justified (at least to some degree), mainly because of its benevolent nature. What, however, if there is no concern that the patient receive equivalent respect for ‘well-being’? I suggest that there are cases where this can be discerned. It is perhaps telling that these cases concerned patients who were imprisoned. In the judgments in Re C [49], Robb [58], and In Re W [32], a level of indifference regarding what is meant by respecting autonomy presents itself. In ex parte Brady [44], where autonomy was considered,²⁸ the judge reasoned around having to afford it the absolute status that the law is held to require. Let us take them individually.

Re C concerned a paranoid schizophrenic who had been imprisoned for stabbing his girlfriend, and had subsequently been moved to Broadmoor; a secure hospital. It was found in this case that Mr C’s likely suicidal refusal of an amputation of a leg should be respected. The judge found that it had not been shown that the patient lacked capacity to understand the treatment, and the refusal was not linked to his mental disorder. In Robb, it was held that prisoners retained sufficient of their right to self-determination to allow those with capacity to refuse nutrition and hydration. In Re W concerned a Category A prisoner at a high security prison who had been convicted of murder and aiding and abetting a suicide. He had cut his leg open and continued to introduce to the wound foreign objects in order to “turn it septic”. He refused any treatment for this because he sought to manipulate the authorities; he wanted to be moved to a special hospital to be treated for his psychopathic disorder. He had also inserted two taps into his anus, which he now wanted removed; a procedure that required surgical intervention. When this procedure was to be performed, he had refused it at the last minute because of his fear of having an injection in his arm. The judge noted this fact, and his enduring willingness for the taps to be removed, the fear of needles notwithstanding. Mr W claimed that his motives for self-harm were linked to a conspiracy theory, but the judge said she doubted the truth of this.²⁹ Ex parte Brady concerned the notorious ‘Moors Murderer’, Ian Brady, who was detained in a secure hospital and feared returning to the normal prison system. He had gone on hunger strike, and the court was asked to rule on the lawfulness of force-feeding him. The case was decided on mental health grounds, although the judge considered the question of capacity too, and found that Brady failed the common law test for capacity.

It is possible to contrast the approach to the patients in these cases with the others we have considered. As above, we see attachment to the different forms of autonomy via distinct understandings of capacity. The crucial difference in the first three is the reluctance of the judges to overrule whichever form of autonomy would be respected. In Re C, Thorpe J seemed to respect the patient’s best desire autonomy. But if Mr C’s social circumstances had differed, it is wholly imaginable that ideal desire autonomy would have been (tacitly) invoked, and he would not have had his expressed will respected. This is intuitively acceptable if we consider a Mr C in the position of the bloodletting Ms T in 2004, in whose case equally ‘disordered’ thought was taken to demonstrate the incapacity to act autonomously. The same could be said of Robb’s case, although his reasoning may be perceived as more rational. The specific means of overruling autonomy that were employed in B v. Croydon HA would have been unavailable in his case, of course. Nevertheless, we see this person being allowed to define good and bad consequences without this bearing on an assessment of his exercise of autonomy, even though that does not accord with other of the decisions. I see no convincing reason for the distinct judicial treatment other than the judge’s distinct evaluation of the person.

In Re W is the most interesting case here because it offers material for analysing all three conceptions of autonomy. We are bound to be unsympathetic towards the manipulative technique the claimant tried to use, but it is noteworthy that his actions cried out for an ideal desire autonomy understanding that would have defined him as mentally unsound: though probably not based on a strict analysis of case law and medical practice, the conduct was an effort to imply a definition of Mr W as someone non-autonomous, and therefore someone inappropriately kept in a prison. There are clearly compelling public policy reasons for not encouraging a jailed person to hold himself to ransom.³⁰ But what of the judge’s neglect to mention the option of an exercise of autonomy in line with Re MB? It may simply be put down to such an action being supererogatory. But it is not, we must assume, right (legally) to suppose that supererogation drives a judge to select an alternative understanding of autonomy in cases where this is done for benevolent purposes. If it is not supererogation that influences the selection of autonomy in cases where a judge sympathises, this suggests that in W the decision contains some sort of punitive or discriminatory message, based on the quality of the claimant.

In Brady, the court did not treat the applicant’s decision with indifference, and instead enforced an ideal desire autonomy ruling on mental health grounds. I suggest, nevertheless, that this does not bring it into line with B v. Croydon HA. The judge’s ruling suggests a different attitude towards the patient, and it does not seem too cynical to submit that the judge had no intention of allowing Brady to achieve his ends.³¹ This view is reinforced by the judge’s finding that the patient had no common law capacity to make the decision. The attitude of the judge is well captured in his comment made (per curiam) at the end of his judgment:

[I]t would seem to me to be a matter of deep regret if the law has developed to a point in this area where the rights of a patient count for everything and other ethical values and institutional integrity count for nothing [44].

The judge seems to think that sometimes what is an absolute right at law³² should be disregarded in the case of certain classes of patient. Whilst this view might please sections of the public, I see no good ground for the differential treatment that it implies. It amounts to the meting out of a punishment out of accordance with law. A person’s status as a convicted criminal may make us seek to limit what manipulative power he has, but we can not do this by inventing arbitrary exceptions to fundamental rules of law.

Quality of the Motivation for the Decision

In some sense, this consideration appears as if it should be welcomed. It apparently contradicts the problem with consequences of a decision being the measure of its wisdom (and thereby acceptability). The contradiction starts to evaporate when this consideration is read with the analysis regarding the quality of the patient. As we saw above, where a judge is able to sympathise with a patient’s reasons for choosing a treatment, this may alter the judge’s propensity to reject, for example, a suicidal refusal of treatment, even if the judge thinks the decision is not the best one.

Where the decision is motivated by a desire that is perceived by the judge to be reprehensible, it may be the motivation that defines the rationality. Here we can contrast what the court may be seeking to achieve. In Brady, the judge was seeking to find the applicant not to be competent, so questioned the motivation behind his decision to go on hunger strike. At paragraph 51, the judge seemed to hold that Brady’s best desire and ideal desire autonomy would not be served by his refusal of nourishment. The judge (rather imperfectly) dichotomised Brady’s ends. We would clearly not like to see manipulation by prisoners of the system, but it is hard to escape the feeling when reading the judgment in Brady that it was felt that justice would not have been served were the prisoner allowed to die. In the case of W—a much less notorious criminal—on the other hand, the prisoner’s reason for harming himself was to improve his circumstances in an inappropriate way, and this provided evidence that he was a rational actor. His harm was done in pursuit of a higher good, as he saw it: the problem seems to be that it was a good that was offensive to society. The risk that he might die as a result of his attempted manipulation did not, therefore, seem to warrant invocation of best or ideal desire autonomy.

Yet Ms T in 2004 also sought a higher good. The harm she caused herself was done for the best as she saw it, and was by that measure rational. A combination of care for her and disrespect for her ability to be in possession of a worthwhile premise caused her best desire autonomy to be flouted. The first part of this combination did not arise in the first three of the cases cited here involving prisoners, and there therefore seemed to be indifference as to whether or not the premises in those cases were worthwhile. Where in the Brady case it was necessary to reason around accepting the patient’s autonomy, the judge did so and achieved the end least offensive to society. We should note that it is striking that Ms T might have been right—her reasons were metaphysical and beyond proof. Mr W, by contrast, was certainly wrong. His reasons were found to be based on an error that could be proven—i.e. his action was not going to have him taken to a special hospital. Nonetheless he was allowed to continue to harm himself because he understood the implications of what he was doing; that it could kill him. Ms T also understood this, but it served her no respect. This can be explained if we see Mr W being punished for having a reprehensible motivation. We also see a punitive element to the decision in the Brady case.³³

Justifiable Inconsistency or Blinkered Moralism?

We see an application of the law that does not adhere perfectly to its precepts. We must ask whether this matters. If the variation in judges’ approaches described here is sometimes inadvertent, do we have cause for concern when it can be shown to happen? I would submit that we do. The principles that support autonomy make loathsome the principles that support state (or third party) interference with a person’s exercise of free will. And here, we see what does not merely amount to pure ‘benevolent paternalism’, which could at least conform with ideal desire autonomy. What emerges in some cases is a sense of overactive sympathy and in others one of just deserts. It is wrong for people with settled, considered value-systems, to be treated against their will. It is also wrong for medical treatment to be used as a tool of punishment (either in its provision or otherwise). It is wrong because it is not in accordance with the basic values that require respect in the society we live in. And it is disingenuous to proclaim judicially that the freedom exists in the same courts that then treat it with disregard. I think it would not be imaginable that Parliament would legislate to allow this differential treatment of prisoners, even were it to accord with public moral disapprobation of them, and anyway such legislation certainly would fall short of requirements of human rights law. In cases regarding prisoners, whilst we may not condone their histories or motivations, we are also acting wrongly if we manipulate institutional systems in order to get the ‘right’ outcome. Medical treatment and non-treatment are not legitimate punishments. Judges would do well to remember the warning of Lord Browne-Wilkinson in Bland:

Where a case raises wholly new moral and social issues, in my judgment it is not for the judges to seek to develop new, all embracing principles of law in a way which reflects the individual judges’ moral stance when society as a whole is substantially divided on the relevant moral issues [2, p. 880].

However repugnant a patient’s behaviour, there can be no legitimate support of misapplication of the law just because it would otherwise offend a proportion of the public. If anything may be done to ameliorate the situation, it is for Parliament to decide. Whilst we cannot expect judges to divorce points of ethics from their rulings on law, we can expect them not to alter fundamental laws on the basis of moral abhorrence.

It may be responded to another aspect of this argument that in cases such as that of Ms T [65], the courts palpably were doing the right thing; that her beliefs were rightly held, of themselves, to betray an unsound mind. I would largely accept that view, but I criticise the decision in order to urge that we must be extremely careful when making legal judgments on matters of metaphysics. As things stand, we can disregard the belief that blood is evil, but can only treat with suspicion the belief that receiving the blood of another prevents our chance of salvation. Even if we escape the tyranny of the majority, we do not always prevent the servitude of the tiny minority. There are multiple ways of understanding good and bad, and benefit and harm. And as I have said, the application of principle is made difficult because there are people whom it would be wrong to allow to make the decision. Nevertheless, on its face, the law allows freedom of thought to everyone. The reality can be seen to differ in a way that is profoundly illiberal. This is shown by the distinction between the courts’ treatment of Ms T in 2004 and Mr W in 2002.

If the courts are aware of what they do, it is hard to imagine how it is justified. With regard to just deserts, it is no better a judge taking advantage of an incidental effect afforded by the power to make a decision than it is for a mob to take advantage of its ability to overpower an enforcer of the law. As for sympathetically made orders to treat, these are potentially justifiable, but they mark the potential for an overstep of a judge’s role. In English law, it is the right of each of us to define the good, and to live (and die) by that definition as long as we do not harm anyone else. Put simply, if we want a law that allows us to make decisions for any reason—rational, irrational, unknown, or non-existent—then that is what must be upheld. If we do not, we should modify the law; not hide behind ambiguities in the way that autonomy can be understood. And we certainly should not take advantage of ambiguity in order to cause harm, or endow unwillingly received benefit.

Conclusions

Each of the conceptions of autonomy discussed here—ideal desire, best desire, and current desire—is based on a tenable theory and has a function in our society. Best desire autonomy is generally the conception to be preferred, at least with regard to serious decision-making, but reality requires that the other two at times be employed. If we merely consider ‘autonomy’ in our analysis of law, we may miss crucial inconsistencies in the approach of the courts. It is important that we do not, for we may otherwise turn a blind eye to unjustifiable differentiation between irrelevantly distinct classes of citizen.

The varied nature of the judgments we have considered has the potential to create problems in medical practice. Medical practitioners, too, will be able to rely on the equivocation in some cases, which may lead to bad decision-making. Furthermore, the mixed messages from the courts have the potential just to create confusion, leading to misunderstandings of what permissible practice involves or requires.