Medical Ethics and Thoroughgoing Autonomy

In a posting on Practical Ethics Charles Foster defends dignity against “the allegation that dignity is hopelessly amorphous; feel-good philosophical window-dressing; the name we give to whatever principle gives us the answer to a bioethical conundrum that we think is right”. He suggests this allegation usually comes from “the thoroughgoing autonomists – people who think that autonomy is the only principle we need”. He further suggests “there aren’t many of them in academic ethics, but there are lots of them in the ranks of the professional guideline drafters”, see http://blog.practicalethics.ox.ac.uk/2011/03/autonomy-amorphous-or-just-impossible/#more-1320 . Foster believes “we need urgently to disown the monolithic fundamentalism of pop-ethics, and embrace a truly liberal pluralism that listens respectfully to the voices of many principles. Proper pluralism isn’t incoherence.” In this posting I want to defend a limited thoroughgoing autonomist’s position.

Before making my defence I must make clear the position I am defending. I do not believe autonomy is the only principle relevant to medical ethics. Even if some form of medical ethics could be based solely on patient autonomy, which I doubt, the ethos of medicine depends on the notion of caring. I accept beneficence is of major importance to medical ethics. Foster characterizes the form of autonomy he is attacking as an icy, unattractive, Millian, absolutist version of autonomy. I do not believe such a form of autonomy should form the basis of most medical ethics. Firstly I will argue a more user friendly concept of autonomy should be applied in medical ethics. I will then argue this concept of autonomy is not subject to the same problems as a more absolutist version. Lastly I will defend the position that respecting the autonomous wishes of a patient should always be given precedence over acting beneficently towards him.

I argued in my posting of 03/09/10 that an autonomous decision is simply a decision with which the agent identifies himself and “cares about”; in this context I am using “cares about” in the same way as Harry Frankfurt, see for instance (1999, Necessity, Volition, and Love, Cambridge University Press). An autonomous decision is one the agent is wholehearted about. Frankfurt argues a wholehearted decision is one with which the agent is satisfied with. He defines satisfaction as an absence of restlessness or any desire for change. If we accept this concept of autonomy then it has consequences for the autonomous decisions patients make. I argued in my posting of 01/07/08 that this form of autonomy is closely connected to the idea of satisficing. It follows patients can make autonomous decisions which are sub-optimal. But don’t some patients such as Jehovah’s Witnesses already make sub-optimal decisions in practice? Moreover a patient may also make an autonomous decision simply to trust his medical team to do what they believe is best for him. In reality I would suggest this is what most patients implicitly do when making a consent decision. This after all is how many of us explicitly make decisions outside a medical context. We simply trust lawyers or financial advisors for instance, without others questioning our autonomy. Are doctors less trustworthy than lawyers or financial advisors?

Foster believes a Millian absolutist version of autonomy causes problems for medical ethics. I agree with Foster. But does a concept of autonomy based on “caring about” cause the same problems? In what follows I will argue it does not. Firstly Foster seems to assume autonomy requires that agents are able to give an unequivocal answer to the question what do you want? He then suggests it is unusual to meet someone who is so well integrated as to be able to do so. I agree many people have difficulty giving an unequivocal answer to the question, what do you want? But how is this difficulty connected to medical ethics? It seems to me Foster must believe that informed consent requires a patient is able to give an unequivocal answer to what he wants. In medical practice a patient is seldom, if ever, asked what treatment he wants. Rather the question usually posed is simply this; “we believe this treatment is in your best interests do you give your consent?” The validity of this consent is of course dependent on the patient being given adequate information about the relevant details of the proposed treatment. If we conceive autonomy as “caring about” linked to satisfaction then autonomous decisions are also not linked to someone being able to give an unequivocal answer to the question what do you want? Autonomy is simply linked to an absence of any desire on the part of a patient to change his decision. Unequivocal answers are only required when a patient doesn’t want some treatment. The above suggests if autonomy is conceived as “caring about” that Foster’s worry about medical practice and respecting autonomy being incompatible because patients cannot always give unequivocal answers when giving informed consent is not justified.

Secondly Foster worries that the giving informed consent in clinical practice is linked to the giving of informed consent in clinical trials. Foster states informed consent requires that a patient in a consultation with a surgeon about his osteoarthritic hip talks in much the same way as a subject in a clinical trial would talk with the trial’s co-ordinator. I once again agree. However I believe if an autonomous decision depends on an agent’s satisfaction with this decision there is no good reason based respect for autonomy why the above linkage should not be broken. For instance I suggested above a patient may make an autonomous decision simply to trust his medical team to do what they believe is in his best interests because doctors are no less trustworthy than lawyers or financial advisors. Different agents may need different amounts of information to make a decision that satisfies them. Moreover an agent might need different amounts of information in order to satisfy him make a decision in different situations. For instance if I am going to have my blood pressure taken all I need to know is I am going to have my blood pressure taken. If I am going to have an operation for my osteoarthritic hip I may need information about the benefits and risks involved. I may only need to understand the risks involved in very broad terms as the pain in my hip means I will discount these to some degree. If however I am consenting to take part in a clinical trial I need to be better informed about any risks involved as I have no factors which will make me discount these risks. In the light of the above I see no reason based on respect for autonomy why the information needed to give consent to treatment should comparable to the information needed to consent to take part in a clinical trial.

Lastly Foster mentions an important paper by Agledahl, Forde and Wifstad (Journal of Medical Ethics 2011; 37) see http://jme.bmj.com/content/37/4/212.full?sid=7a5d7b0f-c5e1-4291-838e-b0d9414fc1d2 . Agledahl, Forde and Wifstad state “patients' right to autonomous choice is upheld as an ideal although the options of both the patients and the doctors are very limited” and then they rightly point out that “in the healthcare setting, choices are often neither explicit nor available.” The implication of the above seems to be that the authors believe a lack of choice means concern for patient autonomy is basically a sham. Let it be accepted all competent patients have some choice. All competent patients can consent or refuse to consent to treatment. I would suggest if autonomy is based on “caring about” linked to satisfaction that in a clinical setting concern for autonomy is not a sham. It is not a sham because a doctor does not have to offer a patient an array of options out of concern for his autonomy. One option is all that is needed. Indeed accepting a concept of autonomy based on “caring about” might mean in a clinical setting too many choices actually erode a patient’s autonomy. Frankfurt argues,
‘For if the restrictions on the choices that a person in a position to make are relaxed too far, he may become, to a greater or lesser degree, disorientated with respect to where his interests and preferences lie.’ (1999, page 109).
In the light of the above I would suggest the fact patients have limited options does not mean it is impossible to respect patient autonomy in practice.

It might be argued the concept of autonomy I have outlined above is an amorphous concept offering little practical guidance. If this is so it might be asked how I am going to defend a thoroughgoing autonomist’s position. Foster argues we need to embrace a truly liberal pluralism that listens respectfully to the voices of many principles. It seems to the soundness of his position depends on what he means by pluralism. I believe if pluralism means some sort of competition between different moral goods the result would be incoherence. I would suggest the only way to avoid this incoherence is to give priority to some moral goods. This prioritization does not imply we must be able to weigh moral goods but that we must be able to rank them. I will now argue this prioritization means giving precedence to respecting autonomy over acting beneficently. If I am going to act beneficently towards someone I must care for him. The basis of my care may be sympathy or empathy. If my beneficence is based on sympathy it seems clear to me that I may act in what I conceive to be his best interests and override his autonomy. If however my beneficence is based on empathy this option is not open to me. If I feel empathy for someone I must focus on what he cares about rather than what I think might be in his best interests. It follows if I want to act beneficently towards someone, and my beneficence is based on empathic concern rather than sympathy and I believe his best interests clash with his autonomy, I should nevertheless give precedence to respecting his autonomy over acting in these interests. Beneficence based on empathy automatically gives precedence to respecting autonomous decisions. It follows if beneficence is based on empathy it is possible to defend a thoroughgoing autonomist’s position. Agledahl, Forde and Wifstad seem to partially support my position because they believe “the right to refuse treatment is fundamental and important”. If autonomy is based on “caring about” and beneficent care based on empathy then a limited thoroughgoing autonomist’s position means doctors need not concern themselves to much with providing choices but must respect all autonomous decisions. However an objector might point out I have provided no reason as to why beneficence should be based on empathy rather than sympathy. We can act beneficently towards animals. Clearly such beneficence is based on sympathy. I would suggest we cannot feel empathy for animals due to epistemic ignorance. I would further suggest we can act more beneficently towards people if our concern is empathic rather than sympathetic. It follows good beneficent care is based on empathic concern. Accepting my suggestions means medical ethics should be prepared to accept a limited thoroughgoing autonomist’s position. Indeed I would argue such a position concurs very well with the practice of medicine with the exception that accepting such a position would mean accepting a fully autonomous decisions by a patient simply to trust his clinicians as to which is the best course of treatment for him.

Brooks, Brudner and Justice

In my previous posting I suggested that any society which insists on compulsory beneficence or the compulsory acceptance of beneficent care is not a truly flourishing society. I further suggested that the only requirement necessary for a flourishing society is that each should be free to do as seems best for herself provided by so doing she does not harm others. This might be classed as the Millian position. In this posting I want to examine what this suggestion means for the basis of law. My examination will consider the position of Alan Brudner, (2009, Punishment and Freedom: A Liberal Theory of Penal Justice, Oxford) and a critique of that position by Thom Brooks (2011, Autonomy, Freedom, and Punishment, Jurisprudence, Forthcoming available at SSRN: http://ssrn.com/abstract=1791538 ).

Brudner believes we should punish others when they damage or threaten to damage our autonomy rather than because of any harm they cause us. Of course if we damage or threaten someone’s autonomy we harm them. In the light of the above I would suggest Brudner’s position might be described as follows. We should punish others only because of the harm they do to our autonomy and not because of any other harm. It is this re-described position I will consider in this posting even though it is by no means certain Brudner would entirely agree with my re-description. I have suggested the only requirement necessary for a truly flourishing society is that each of us should be free to do as she sees best provided by so doing she does not harm others. In the light of this suggestion it might appear to follow that I believe the basis of the law should be harm to others rather than restricted to harm an individual’s autonomy. This appearance is deceptive. Any harm to an autonomous agent always also harms her autonomy. If I harm an autonomous agent I do something to her which she would not do to herself. My actions prevent the agent from making a choice she would identify herself with or which satisfies her. My actions damage her autonomy. It follows all actions that harm autonomous agents also do harm to their autonomy.

Brooks believes Brudner’s position is unconvincing in some contexts. He points out “that there may be many actions we may want criminalized (e.g., traffic offences) that would not be clearly included” in any criminalization as defined by Brudner. Brooks also points out suicide, which ends the victim’s autonomy, and moderate alcohol use, which temporarily affects the user’s autonomy, both of which we would intuitively not want to be criminalized, might well be criminalized provided criminalization is based on harm to an agent’s autonomy. Someone might argue some traffic offences such as speeding have the potential to damage the autonomy of another and hence their criminalization can be justified on Brudner’s account. She might then proceed to point out it is possible to respect autonomy in different ways. Firstly it is possible to respect someone’s autonomous decisions and secondly to protect her capacity to make autonomous decisions. Let us assume a patient suffers from some incurable disease and makes an autonomous decision to commit suicide. If criminalization is justified by damage done by not respecting autonomous decisions rather than harm done to an agent’s capacity to make autonomous decisions then suicide and the moderate use of alcohol should not be criminalized. If criminalization is justified by damage to an agent’s capacity to make autonomous decisions then these actions should be prohibited. How do we choose between respecting autonomous decisions and protecting someone’s capacity to make these decisions? I have argued in previous postings respect for autonomy is intimately tied to respect for persons and it is impossible to respect persons if one doesn’t respect their decisions, see for instance http://woolerscottus.blogspot.com/2008_08_01_archive.html . If my arguments are accepted then it follows that whilst we must respect someone’s autonomous decisions and protect her capacity to make these decisions that when these two ways of respecting autonomy are incompatible we should give precedence to respecting autonomous decisions. It further follows Brudner’s justification of criminalization may be able to account for the cases Brooks raises.

Brooks states that he strongly agrees with Brudner’s focus on offering a political, not a moral, theory of punishment. I’m by no means sure this focus on a purely political basis is possible. It seems to me if we accept Brudner’s position, that the legality of some action depends on whether it harms or threatens to harm our autonomy, we must also implicitly accept that the domain of legal responsibility is the same as the domain of legal concern. Accepting the above means we must accept only all autonomous agents can be held legally responsible and it also means we must accept the domain of legal concern only extends as far as these agents. Accepting the domain of legal concern only extends as far as autonomous agents seems problematic. Clearly children are not fully autonomous agents and equally clearly a child is of legal concern even though she cannot be held to be legally responsible for her actions. In the case of children this problem might be overcome in the same way as traffic offences above. Brudner’s position might be slightly modified so that the legality of some action depends on whether it harms or threatens to harm our potential autonomy. However even if this modification is accepted a problem remains. Let us assume that animals cannot be autonomous. Clearly if a farmer or pet owner lets her animals starve to death the fate of these animals should be of legal concern. Even more clearly severely handicapped children and old people suffering from dementia should be of legal concern even though these people will never be autonomous.

In the light of the above it seems to me that the legality of an action cannot be based solely on political concerns about autonomy. Let it be accepted the only legal limitations a truly flourishing political society may impose on an agent’s autonomous decisions must be to limit the harm her decisions do to others. However what counts as harm is based sometimes depends on moral concerns completely unconnected to concerns about autonomy. Depending on the moral interpretation of what counts as harm the constraint of harm on the unbridled exercise of autonomy may vary greatly.