Anaesthetists are involved in the care of two thirds of inpatients, underpinning £10 billion of National Health Service income at a pay cost equivalent to 3% of this sum.¹ Although primarily responsible for the administration of anaesthesia during surgery, anaesthetists are the lead clinicians in intensive care and pain services; they are also involved in resuscitation and trauma teams, and non-theatre provision of anaesthesia—for example, during electroconvulsive therapy.

Contrary to public perception,² general anaesthesia is very safe, with an attributable mortality of less than 1:100 000 in the UK.³ High quality delivery of anaesthesia has also been shown to reduce perioperative morbidity.⁴ Complex surgery is increasingly performed on sicker patients, resulting in more technically challenging anaesthesia. Anaesthetists often perform interventions that are separate from treatment administered by other specialties—for example, the provision of epidural pain relief during labour.

In addition to the above trends, the growing threat of consumerist litigation (which places great emphasis on individual autonomy) has led anaesthetists to reconsider their position concerning separate consent for anaesthesia. Traditionally, consent for surgery was obtained by junior surgeons. It is now recognised, however, that junior surgeons rarely perform the surgery for which they obtain consent, and that they do not know enough about the nature and risks of operations to be able to convey this information to patients when eliciting their consent.⁵ The nature and risks of anaesthesia are usually discussed as part of the general information given about surgery. Anaesthesia is, however, associated with its own particular set of risks and consequences, which are quite separate from those associated with surgery. Similarly, the nature and purpose of anaesthesia are different from those of surgery, facilitating rather than delivering definitive medical treatment. It is therefore nonsensical for doctors other than anaesthetists to advise patients about anaesthesia when they will not be administering the anaesthetic, and have little concept of what is involved in the process. This is particularly important in complex cases involving sick patients with illness limited autonomy, for whom consent affords the greatest protection.

In July 1999, the Association of Anaesthetists of Great Britain and Ireland published guidelines entitled Information and Consent for Anaesthesia,⁶ which reflect current legal opinion, but provide no discussion of why anaesthetists should obtain separate consent (if at all). The purpose of this paper, therefore, is to demonstrate not only that anaesthetists should always seek consent from patients (separate from that normally obtained), but also that, far from merely providing a legal shield, consent should be viewed as a valuable tool that inspires discussion about the proposed anaesthetic intervention.

Respect for autonomy—the basis of consent

Consent is an ethicolegal concept which reflects the respect given by a society towards the autonomy of its citizens. In a medical setting, consent allows an autonomous patient—that is, one who has the capacity to think, decide, and act on the basis of such thought, independently and without hindrance⁷—to define and protect his or her own interests and to control body privacy. In law, consent is a device which protects autonomy from third party interference. An anaesthetist—for example, may be liable in battery if he administers a general anaesthetic to a patient without their consent. Legal sanctions are employed to ensure respect for autonomy in society. Recourse to the law does not, however, address the thornier moral problem of why anaesthetists should respect patients’ autonomy, and why they should respect it in preference to other principles that influence the relationship (such as paternalism).

Ethically, the deontologist envisages a “duty of respect for autonomy” in this instance, whereas the utilitarian suggests that respect for autonomy maximises general happiness (although a utilitarian might also argue for the rejection of an individual’s autonomy in circumstances where its recognition could increase unhappiness). Both approaches recognise that there appears to be some intrinsic value in self determination, without which individuals are vulnerable to treatment for treatment’s sake and are denied the autonomy on which much of human happiness is founded.

In the majority of cases, the patient and anaesthetist will agree about the proposed treatment. Respect for self determination becomes problematic, however, when conflict arises, particularly when patients reject advice that is medically in their best interests—for example, choosing local anaesthesia instead of general anaesthesia. Proponents of patient autonomy provide a strong argument in favour of self determination in this instance⁸: in rejecting medical advice, patients may determine that the benefits and risks of treatment do not accord with their sense of self. This is a value judgment that only the patient can make, because it is the sum decision reached through the self integration of all the abstract components of an individual’s personality, components which could never be discerningly evaluated by the anaesthetist in determining best interests.

The limits of autonomy—partial autonomy

At first sight, consent appears to enshrine a core value of contemporary medical practice, empowering patients to remain in control of their fate and bodily integrity, free from unwarranted interference from others. Indeed, the retention of autonomous medical decision making capacity is associated with both improved patient satisfaction and more favourable medical outcomes.⁹

However, not all patients are fully autonomous. Children—for example, are autonomous in that they are capable of independent thought and deed, but the degree of autonomy they possess is not that of a competent adult. A continuum may be envisaged along which mental, physical, and moral development matures towards full autonomy, but this results in applying either a status approach to autonomy—that is, autonomy above a certain age, or a functional approach—that is, autonomy in making some decisions but not others. Similarly, mental illness may transiently or permanently limit personal autonomy, but again, such limitations may be circumstantial—a suicidal patient may still retain the capacity to consent to appendicectomy.

More contentiously, it could be argued that illness itself limits true autonomy, and to varying degrees.¹⁰ Again, one might envisage a continuum of illness severity: one end represents minor conditions which are associated with minimal effects on personal autonomy—for example, removal of a sebaceous cyst under local anaesthesia; the other end represents severe medical conditions that necessitate a dependency on medical treatment—for example, surgical repair of a ruptured aortic aneurysm.

Three scenarios are of special interest to anaesthetic practice. Anaesthetists provide obstetric pain relief in about a third of pregnancies: labour can undoubtedly be very painful, to the extent that it erodes personal autonomy, which might invalidate any maternal consent given. Similarly, patients who experience chronic pain are often physically and psychologically dependent on anaesthetic intervention. Patients may also experience a diminution of their autonomy after the administration of certain drugs.¹¹ notably sedatives, for example, benzodiazepines and opioid analgesics, for example, morphine, drugs which are frequently used as preoperative medication.

A common problem arises in all the above situations. If patients are never more than partially autonomous (through illness, treatment or dependency on treatment), then there must be a threshold level of capacity above which the autonomy of a patient should be respected, and below which a patient is considered insufficiently autonomous to decide for themselves about treatment. The question is—who decides what this level should be? It cannot be the patient; capacity, then, must be determined by a third party (such as the anaesthetist), which involves an inevitably paternalistic process. Libertarian critics of this conclusion argue that patient autonomy should always prevail—even sick patients should be allowed to assert their autonomy through consent. This cannot be right: the septic trauma victim who refuses anaesthesia for the surgical stabilisation of a fractured pelvis would be extremely unlikely to refuse anaesthesia were he sufficiently “mature in his faculties”.

A final conceptual problem involves the quantity and quality of information that is required by a patient in order to form an autonomous opinion. Can a person ever be fully in receipt of all the facts that might influence their decision? If not, is their autonomy compromised? To the former question I would answer surely not; to the latter—yes, their autonomy is compromised. Even the most rigorous research by a patient will not reveal the quantity or quality of medical information that is possessed by the anaesthetist.¹² Moreover, the anaesthetist has had time to assimilate the information, rejecting that which is false or irrelevant, and refining that which appears to be true, a process of reflection that is a function of experience, and which incorporates deliberation of all the subtle nuances of medical fact. The only way of preserving patient autonomy would be for the anaesthetist to act as a dispassionate conduit for facts, leaving the patient to assimilate the knowledge for themselves. This, however, reduces the doctor patient relationship to one that is based on data transfer, which is clearly not why the majority of patients wish to meet their anaesthetist preoperatively: they are seeking both information and opinion. Therefore, their decisions are always to some extent non-autonomous, because they involve the opinions of their anaesthetist, opinions which will be biased in favour of the patient’s best medical interests.

Paternalism

“Partial” autonomy can have one of two possible consequences for medical decision making: either the patient is deemed to be partially self determining in every decision (what might be described as a status approach—for example, whether to have general anaesthesia or regional anaesthesia, or both, during surgery), or it is deemed there are some decisions that the patient is incapable of making (what might be described as a threshold approach—for example, which drug they would prefer for induction of anaesthesia). In either instance, another opinion is required in order to reach the final decision. Usually, this opinion is supplied by the anaesthetist, in the best interests of his patient. Such beneficence is a core philosophical tenet of medicine, and is the rationale behind medical paternalism.

Paternalism has popularly come to represent a deliberate attempt by a knowledgeable elite to limit personal rights and curb consumer choice, and is seen as the direct moral antithesis of autonomy.¹³ Indeed, paternalism is open to legitimate criticism. The argument most commonly advanced states that doctors can never know enough about their patients’ best interests to make decisions for them.¹⁴ Best interests are more than just medical best interests, and involve other values that are of importance to the patient, values which a doctor cannot and will not be able to ever appreciate. This, however, is a theoretical barrier that exists whatever the philosophical position.¹⁵

Another argument suggests that paternalism is open to professional abuse by inviting decisions based on medical self interest rather than patient interest. Consider the example of the anaesthetist who prescribes regular injections of morphine to consenting NHS patients after laparotomy because it is quick and effective, but persuades private patients to consent to epidural analgesia, because this attracts a larger fee for service. Critics might cite this practice as an abuse of power and an indictment of paternalism—the anaesthetist has put self interest before patient interest in both instances. Such practices, are, however, more likely to represent a flaw of character rather than a flaw of paternalism.

The most divisive argument concerns whether beneficence is of greater moral importance than respect for autonomy. Both utilitarianism and deontology can accommodate paternalism. Paternalism has utilitarian moral worth if it maximises welfare, and can only be subordinated by autonomy if the consequences of paternalism produce less happiness than autonomous action (or if autonomy is viewed as an independent determinant of happiness). Paternalism sits less well in deontology, because it tends to advocate the use of people as means to an end. The practice of medicine is, however, fundamentally deontological in nature: “do no harm”, “act in the patients best interests” etc. Respect for autonomy is merely one of a number of duties that the anaesthetist should strive to follow (“pluralist” deontology). Whether the anaesthetist respects patient autonomy ahead of paternalistic action, or vice versa, depends on which of the competing duties is the most compelling. Current medical thinking makes respect for autonomy more imperative than beneficence, the converse only being permissible if a patient clearly lacks autonomy, or explicitly entrusts her best interests to the doctor. In the first instance, consent may be impossible to obtain—for example, an unconscious patient. In the second, the patient may tell the doctor to do “what you think is best”, or implicitly entrust their welfare to the doctor by—for example, signing a consent form without reading it.¹⁶

Nevertheless, there are a number of instances which justify paternalism in anaesthetic practice.¹⁷ Firstly, the welfare of the patient may be served best by paternalistic intervention. In cases involving severe injury or unconsciousness, the anaesthetist might reasonably expect that the patient would consent to intervention were he conscious or fully competent (“predictive” consent). Not only are there potentially life saving benefits for the patient, but there is no loss of autonomy, because the patient is unable to consent, and there is no reason to suppose the patient would object. Intervention could be justified even if the patient verbally refused treatment—for example, suturing lacerations in inebriated patients, or in order to return the patient to normal health—for example, intubating a patient with facial injuries resulting from a car accident, or to prevent further deterioration in their condition—for example, intubating a severe asthmatic who is conscious but unable to speak.

Secondly, patients may waive their autonomy.¹⁸ If the patient refuses to entertain any information about the intervention proposed, any consent given would be invalid. Nevertheless, waivers should be treated with scepticism, particularly when apparently subject to denial or fear (common emotions amongst patients, preoperatively). If the patient, however, waives their autonomy because they feel unfamiliar with the decision making process, are badly informed, or think their doctor knows best, there is an onus on the doctor to seek their consent. The information given to such patients may be more paternalistic in nature, because what the patients seemingly require in this instance is opinion rather than sufficient information on which to form their own, independent decisions.

Thirdly, degrees of paternalism may be justified according to individual patient competence. The benefit may be absolute (if the patient has never been competent or is unlikely to return to a state of adequate competence) or relative, enabling either a return to health—for example, intensive care or continued health until a level of competence is reached—for example, paediatric anaesthesia. A sliding scale of risk dependent, functional competence that justifies compensatory paternalism has been proposed by some authors,^19,20 but rejected by others²¹ who argue for a status approach (the patient is either competent or not: greater risk requires more intense assessment of competence).

Finally, the patient may just make “bad” decisions. Pro-autonomists might argue that a patient can never make a bad decision, merely one that contrasts drastically with the anaesthetist’s opinions. This is plainly wrong—for example, a competent patient may decide after receiving relevant information that he does not wish to receive any analgesia after a major bowel operation, because all methods of analgesia proposed have side effects; the decision is autonomous, but undoubtedly a poor one, and justifies coercive attempts by the anaesthetist to get the patient to change his mind.

Savulescu has noted that “patients can fail to make correct judgments of what is best, just as doctors can”, by failing to make choices that best satisfy their own values—for example, refusing postoperative analgesia, by making choices that frustrate rather than facilitate their autonomy— for example, accepting the need for general anaesthesia during major surgery, but refusing intravenous cannulation, or by making incorrect value judgments—that is, failing to attach the appropriate significance to the relevant facts when making a decision.²² In these situations, further discussion with the patient may reveal the reasons behind their “poor” judgment, and the provision of additional information may resolve any contention between both parties.

Nevertheless, there will be instances when the anaesthetist may continue to believe that a degree of coercion is warranted. Consider the case of a patient, Mr A, who is due to undergo surgical repair of a 10 cm abdominal aortic aneurysm, but who refuses intraoperative blood transfusion, because he is worried about the infinitesimally small risk of contracting variant Jakob-Creutzfeld disease (vCJD) through transfusion. The anaesthetist may explore these fears in a preoperative visit, and may discuss alternative methods of fluid replacement or conservation during this potentially very bloody operation. If the patient still refuses blood transfusion, however, the anaesthetist is faced with a conundrum—it would be morally and professionally very difficult to justify proceeding without potential recourse to transfusion, because of the markedly greater risk of severe patient morbidity or mortality. This example differs from the problems posed by blood refusing Jehovah’s Witnesses, in that such patients refuse blood on the basis of a strongly held religious belief, a belief that to them, forms a core value, and is therefore to be respected. Mr A, however, although making an autonomous decision about an admittedly possible but realistically negligible risk is undoubtedly making a poor decision, and one that may be viewed as being at odds with his normal beliefs and values. The easiest course is to respect Mr A’s decision, and proceed. Alternatively, the anaesthetist may respect Mr A’s decision, but refuse to anaesthetise him because of the substantially increased perioperative risk to Mr A. Hard cases make hard decisions, however, and in this case, the anaesthetist would be justified in coercing Mr A into accepting blood; untreated, a 10 cm aneurysm would be likely to rupture within a year, with a 90% mortality if this occurred outside hospital, an occurrence that the anaesthetist may decide does not conform with Mr A’s values (Mr A being an otherwise happy family man). The anaesthetist may feel that Mr A has attached undue weight to the risk of vCJD, and may continue to try and convince Mr A to accept blood.

Are there limits to the extent to which the anaesthetist should attempt to coerce patients into accepting their advice? Certainly—although it is a fine (and often indistinct) line between coercing a patient to accept treatment (which maintains or enhances their autonomy), and compelling them to accept treatment (which erodes their autonomy). Savulescu’s concept of “rational, non-interventional paternalism” acknowledges this, maintaining that the anaesthetist should make a value judgment of what is best for the patient. This fulfils his duty as a moral agent, and may benefit patient autonomy through rational presentation of treatment options, but rejects the use of compulsion in the therapeutic relationship—that is, it is “non-interventional”.

A loss of trust

Both autonomy and paternalism have practical and philosophical limitations. Why, then, has society sought to promote the principle of respect for autonomy at the expense of medical paternalism, even though the process of consent often involves little more than obtaining a patient’s signature?²³

Traditionally, the doctor patient relationship consisted of a paternalistic association based on trust. Patients trusted their doctors’ advice because they had no other sources of information, respected the knowledge that doctors possessed, and believed that their doctors always acted in their best interests. This situation has changed, however: patients are much better informed about their condition; the advance of technology has outstripped many doctors’ ability to keep pace with change; “quality assurance” initiatives provide league tables of doctors, which are available for public scrutiny, and medical litigation promotes unconditional respect for patient autonomy whilst exposing medical “errors” (which are sensationalised by the media).

Although the changes have for the most part been welcomed by the medical profession, they have undoubtedly led to an erosion of trust between patients and doctors, such that consent has been transformed from a mechanism of doctor patient communication into a defensive legal instrument that hinders respect for patient autonomy. Consent has effectively replaced trust as the basis for doctor patient relationships.²⁴ This should not be viewed as the victory of autonomy over paternalism, however: rather that the loss of trust in the doctor patient relationship has been replaced by patients themselves determining where to place their trust. Patients may trust their own judgment based on assessments of information acquired, or may request further information from doctors and choose whether to accept or reject the advice. This choice places a greater responsibility for decision making on the patient, but that responsibility is lessened if there is trust between patient and doctor, because where there is trust the patient readily accepts that the advice given by the doctor is sound and that the information he conveys is trustworthy. Consent in a trusting relationship maximises patient welfare by respecting both patient autonomy and medical beneficence.

A moral compromise

Consent, therefore, is not needed to shield patients from the “injustices” of medical paternalism, provided there is a satisfactory process of shared decision making between the anaesthetist and patient, based on trust. What might the process involve? The re-establishment of trust in the relationship would facilitate an active, reciprocal, and fluid dialogue between both parties that would enable exploration of the overall best interests of the patient. For the most part, patient autonomy would be respected; under certain circumstances (such as severe illness or loss of competence), anaesthetic opinion would prevail—constant, mutual, and explicit reappraisal of patient competence would tailor the degree of input required by the doctor. This approach could permit the waiver of autonomy under certain conditions, but would make it incumbent on the doctor to advise the patient appropriately according to previously revealed values.

This compromise position is popular among ethicists,^22,25,26 as it provides a practical “third way” between the philosophical extremes of autonomy and paternalism. In anaesthetic practice, this pragmatic approach most closely reflects actual practice. The legal imperative for obtaining consent acts as a catalyst for the development of a trusting relationship, one in which the anaesthetist provides honest, comprehensible information, listens to patient concerns, and adjusts the proposed treatment plan accordingly.

The following example illustrates the approach in practice. An opera singer presents for emergency caesarean section. Preoperatively, the patient is in pain, but is rational enough to make decisions. The anaesthetist suggests spinal anaesthesia, and describes the risks and consequences of the procedure, one of which is paralysis below the waist (less than one in ten thousand spinals). This alarms the patient, who states that paralysis would make her life “unbearable”. The anaesthetist accepts this concern, but informs the patient that the alternative is general anaesthesia. The patient is happy with this, but after inquiring about what is involved in general anaesthesia, she is concerned that the anaesthetist intends to intubate her—this happened to a colleague, whose voice never fully recovered. The anaesthetist stresses the risk of aspiration, but the patient adamantly refuses intubation. The anaesthetist presses the issue because of the risk involved, but finally accepts the patient’s decision, and formulates an alternative method of airway maintenance that is acceptable to the patient, although he insists the patient both discuss her decision with her husband first and accept that further medical intervention may be required. In the event, mild aspiration occurs and the patient requires 24 hours’ care on the intensive care unit, from which she recovers completely. Postoperatively, the anaesthetist explains the complication to the patient and she agrees to continue antibiotics for a week in order to prevent a possible chest infection, even though this prevents her from breast feeding her baby.

This example illustrates how fluid reciprocity recognises both respect for the patient’s autonomy—for example, refusal of intubation, despite the risks, even when autonomy is limited—for example, by labour pain, and the justifiable application of limited paternalism—for example, compliance with antibiotic therapy. The relationship works because it is based on trust—the patient recognises that the anaesthetist is concerned for her best interests (he alters his management plan based on her value judgments about spinal anaesthesia and intubation), and the anaesthetist realises that his management plan is accepted except where there are valid, explicit reasons for rejection. Notably, both sides fulfil their moral obligations without the signing of a formal consent form.

An obvious criticism of this approach is that it is time consuming and impractical given the constraints of an overstretched National Health Service. Due to an ongoing underprovision of hospital beds—for example, patients are rarely admitted to hospital until the day of their surgery, even in the case of major surgery; on admission, they are subjected to a bewildering array of people and paperwork, at a time when they are most anxious about their health and imminent surgery. “Active, reciprocal, and fluid discussion” about the proposed treatment, in this author’s experience, is therefore rarely possible: it takes time to explain anaesthesia to patients, and time for them to reflect on this information and ask further questions, before giving valid consent—time which is just not available if surgical lists are to start on time and progress without interruption in order to be completed. Preadmission clinics and leaflet information might go some way to resolving this issue, but such is the crisis in manpower in the NHS, that there are rarely sufficient anaesthetists to cover surgical lists, let alone such clinics. There is a dichotomy, therefore, between what is ideal—that is, “active, reciprocal, and fluid discussion” and what is possible—that is, the anaesthetist briefly stating what she intends to do, followed by “Is that OK?” Consequently, society (and anaesthetists) must decide what is to be considered more important: “real” consent that fulfils morally and legally acceptable criteria, or patient throughput, that to an extent steamrolls patient autonomy in order to maximise service provision. In reality, the current (and foreseeably future) political agenda favours the latter option, but notably at a time when doctors are being pressured by both the courts and government to improve their legal accountability in relation to consent for treatment.

In summary, unconditional respect for patient autonomy through consent has several disadvantages—patients may only ever be partially autonomous, and they may make bad, uninformed or impractical decisions. Paternalism also has a number of disadvantages, but its limited implementation under certain circumstances may provide very real benefits to patients undergoing anaesthesia, either in terms of medical outcome or by enhancing autonomy in situations in which patients find themselves vulnerable. Anaesthetists should always try to obtain consent from their patients, but should view the process of consent as a stimulus for (rather than an outcome of) decision making, and should always aim to provide a pragmatic ethical compromise when seeking consensual agreement to a proposed plan of treatment.