Paternalism and Partial Autonomy Notes

Paternalism and Partial Autonomy

Abstract

• The paper addresses the contrast between standard adult autonomy and patients' reduced autonomy, suggesting that the latter requires paternalistic treatment.
• It argues that human autonomy is limited for everyone, not just patients.
• An adequate account of paternalism must consider the variable and partial nature of actual human autonomy.
• Autonomous action is self-legislated action, requiring understanding and choice.
• Autonomy is reduced or impossible in infancy, early childhood, unconsciousness, senility, some illnesses, mental disturbances, and severe retardation.
• Illness often damages autonomy, making it a less promising fundamental principle for medical ethics than concern for well-being.
• However, beneficence alone is insufficient as a fundamental aim, as it may disregard patient autonomy.
• Debates about paternalism, autonomy, consent, and respect for persons are articulated differently in various ethical theories.

1. Paternalism and Autonomy in Result-Oriented Ethics

• Consequentialist moral reasoning does not typically view patient autonomy as a fundamental constraint.
• Utilitarianism prioritizes welfare or well-being, requiring respect for autonomy only when it maximizes welfare.
• Paternalism is not inherently morally wrong in this view, but actions disregarding autonomy are wrong if non-paternalistic actions would have maximized welfare.
• Only an 'ideal' consequentialism that values autonomy independently could view subordinating autonomy to beneficence as wrong.
• In utilitarian ethics, autonomy has marginal importance, and paternalism is only misplaced when it reflects miscalculation of benefits.
• John Stuart Mill's On Liberty is discussed as a classical utilitarian perspective. Mill believed that individuals are the best judges of their own happiness and that autonomous pursuit of goals is a major source of happiness.
• Mill argued that happiness is seldom maximized by thwarting others' goals; thus, paternalists are often well-intentioned but miscalculate.
• The author argues that Mill's claims are empirically dubious: many people might be happier under beneficent policies that reduce autonomous action.
• Many patients desire relief from difficult decisions and the burden of autonomy; even when they don't want decisions made for them, they may be unable to make them well.
• Mill's emphasis on autonomy is somewhat anomalous in result-oriented ethical thought.

2. Paternalism and Autonomy in Action-Oriented Ethics

• Autonomy can have a more central place in action-oriented ethics, where the preconditions of agency are fundamental.
• Since autonomy is a presupposition of agency, action-centered ethics must make the autonomy of agents a basic moral concern.
• This concern can be expressed as not using others, respecting them, treating them as persons, securing their consent, and avoiding coercion, including paternalistic coercion.
• A key difficulty is that some humans lack the cognitive and volitional capacities to be considered autonomous.
• The moral ground for insisting on respect or support for human autonomy in such cases is questioned.
• Patients often have reduced capacities for agency, undercutting some, but not all, possibilities for action.
• Agent-centered moral theories are relevant to medical ethics only if based on an accurate view of human autonomy.
• Enlightenment political theory, especially Locke, argues against paternalism and for human autonomy, where consent legitimizes government action.
• The sovereignty of the people is understood as their consent to the laws by which they are ruled, retaining autonomy while obeying laws.
• Debates arise over whether consent must be express or can be tacit. In politics, this concerns explicit allegiance versus tacit acceptance through continued residence.
• In medical ethics, the parallel debate concerns explicit consent via signed forms versus placing oneself in the hands of a doctor as implicit consent.
• Advocates of 'informed consent' invoke a contractual model of human relations, analogous to commercial contracts and limited government action.
• Critics of informed consent in medicine reject the intrusion of commercial and contractual standards into medical care.
• The contractual picture is questionable in medicine because the notion of 'ideal rational patients' cannot withstand scrutiny.
• Opposition to medical paternalism may reflect an abstract and inaccurate view of human consent, irrelevant in medical contexts.

3. The Opacity of Consent: A Reversal of Perspective

• Human autonomy is limited and precarious in many contexts, and consent is standardly selective and incomplete.
• Consent is given to proposed actions or projects under certain descriptions; it does not necessarily extend to logical implications, likely results, or unavoidable corollaries.
• Consenting is opaque: individuals do not necessarily 'see through' to the implications of what they consent to.
• Patients consenting to operations may be unaware of further implications or risks.
• The opacity of patients' consent is not radically different from all human consenting.
• Even in transparent, regulated contractual arrangements, consent has limits, recognized when contracts are voided due to cognitive or volitional disability or unmet expectations.
• Medical cases highlight typical limits of human autonomy and consent, rather than being anomalies.
• The limitations of actual human autonomy aren't usually considered constraints on respect for autonomy but as aberrations from ideal autonomous choosing.
• The rhetoric of the liberal tradition often forgets the finitude of rational beings in favor of abstract perspectives.

4. Actual Consent and 'Ideal' Consent

• Debates on consent theory often shift from concern with dubious actual consent to hypothetical consent by an ideally autonomous agent.
• This shift allows us to treat impairments of autonomy due to illness as irrelevant, asking what the ideally autonomous patient would consent to.
• This combines ostensible concern for human autonomy with paternalistic medical practice.
• Focus shifts from what has been consented to to what would be ideally consented to, replacing concern for others' autonomy with concern for hypothetical, idealized agents.
• To avoid this replacement, we must define when actual consent is genuine and significant versus spurious or misleading.
• Instead of focusing on idealized conceptions of human choosing, we must look at messy actual choosing.
• The point of concern for autonomy and genuine consent is that it is up to those affected to choose whether to accept or reject proposals.
• Respecting others' autonomy requires making consent possible, considering their partial autonomy.
• Medical practice respects patients' autonomy when it allows them, as they actually are, to refuse or accept proposals.
• Some impairments prevent refusal or acceptance, requiring paternalistic treatment.
• For patients with some capacities for autonomous action, anything comprehensible and refusable can be treated as subject to their consent or refusal.
• Doctors must avoid haste, pressure, and intimidation in unfamiliar medical environments.
• Without such care, patient 'consent' lacks the autonomous character that shows they have been treated as persons rather than paternally.

5. 'Informed Consent' and Legitimating Consent

• There is a temptation to formalize consent procedures to guarantee the avoidance of paternalism.
• Given the varied ways human autonomy is limited, no set procedure can guarantee genuine consent.
• Examples of colonialists 'negotiating treaties' with barely literate native peoples demonstrate a lack of respect for autonomy.
• Medical practice relying on routine signing of 'consent forms' may meet litigation conditions but doesn't show concern for actual human autonomy.
• Such procedures are disreputable, given the difficulties even the most autonomous have in assimilating distressing information or making hard decisions.
• Serious respect for autonomy demands that patients' refusal or consent to fundamental treatment aspects be possible.
• Practitioners must ensure that patients understand the basics of their diagnosis and proposed treatment and feel secure enough to refuse treatment or insist on changes.
• If a proposal is accessible and refusable for an actual patient, silence or compliance can reasonably be construed as consent.
• Seeking consent and respecting autonomy are discredited when the 'consent' obtained doesn't genuinely reflect the patient's response.

6. Partial Autonomy, Coercion, and Deception

• Consent to all aspects of proposed treatment is neither possible nor required, only the ideally autonomous could offer such consent.
• Respect for autonomy requires that consent be possible for fundamental aspects of actions and proposals, but not necessarily trivial or ancillary aspects.
• Treatment undertaken without consent, when a patient could have made their own decisions with care and respect, may fail in many ways.
• In the most serious cases, the action undertaken uses patients as tools or instruments, precluding consent even for ideally autonomous patients.
• Where a medical proposal hinges on coercion or deception, not even the most rational and independent can dissent or consent.
• In deception, 'consent' is spurious because cognitive conditions are not met; in coercion, it is spurious because volitional conditions are not met.
• Some non-fundamental aspects of treatment may include elements of deception or coercion.
• Use of placebos or reassuring but inaccurate accounts of expected pain might be permissible deceptions.
• Restraint during a painful procedure might be permissible coercion.
• Using patients as unwitting experimental subjects, concealing fundamental aspects of illness, prognosis, or treatment, or imposing treatment and preventing refusal, uses patients and fails to respect autonomy.
• At best, such imposed treatment might be impermissible paternalism; at worst, it might be assault or torture.

7. Partial Autonomy, Manipulation, and Paternalism

• Use of patients is an extreme failure to respect autonomy, precluding consent even of the ideally autonomous.
• Respect for partial autonomy requires avoiding treatment that, while refusable by the ideally autonomous, would not be refusable by a particular patient in their condition.
• Manipulation and questionable paternalism fail to meet these requirements.
• Patients are manipulated if they are 'made offers they cannot refuse,' given their cognitive and volitional capacities.
• Ensuring 'consent' is not manipulated requires spelling out available alternatives and showing that refusal of treatment is a genuine option.
• 'Consent' achieved by misleading or alarmist descriptions of prognosis or uninformative accounts of treatment does not show genuine respect.
• Only patients unable to understand or decide need complete paternalistic protection.
• When there is unequal power, knowledge, or dependence, avoiding manipulation and unacceptable paternalism demands a lot.
• Manipulators use knowledge of others' weaknesses to impose their goals; paternalists may not recognize others' goals.
• Patients may be helped by advice and information to achieve their aims, but if others' aims determine the limits and goals of intervention, it will be unacceptably paternalistic.
• Imposing others' goals on patients capable of some autonomy does not respect patients.
• Patients' own goals, medical and non-medical, and their plans for achieving these, are constraints on any medical practice respecting their autonomy.
• Where patients' goals differ from doctors' goals, respect requires that these goals not be overridden or replaced.
• Paternalism is simply the imposition of others' goals on patients capable of some autonomy.
• The contextually-sensitive, action-oriented framework insists that judgements of human autonomy must be contextual.

8. Respecting Limited Autonomy

• Medical paternalism has been considered within three frameworks:
  • Result-oriented: Concern for autonomy is subordinated to total welfare.
  • Action-oriented (abstract): Medical practice rules out all paternalism, permitting only treatment consented to by 'idealized' autonomous agents.
  • Action-oriented (partial autonomy): Boundaries are drawn between permissible and impermissible paternalism in given contexts.
• This account yields no formula for avoiding coercion and deception.
• Trying to incorporate concern for actual, partial capacities for autonomous action into an account of respect for patients and medical paternalism leaves us without a single boundary line.
• There are patterns of reasoning yielding different answers for different patients and proposals.
• Attempts to provide uniform guidelines are insensitive to the radical differences in capacity of different patients.
• A theory of respect for patients must rely heavily on actual medical judgements to assess patients' current capacities to absorb and act on information.
• 'Professional judgement' or 'current medical standards' alone cannot provide appropriate criteria for treating patients as persons if they don't consider the varying ways patients can exercise autonomy.
• Professional judgement determines respect for patients only when guided by concern to communicate effectively what patients can understand and to respect the decisions they can make.

9. Issues and Contexts

• Consideration of determinate cognitive and volitional capacities of particular patients at particular times provides a framework for working out boundaries of permissible medical paternalism.
• If such judgements are contextual, there is no way to demarcate unacceptable paternalism in the abstract.

A. Temporarily Impaired Capacity for Autonomy

• If respect for autonomy is morally fundamental, restoring capacities is morally fundamental, survival is necessary, but not sufficient to restore autonomy.
• If patients' autonomy constrains practice, survival can never be foregone in favor of autonomy, but it is an open question whether survival with no or greatly reduced capacities for autonomy can be a permissible goal.
• Risky surgery may reasonably be imposed for restoring capacities even when mere survival would be surer without surgery.
• Temporary loss of autonomy offers grounds for paternalistic intervention to restore autonomy - but not for all paternalistic interventions.

B. Long Term or Permanent Impairment of Autonomy

• This is the standard situation of children, and so the original context of paternalism.
• Those with long and debilitating illnesses may suffer varied impairments of autonomy.
• Consideration of parental paternalism may illuminate these cases.
• While the law fixes an age to end minority, parents adapt to constantly altering capacities for autonomous action.
• Choices which cannot be made at one stage can at another; autonomy develops unevenly.
• Medical trajectories may not be towards fuller autonomy.
• Medical and other decisions may then have to be imposed.
• There is no general reason to think that those who are unable to make some decisions are unable to make any decisions.
• Even when full return of capacities is unlikely, patients, like children, may gain in autonomy when an optimistic view is taken.

C. Permanent Loss of Autonomy

• Decisions must be made that go beyond restoring autonomy.
• Medical staff and relatives may use a notion of hypothetical consent.
• They are likely asking, 'What would this patient have chosen in this situation?' rather than 'What would the ideally autonomous choose?'
• If this can be answered, it may be possible to maintain elements of respect for the particular patient as he or she was in former times.
• Usually, this provides only vague indications for medical or other treatment, and respect for absent autonomy can be at best vestigial.

D. Lifelong Incapacity for Autonomy

• For those who never had or will have capacities for autonomous action, the notion of respect is vacuous.
• There is no answer to the hypothetical question 'What would he or she have chosen if able to do so?' and the question 'What would the ideally autonomous choose?' may have no determinate answer.
• Here, paternalism must govern medical practice indefinitely.
• The main questions concern dividing authority for paternalistic decisions between relatives, medical staff, and legal guardians.

References and notes

• (1) Mill J S. On liberty. In: Warnock M, ed. Utilitarianism and on liberty, etc. London: Fontana, 1972.
• (2) This has been a recurrent criticism of Mill from Stephen J F, Liberty, equality, fraternity, London: Smith, Elder, 1873, to Dworkin G. Paternalism. The monist 1972; 56: 64-84 and reprinted in Sartorius R, ed. Paternalism. Minneapolis: University of Minnesota Press, 1983: section IV.
• (3) Broader worries mushroom here too: what grounds the moral status of non-autonomous humans in action- oriented ethics? For recent discussion see Haksar V. Liberty, equality, perfectionism. Oxford: Clarendon Press, 1979; Clark S. The moral status of animals. Oxford: Clarendon Press, 1977; Dennett D Conditions of personhood. In: Rorty A, ed. The identity of persons. Berkeley and Los Angeles: University of California Press, 1976, and reprinted in Dennett D. Brainstorms. Hassocks, Sussex: The Harvester Press, Ltd 1979.
• (4) Here US and British practice differ. US legislation and debates often stress the need to secure informed consent from patients (or their guardians). Cf. discussions and bibliography in Veatch R M. Case studies in medical ethics. Cambridge Mass: Harvard University Press, 1977. British law holds that 'what information should be disclosed, and how and when, is very much a matter of professional judgement', and that 'there is no ground in English law for extending the limited doctrine of informed consent outside the field of property rights'. See Sidaway v Board of Governors of the Bethlem Royal Hospital and the Maudsley Hospital and Others, Law Report, The Times, 1984 Feb 24. However, medical paternalism may be more practised in the US than it is praised by those who write on medical ethics. See Buchanan A E Medical paternalism. Philosophy andpublic affairs 1978; 7: 371-390, and reprinted in Sartorius, see reference (2).
• (5) For further comments on the limitations of 'normal' abilities see Wikler D. Paternalism and the mildly retarded. Reprinted in Sartorius, see reference (2).
• (6) A point made long since by Isaiah Berlin in Two concepts of liberty. Four essays on liberty. Oxford: OUP, 1969.
• (7) For the interpretation of Kantian ethics offered here see also O'Neill 0. Kant after virtue. Inquiry 1984; 26: 387- 405; Consistency in action. In: Potter N, Timmons M, eds. New essays in ethical universalizability. Dordrecht, the Netherlands: Reidel publishing company, forthcoming, and Between consenting adults, unpublished.
• (8) Bok S. Lying: moral choice in public and private life. New York: Harvester Press, Random House, 1978: 234. Bok points out that sometimes the use of placebos may be more than ancillary, (61-68), and also discusses fundamental forms of deception such as hiding from the patient that the illness is terminal. On the latter point see also Kubler-Ross E. On death and dying. New York: Macmillan 1969, and the bibliography in Veatch, reference (4).
• (9) For discussions of some distinctive features of children's partial autonomy see Leites E. Locke's liberal theory of fatherhood; Slote M A. Obedience and illusions and Katz S N, Schroeder W A, Sidman L. Emancipating our children - coming of legal age in America. In: O'Neill 0, Ruddick W, eds. Having children: philosophical and legal reflections on parenthood. New York: OUP, 1979.