Advance treatment planning for mental illness by way of written advance directives is one of the issues of debate in contemporary psychiatry in a number of western civilized societies (Howell, Diamond & Wikler, 1982; Dresser, 1982; Brock, 1993; Backlar, 1995; Savulescu & Dickenson, 1998) (1). As is well known, there are certain mental illnesses that have periodic features. The most prominent are the manic-depressive psychoses, but we can also think of young chronic schizophrenics (Brock, 1993). People who suffer from manic-depression often can benefit from treatment and live a life that can be in many respects satisfying. However, it is a well known fact that sometimes these patients stop their medication, and that by doing so a manic condition can be triggered (Goldman & Post, 1991). In such cases, some of these patients refuse treatment or supervision because they lack insight into their actual condition. They may undertake actions that are harmful to themselves or to people in their environment. If not in the short run, in the long run this behaviour may be extremely destructive for individual self-esteem, private and social relations, employment, etc... However, in a number of cases the patient does not meet the legal criteria for involuntary commitment or coercive treatment. His or her actions may be harmful, but not (extremely) dangerous to himself or others (Berghmans, 1997b). Through the use of psychiatric advance directives, it would be possible for mentally ill persons who are competent and with their disease in remission, to give prior consent to treatment at a later time when they are incompetent, and are refusing treatment (Brock, 1993).
In this paper we discuss a number of ethical issues and concerns that are raised by psychiatric advance directives. We will show that the dominant, analytical-principlistic approach falls short in settling these issues. Therefore, we will present an alternative narrative ethical approach. Psychiatric advance directives can have different forms. In this paper we focus on advance directives in which a wish to receive psychiatric treatment is formulated (as opposed to the different issues that are raised by advance directives involving refusal of commitment and/or treatment).
Before considering a number of ethical issues, we first discuss the general model of a psychiatric advance directive or so-called 'Ulysses contract' or 'self-binding contract'. We will argue that this contract need not be interpreted as an act of self-determination, as it is usually done from a principlist perspective. From a narrative perspective, it can be regarded as an element in the coconstruction of the patient's life-story.

Ulysses contracts in psychiatry: self-determination or narrative work?

Different authors have proposed a so-called 'Ulysses contract' as an instrument of consent-in-advance ('prior consent') in psychiatric care (Howell, Diamond & Wikler, 1982; Culver & Gert, 1982; Lavin, 1986; Rosenson & Kasten, 1991; Brock, 1993). They typically argue that by issuing a Ulysses contract, people with recurring psychosis can protect themselves from the harmful results of a psychotic episode. Ulysses contracts are considered a form of justified self-paternalism, just like Ulysses acted paternalistically towards himself by instructing his crew to bind him to the mast of his ship before they sailed past the irresistible Sirens, and to ignore his requests for release (Lavin, 1986). Thus he was able to enjoy the beautiful singing of the Sirens without suffering the disastrous results that would normally have appeared (Elster, 1979).
The particular harms to be prevented by treatment in accordance with such a contract would be those identified by the individual and his or her doctor (Dresser, 1982). Above that, the individual and his/her psychiatrist would agree to the types of behavior of the patient that signal the outbreak of a psychotic episode (i.e unusually excessive spending of money, sexual promiscuity, bizarre thinking). Typically, in the contract the patient and psychiatrist prospectively can agree with regard to the interventions that are considered necessary and justified in case these signs and symptoms express themselves, even if the patient then refuses these interventions. For example, a patient could consent in advance to clinical commitment, to electroconvulsive therapy and/or particular psychotropic medication.
In sum, the devastating results of recurrent psychotic behavior - which often cannot be avoided under current civil commitment laws in a number of jurisdictions - could be minimized by means of a self-binding contract (Dresser, 1982). This interpretation of Ulysses contracts is centered around the notion of self-determination. The contract is a way to rule over one's own future, and autonomously direct one's own life. It is essentially a token of power. Ulysses commands his crew to do what he has cleverly thought out in advance. The crew merely obeys his orders. From this perspective, the self binding contract is the act of an individual determining his or her own future, in difficult circumstances. Yet, we think the use of advance directives in psychiatry can be regarded quite differently. The psychiatric patient is not an autonomous individual directing his or her own life. He or she is a person in distress, and in need of care. The distress is not only apparent when the patient is in crisis. The danger of future psychosis is always lingering, and induces the patient to seek help. In discussing future treatment options, including a possible Ulysses contract, the patient tries to get hold on his or her life. This is not an act of self-governance, but a practice of caring (Tronto, 1993). In discussing the future, the patient and the psychiatrist try to find ways to maintain and repair a world which is in need of joint caring activities.
From this alternative perspective, the contract is part of the life-story of the patient, a story which is constructed in interaction with others (Ricoeur, 1986, 1991; Widdershoven, 1993; Widdershoven & Smits, 1996). It is not an act of power, but an attempt to relate future events to past experiences, and to create a practice of dealing with psychosis by integrating them into one's life-project, a project which is not a given plan, but a joint endeavour which is in need of care and support. The Ulysses contract need not be regarded as an order to the physician, it is rather the expression of a mutual engagement to support and trust one another in precarious conditions. This interpretation of the Ulysses contract is much more in line with the ancient Greek worldview than the common interpretation in terms of self-governance. The Greeks emphasized the vulnerability of life, and the awareness that human life is dependent on things that human beings do not control (Nussbaum, 1986). This is also apparent in Homer's Odyssea: although Odysseus is certainly clever, he is also dependent upon fate, and upon the protection by the Gods. It is the nymph Circe who gives him the advice to let himself be tied to the mast of the ship, and he trusts her in this (not blindly, but always aware of the possibility that things might turn out less managable than hoped for).

The use of Ulysses contracts and other psychiatric advance directives, however, raises a number of ethical, legal and practical questions. We now discuss some of the ethical issues that have been raised in the debate. We will compare the principlist and the narrative approach to each of these issues.

Right to refuse treatment

In many jurisdictions, the (psychiatric) patient has a legal right to refuse treatment. Ulysses contracts undermine this right to refuse treatment, because the prior consent (if given legal authority) leads to the possible overruling of a later refusal of commitment/treatment by the patient (Radden, 1988; Rogers & Centifanti, 1991). The issue, then, is what moral justification can be given for overruling later refusals of treatment on the basis of the formerly expressed wishes of the patient.
Most commentators who are in favour of Ulysses contracts argue that an overruling of a refusal of treatment can be justified if the patient has become incompetent. The prior wishes then become authorative. This principlist argument is based upon the assumption that competent decision-making by the patient is crucial for the moral status of treatment. The dichotomy between competency and incompetency which is implied in this approach is however, problematic. In the first place, one may doubt whether the patient in non psychotic conditions is competent to decide about later treatment. Does he or she really have enough information and distanciation to make a competent decision? Again the story of Ulysses may serve as an antidote against too much emphasis on individual rational decision making. Ulysses did not know what to expect; he was told about the sirens by Circe, and had to rely upon her wisdom. One might object that psychiatric patients for whom the Ulysses contract is meant, know their psychosis, since they have prior experience. Yet, this knowledge is not secure, as a new psychosis will not always be exactly like the former one(s). Critics of advance directives emphasize the vagueness and arbitrariness of concepts and criteria in psychiatry (Dresser, 1982; Radden, 1988; Berghmans, 1994). It is difficult for psychiatrists as well as for patients to foresee future conditions of illness. In the second place, it may be doubted whether the patient during psychosis is totally incompetent. Again, not every psychosis is the same. One should at least be open to the possibility that the patient might refuse treatment for reasons which are different from those considered to be irrelevant prior to the psychosis.
Instead of making patient competency the moral basis for advance directives, we think that advance directives can better be regarded as based upon a narrative work of patient and physician, making sense of joint experiences with critical events. There is no need to have certain knowledge about the future to make arrangements, rather such arrangements can be justified by referring to the practical necessity of developing a scenario to deal with adverse situations, together with the willingness to adapt the scenario on the basis of future experiences. This implies that the physician will always have to adress the issue how to interpret a refusal of treatment. It cannot simply be regarded as a prove of the incompetency of the patient, but has to be related to the patient's history. Criteria of competency and incompetency have to be part of the life narrative of the patient, and open to discussion. Thus, each specific way of dealing with refusal during an episode of psychosis will have to be evaluated later on, and will have to result in a further refinement of the scenario. From a narrative perspective, a moral justification for overruling refusal cannot be found in a specific act of self-determination in the past, but only in the process of narrative work which is the broader context of treatment, and which reaches from past to future agreements.

The concept of autonomy

Central to the moral and legal status of Ulysses contracts is the question how we view the autonomy of the patient concerned. Questioning the moral authority of the wishes of the individual that issues the Ulysses contract and the wishes of his/her 'later self' that refuses psychiatric interventions, autonomy becomes a relevant consideration in the dominant approach of Ulysses contracts.
Those arguing in favour of the Ulysses contract typically refer to 'split level' approaches of autonomy, as for example defended by Dworkin. In Dworkin's view, autonomy as 'procedural independence' is the second-order capacity of persons to critically reflect upon their first-order preferences, desires, and wishes, and the capacity to accept these or to try to change them in the light of higher-order preferences and values (Dworkin, 1988). In line with Dworkin's approach it can be argued that the critically reflected preferences of a person (as formulated in a Ulysses contract) must receive priority above his or her first-order desires (a refusal of treatment in case of crisis) because the former are an expression of second-order autonomy.
Another way of framing the question of individual autonomy in the context of self-binding contracts is to ask what are the 'authentic' or 'true' wishes and preferences of the person involved (Feinberg, 1986). Is it possible to decide whether Dr. Jekill or Mr. Hyde represents the authentic self of the individual? Or, if we use the model of a 'hierarchy of selves' (Elster, 1979), the question is how other persons can pick out the authorized spokesman for the hierarchy, and how they can decide whether the revocation of an order is issued by the lowest or the highest self. Feinberg (1986) argues that the voluntariness of a decision is the decisive criterion, admitting that the decisions of the 'former' and 'later' self of the person can be equally voluntary. Schelling (1984), presuming that both selves may be equally authentic, claims that here a distributive issue arises. The issue then is how to deal fairly with two selves that have opposite needs.
All these concepts of autonomy focus upon the individual's wishes and preferences. The issue adressed is how to choose between conflicting or contradictory wishes of the same individual (a wish for care versus a refusal of treatment; a wish for being helped to live a normal life versus a wish to experience utter freedom and superiority in a period of manic behaviour). From a narrative perspective, another concept of autonomy can be developed, emphasizing the crucial role of biographical work and the embeddednes in social relations (Gubrium & Buckholdt, 1977; Agich, 1993; Moody, 1992). A narrative approach to autonomy assumes that autonomy is not equal to independence, but that it is developed in relations of dependency. Given conflicting urges, autonomy is not the ability to choose between them, but the ability to integrate them in a meaningful way into one's life history. This means that one has to find a story which deals with both sides, the 'good' as well a the 'bad' side of one's personality. From this perspective, the crux of the Ulysses contract is not that it secures the patient's autonomy because it expresses the priority of the higher wish over the lower one, but that it enables patient and physician to integrate various phases - 'good' as well as 'bad' - into the life history of the patient, not in a theoretical, but in a practical way of keeping life sustainable.
From a narrative perspective, autonomy is not the basis for interventions (such as medication or the use of force), but the goal of such interventions. This implies a different view on the balancing of various aspects of the self. It is not by balancing preferences or states that one can justify certain therapeutic interventions; on the contrary: the interventions are justified in so far as they contribute to a balance between extreme states.

Personal identity

A number of commentators challenge the moral authority of (psychiatric) advance directives from the perspective of personal identity. The central challenge presented by the problem of personal identity is to articulate the conditions under which stages of a person's life are stages of the same person or, conversely, to explain the circumstances which signal the development of a different person (Dresser, 1989). Psychological changes through time, as is typically the case in people suffering from manic-depression (or, differently, in case of dementia), raise questions about psychological continuity and connectedness between different stages in a person's life and the possibility that a different person (or successive selve) has replaced (or joined) another in a single body (Parfit, 1984; Radden, 1996; Berghmans, 1997a). The normative question then is what the moral authority is of previously expressed wishes of a person in case deep psychological changes have taken place (Buchanan, 1988).
Such arguments from personal identity against (not only psychiatric) advance directives are a radicalisation of the moral issue of autonomy discussed above. The problem of choosing between sets of preferences of a person is transformed into the problem of choosing between sets of preferences of different persons. From a narrative perspective, these arguments rightly question the assumption of a given unity of the person, underlying principlist pleas for advance directives. Yet, by hypothesizing two different personalities, these counterarguments in fact are based upon the same, rigid notion of personal identity. A narrative concept of identity does not imply that the person stays the same, but neither does it entail the assumption that drastic changes in a person's psychological condition make it impossible to use the concept of identity any longer. From a narrative perspective, a person's identity is formed in stories, which both express and create the unity of a person's life (Ricoeur, 1986, 1991; Widdershoven, 1993). As stories, advance directives presuppose the unity of the patient's life, and try to contribute to that unity, not by making the different phases identical, but by trying to create a meaningful whole which covers all of them.

Misuse and abuse

Psychiatric advance directives are open to misuse and abuse. Several limitations and procedural safeguards suggested by advocates of self-binding contracts would reduce its potential for abuse (Howell, Diamond & Wikler, 1982; Dresser, 1982). Contracts would be permitted only when the individual's illness is recurrent, interspersed with periods during which behavior was relatively symptom-free. Above that, it would be required that the person involved has experienced a specified number of psychotic episodes in the past, and contracts would be permitted only when the individual's disorder was responsive to treatment. Moreover, it would be required that the patient's disorder be in remission at the time the contract was made, and the individual's legal competence at the time of the contract formation would have to be established. Contract intervention would be permitted only when the individual's behavior reached psychotic proportions. The contract proposal envisions some third party involvement to ensure that the patient's best interests are served (during contract negotiations as well as during the execution of the contract). The contract would be valid for a limited time, and patients could renegotiate or revoke the agreement at any time other than during a relapse as defined in the contract. Contract sanctioned commitment and/or treatment would be allowed to continue for only a short time (for instance 3 weeks). Court involvement would remain a right of the patient (i.e. the right to a hearing within seventy-two hours of commitment, the purpose of which being to determine whether the terms of the written agreement had been met).
All these provisions are sound, in so far as they might protect the patient against the abuse of power by health care professionals. Yet, such answers to the problems of potential misuse again are based upon a conception of advance directives as individual acts of self-determination, rather than as elements of the patients life-story, which is developed in interaction with others, including the psychiatrist. They presuppose that the self-binding contract should ideally be based upon the informed and voluntary consent of the patient. From a narrative perspective, this notion of informed consent as the basis of treatment decisions is problematic. It starts from individual choices, rather than from joint processes of negotiaton and deliberation (Moody, 1992; Emanuel & Emanuel, 1992). It focuses on individual decision-making, and overlooks the narrative work which is part of creating a practice of care (Tronto, 1993; Widdershoven & Smits, 1996). If the advance directive is conceptualized as the temporary sedimentation of an ongoing process of joint narrative work, involving patient and physician alike, the focus is not upon potential infringements upon the patient's free will, but on possible frustrations to the relationship of mutuality and trust between patient and physician. Moreover, from a narrative perspective, what counts as misuse cannot be seen as something to be determined by external procedures, but as an issue to be discussed in the ongoing discourse between patient and physician. The issue of abuse is not something external to the Ulysses contract, it is an essential part of the narrative construction of the advance directive itself, as the issue of dealing with the risk of abuse is part of practices of care in general.

Conclusion

We have presented a narrative approach to advance directives in psychiatric care as an alternative to the principlist approach, common in medical ethics (see also Widdershoven, forthcoming). It was argued that the narrative approach, by focusing upon the mutual responsibility op patient and physician in a caring practice, opens new perspectives on the ethical problems related to advance directives in psychiatry. From a narrative perspective, such contracts are not to be regarded as expressions of self-determination, but as attempts to develop a practice of care, based upon trust and cooperation between physician and patient. This does not imply that psychiatric advance directives are morally justified per se. It rather means that advance directives can be morally justified in so far as they fit in with and contribute to the patients life narrative, and enable those involved to transform the patient's life history into a meaningful unity, including elements of frustration, pain and sorrow.

Literature
- Agich, G.J., Autonomy and long-term care. New/York / Oxford: Oxford University Pres, 1993.
- Backlar, P., The longing for order: Oregon's medical advance directive for mental health treatment. Community Mental Health Journal, 32, 1995, 103-108.
- Backlar, P. & B.H. McFarland, A survey on use of advance directives for mental health treatment in Oregon. Psychiatric Services, 47, 1996, 12, 1387-1389.
- Berghmans, R.L.P., Zelfbinding in de psychiatrie. Ethische aspecten. [Self-binding in psychiatry. Ethical aspects] Tijdschrift voor Psychiatrie, 36, 1994, 9, 625-638.
- Berghmans, R.L.P., Ethical hazards of the substituted judgement test in decision making concerning the end of life of dementia patients. International Journal of Geriatric Psychiatry, 12, 1997a, 283-287.
- Berghmans, R.L.P., Coercive treatment in psychiatry. In: Encyclopedia of applied ethics. Volume 1 (R. Chadwick, Ed.). New York: Academic Press, 1997b, 535-542.
- Brock, D.W., A proposal for the use of advance directives in the treatment of incompetent mentally ill persons. Bioethics, 7, 1993, 2/3, 247-256.
- Buchanan, A.E., Advance directives and the personal identity problem. Philosophy & Public Affairs, 17, 1988, 277-302.
- Culver, C.M. & B. Gert, Philosophy in medicine. Conceptual and ethical issues in medicine and psychiatry. New York/Oxford: Oxford University Press, 1982.
- Dresser, R., Ulysses and the psychiatrists: a legal and policy analysis of the voluntary commitment contract. Harvard Civil Rights - Civil Liberties Law Review, 16, 1982, 777-854.
- Dresser, R., Advance directives, self-determination, and personal identity. In: C. Hackler, R. Moseley & D.E. Vawter (eds.), Advance directives in medicine. New York: Praeger, 1989, 155-170.
- Dworkin, G., The theory and practice of autonomy. New York: Cambridge University Press, 1988.
- Elster, J., Ulysses and the Syrens. Studies in rationality and irrationality. Cambridge: Cambridge University Press, 1979 (Rev. edition 1984).
- Emanuel, E.J. & L.L. Emanuel, Four models of the physician-patient relationship. JAMA, 267, 1992, 2221-2226
- Feinberg, J., The moral limits of the criminal law, Volume III: Harm to self. Cambridge: Cambridge University Press, 1986.
- Goldman, D.L. & R.M. Post, The puzzle of noncompliance in the manic patient. Bulletin of the Menninger Clinic, 55, 1991, 248-253.
- Gubrium, J.F. & D.R. Buckholdt, Toward maturity, San Fransisco: Jossey-Bass, 1977.
- Howell, T., R.J. Diamond en D. Wikler, Is there a case for voluntary commitment? In: T.L. Beauchamp en L.R. Walters (Eds.), Contemporary issues in bioethics. Second edition. Wadsworth Publishing Company, Belmont, 1982, 163-168.
- Husak, D.N., Paternalism and autonomy. Philosophy & Public Affairs, 10, 1981, 27-46.
- Lavin, M., Ulysses contracts. Journal of Applied Philosophy, 3, 1986, 89-101.
- Moody. H.R., Ethics in an aging society. Baltimore: The Johns Hopkins University Press, 1992.
- Nussbaum, M.C., The fragility of goodness. Luck and ethics in Greek tragedy and philosophy. Cambridge; Cambridge University Press, 1986.
- Parfit, D., Reasons and persons. Oxford: Clarendon Press, 1984.
- Radden, J., Choosing to refuse. Patients rights and psychotropic medication. Bioethics, 2, 1988, 83-102.
- Radden, J., Divided minds and successive selves. Ethical issues in disorders of identity and personality. Cambridge/London: The MIT Press, 1996.
- Ricoeur, P., Time and narrative. Chicago: University of Chicago Press, 1986
- Ricoeur, P., Narrative identity. In: D. Wood (ed.). On Paul Ricoeur. Narrative and interpretation. London: Routledge, 1991, 188-199.
- Rogers, J.A. & J.B. Centifanti, Beyond 'self-paternalism': response to Rosenson and Kasten. Schizophrenia Bulletin, 17, 1991, 1, 9-14.
- Rosenson, M.K. & A.M. Kasten, Another view of autonomy: arranging for consent in advance. Schizophrenia Bulletin, 17, 1991, 1-7.
- Savulescu, J. & D. Dickenson, The time frame of preferences, dispositions, and the validity of advance directives for the mentally ill. Philosophy, Psychiatry & Psychology, 5, 1998, 3, 225-246.
- Schelling, T.C., Choice and consequence. Cambridge/London: Harvard University Press, 1984.
- Tronto, J.C., Moral boundaries. A political argument for an ethic of care. London: Routledge, 1993.
- Widdershoven, G.A.M., The story of life: hermeneutic perspectives on the relationship between narrative and life history. In: R. Josselson & A. Lieblich (eds.). The narrative study of lives. Vol. 1. Newbury Park: Sage, 1993, 1-20.
- Widdershoven, G.A.M., Alternatives to principlism in bioethics. In: K.W.M. Fulford, D. Dickenson, T. Murray (eds.). The Blackwell reader in health care ethics and human values. Oxford: Blackwell, forthcoming.
- Widdershoven, G.A.M. & M.J. Smits, Ethics and narratives. In: R. Josselson (ed.). Ethics and process in the narrative study of lives. Vol 4. Newbury Park: Sage, 1996, 275-287.

Notes

(1). in a number of states in the U.S. (a.o. Oregon and Maine) psychiatric advance directives have a legal status (Backlar, 1995; Backlar & McFarland, 1996).