Abstract
Community psychiatry, as adopted in many developed countries, involves radical changes in practice. It differs distinctively from previous hospital-based or office-based practice, and some treatment approaches depart dramatically from what we have previously known. Developments have occurred rapidly; in the first two editions of this volume on psychiatric ethics no need was perceived for a separate review of the ethics of community psychiatry.
Just how far-reaching the changes might prove is attested by a 1993 editorial by Bell [1]: ‘By 2000, mental health laws will have swung back in the direction of permitting involuntary treatment of severely and persistently mentally ill patients based on psychiatric need: stringent commitment criteria will no longer allow patients who seriously need help to go untreated. Further, the restrictive confidentiality Iaws that inhibit mental health professionals’ communication with family members about patients' conditions and treatment plans will be long gone… Problems with continuity of care will be lessened by a unified computer network that provides authorised access to medical records and permits knowledgeable treatment of patients regardless of where they seek care… More CMHCs (Community Mental Health Clinics) will be involved with residential alternatives and vocational training… (There will be) linkages with correctional facilities that will make after care available to incarcerated mentally ill persons after their release… The increase in partnership between family members and CMHCs will form a strong lobby that will bring a common sense approach to the care of patients most in need’.
Ethical implications of the new kinds of practice in community psychiatry have been neglected. It is not clear why this is so; perhaps the forces, ideological and fiscal, driving change are too powerful to be questioned.
My aims in this paper are: (i) to outline the new model of care and the climate in which it has developed; (ii) to identify the ethical challenges resulting from this new model of care and the need for an appropriate ethical framework; and (iii) to look at practical approaches to these challenges.
There will be a British bias in this paper, but the examples presented will almost certainly be paralleled by similar ones in most other countries and services.
Philosophy and practice of community psychiatry
In this paper I present a fairly detailed account of community psychiatry since many are still not aware of its philosophy, organisation and objectives.
Community psychiatry involves a change in the locus of care (from hospital to community), funding arrangements, and treatment techniques (many derived from theories of social psychiatry). It aims to establish a network of services offering crisis intervention, continuing treatment, accommodation, occupation and social support which together help people with mental health problems to retain or recover social roles, as close to normal as possible for them [2]. In most countries, the focus of the services is on those with serious mental illness, especially the psychoses. A useful perspective is to see community services as attempting to provide outside the old mental hospitals, the complex range of care once provided within. These covered not only health needs but also social, accommodation, occupational and leisure facilities, as well as the provision of asylum.
The scale of deinstitutionalisation has been astonishing. In England, the number of psychiatric beds decreased from a peak of 152 000 in 1954 to 39 500 in 1993, a reduction of 74%. This has been associated with a large increase in admissions but also a greatly decreased number of patients with a protracted stay in hospital. In the US, occupied psychiatric beds in state hospitals have fallen from 339 per 100 000 population in 1955 to 29 per 100 000 in 1994 (cited in Lamb) [3].
Case management and assertive community treatment
To ensure that patients in the community receive the benefits of the range of services which they may require, the widespread practice of ‘case management’ has been adopted. The aims of case management are to ensure continuity of care, accessibility to often fragmented and independently managed services, accountability and efficiency. The core functions usually include: assessing of patient needs; developing a comprehensive care plan; arranging service delivery; monitoring and assessing services; evaluating progress and follow-up [4, [5]].
Although the practice of case management varies, two general approaches can be identified. Service ‘brokerage’ case management sees the ‘case manager’ as an enabler, systems coordinator or broker of services. In ‘clinical’ case management, however, the professional has a direct treatment relationship with the patient, often being personally involved with aspects of the patient's psychological, physical and social care [6]. In the public sector, case management is seen as ensuring access to care for those too disabled to seek it themselves. Within the private sector, as sometimes in the United States, case management may be used as a mechanism of utilisation review, limiting access to some services in order to contain costs [7].
Research on simple implementations of case management tends to show relatively poor effectiveness [8, [9]]. As a result a more intensive model of case management is being increasingly adopted in community mental health services, usually known as assertive community treatment (ACT). Assertive community treatment aims to provide a comprehensive care package including treatment and support services via a multidisciplinary team within the community. It includes frequent contacts with patients in the community (often at home), 24 h availability, direct responsibility of staff for a broad range of interventions and low staff:patient ratios.
The ACT model was first evaluated in the early 1970s in Madison, Wisconsin [10], and has been followed by a number of replication studies in Australia, Michigan, California, and London [11]. Most ACT programs result in a reduction in the use of psychiatric inpatient care partially offset by additional use of other community based services. Positive effects on symptoms, social functioning and disability are less evident. They do have a positive impact on residential stability and result in better retention in treatment [9]. Cost savings result where there has been a reduction in the use of inpatient care. Costs are shifted to community-based services but there is often a net saving despite this [12, [13]].
A major focus of ACT is to prevent the severely mentally ill patient from dropping out of treatment since loss of contact is likely to lead to relapse and the need for admission to hospital. Admission tends to be seen as a last resort, and as often representing a failure of continuity of care. An assertive approach, bringing treatment to the patient aims at preventing the patient falling through gaps in the net of services. Lapsing from treatment may be an early sign of relapse. The community team may then actively seek out the patient to re-establish contact. Uninvited visits to the patient's home and sharing concerns about the patient's risk with other agencies may follow. Interventions may occur in public places so that patients, their treatment, and the role of mental health teams become more visible to the public.
In contrast to the hospital, the community is a complex environment in which to provide care. Patients with severe mental illness have a complex range of needs that can only be met by a range of separate services and agencies. For the patient to have access to these, there needs to be a substantial flow of information between them concerning the patient. Furthermore, the key worker or other members of the multidisciplinary team tend to develop a special kind of relationship with the patient. They provide a broad treatment approach. As well as medication and conventional psychological interventions, they work with the patient in their ordinary community settings to rehabilitate basic living skills. They may: help with budgeting, shopping and cooking; attend appointments with the patient to see other professionals; advocate for services; and work closely with carers, housing officers and other figures in the patient's social network. This special relationship may be used to encourage the patient to comply with treatment, especially prophylactic medication.
In the background may loom important issues of another kind. The shift from hospital to community may exert pressure to discharge patients as quickly as possible. Still symptomatic and disabled patients then require intense and controlling management in the community. Lack of responsiveness from other agencies (e.g. finding suitable accommodation) may provide a temptation to withdraw some responsibilities in an effort to stimulate a response.
The community's response to community psychiatry
Public fears are common that care in the community for mentally ill persons will be a failure. At times, this fear may amount to a ‘demonisation’ of the mentally ill who are seen as posing a serious risk to innocent members of the community. Responses to these fears by government and others may significantly affect the practice of community mental health teams.
In England, for example, such fears have been fuelled by highly publicised incidents involving people with mental illnesses. In December 1992 Christopher Clunis, a man with schizophrenia, stabbed to death a complete stranger on an underground station platform. This led to a prominent public inquiry, the Ritchie Inquiry, which described Clunis' care as ‘a catalogue of failures’ [14]. Among the report's many recommendations were the need for improvement in care planning and interagency working, for highly intensive care for difficult patients, for training of ‘key-workers’ (case managers) and for better assessments of risk. Subsequently, almost every homicide by a person with a history of contact with mental health services has been publicised, often sensationally, and has been the subject of an independent public inquiry. These have all reached more or less the same conclusions: that contributing to the tragic outcome have been poor communication between agencies, poor assessment of the risk of violence, poor liaison with police and probation services, and barriers posed by ‘confidentiality and professional ethics’ [15].
In England, the government's response to the public's fear has been to issue ‘guidance’ and new legislation. The former has taken the form of the ‘Care Program Approach’ (CPA) and the ‘supervision register’. The CPA essentially institutionalises the core principles of case management: assessment of the health and social needs of people accepted by specialist psychiatric services; formulation of a care plan which addressed these needs; appointment of a key-worker to keep in close touch with the patient and monitor care; and, regular review, and if need be, agreed changes to the care plan. Health-purchasing authorities monitor the implementation of this approach including the requirement that registers of patients on the CPA be maintained. In 1994, the government introduced guidance for ‘supervision registers’ for patients, identified during a care program review meeting, as being at ‘significant risk of committing serious violence or suicide, or of severe self-neglect in some foreseeable circumstances which it is felt might well arise in this particular case (e.g. ceasing to take medication, misusing drugs or alcohol, loss of a supportive relationship, or of accommodation)’ [16, [17]]. In practice, their focus has usually been on the risk of harm to others. Services are required to identify and give priority, of an only vaguely suggested nature, to these patients.
In 1995, legislation was passed establishing ‘supervised discharge’, ‘after care under supervision’ for vulnerable former inpatients previously detained compulsorily under the Mental Health Act. This provides powers ‘to convey’ a patient to a place of treatment, occupation, education or residence by a community supervisor (e.g. case manager) but does not permit compulsory treatment itself. Most clinicians view supervised discharge as ill-considered and unworkable [18].
Most commentators see the CPA and supervision register as having two implicit functions: first, identifying individuals (key-workers) who can be held responsible for the dangerous behaviour of their patients, while deflecting attention from resource deficiencies [19]; second, a political purpose, to reassure the community that they are being protected from dangerous patients. The term adopted, ‘supervision’ register rather than say, ‘at-risk’ or ‘priority’ register, supports this contention.
As a result of such fears and reactions to them clinical practice has changed in ways which have ethical import. Between 1989 and 1993 the number of compulsory admissions to hospital increased by 27%. Mental health professionals recognise that a public inquiry following a homicide by one of their patients (or ex-patients) threatens their professional reputation and badly damages the morale of their service. Under such circumstances, it would not be surprising to find their practice moving in the direction of greater control over their patients at the expense of autonomy [19].
Responses also occur at an institutional level. Increasing attention is being given to ‘clinical risk management’ in psychiatry. Lipsedge [20] has summarised the factors contributing to the clinical risk in psychiatry. These include: ‘professional arrogance combined with a reckless tolerance of deviance leading to a failure by mental health professionals to heed reports by carers and members of the public about disturbed behaviour’ ‘undue emphasis on the civil liberties of psychiatric patients at the expense of tolerating grave suicidal risk and the danger of violent behaviour’ ‘failure to pass on information about potential dangers to other professionals, such as hostel staff, for reasons ranging from inertia, inefficiency, or over work to a misguided over protective view of the patient at the expense of the safety of potential victims’. Ethical implications abound, as we shall see in the next section.
Harrison [21] has summarised the current position in England as follows: ‘The danger is that given the low predictability of risk factors in relation to individual patients, large numbers may be drawn into supervision programs having a custodial and coercive therapeutic focus. …Many patients require intensive care, perhaps over long periods of time. But there is a significant risk that mental health professionals will resort to inpatient care, or to over-restrictive styles of therapeutic care, because of risks not to their patient but to themselves should something go wrong’.
While especially prominent in England, similar pressures inflate anxieties of psychiatrists elsewhere. Appelbaum [22], from an American perspective, has noted the tendency to err on the side of involuntary commitment because of societal pressure. In some states in the US, there is a ‘Tarasoff-type’ duty to protect, and if necessary to warn, potential victims when a patient poses a serious threat of physical harm. At least one State, Vermont, has extended the duty to protect to include property [23]. The potential for litigation by third parties harmed by patients has probably changed clinicians' behaviour significantly, but the negative effects of Tarasoff-like court decisions have in the main been less damaging than feared. Subsequent judicial, legislative and clinical responses have muted the more unrealistic expectations [24].
In relationship to community psychiatry, it is noteworthy that the original Californian Tarasoff decision occurred at a time when mental hospitals were being viewed increasingly negatively and commitment laws were becoming more restrictive [25]; there nevertheless remains a persistent demand (if not met by walls then by duties to protect imposed on professionals charged with the care of the mentally ill) for the public to be kept safe from the ‘insane’. Well-publicised local incidents may result in a temporary increase in admissions; in England, where nearly all are publicised, the effect becomes persistent.
Ethical dilemmas
In this section, I attempt to characterise the core ethical issues arising in community clinical practice. They can be grouped under four headings: privacy, confidentiality, coercion, and conflicts of duty.
Privacy
Assertive treatment programs bring treatment to the patient, often in the patient's residence whether it be home, hostel or boarding house. If the patient is regarded as being at risk of relapse, visits may be made by members of the community mental health team even when uninvited. Indeed, visits may continue even when the patient's explicit desire is that they cease. Since much treatment occurs in the community, there is also an increased risk that it becomes public. The curiosity of neighbours may be aroused, particularly with repeated visits, and especially if attempts to gain entry are rebuffed by the patient. Neighbours and other members of the public may deduce that a mental health team is involved and that the person being visited is a patient.
Furthermore, as treatment becomes more visible to the public, new expectations may be generated that a community mental health team can be called to deal with a disturbed or difficult person suspected to be a patient. Even if a public assessment is not carried out, an acknowledgement by the mental health team that they may have a role may reveal to bystanders that the difficult person is a mental patient (if already so) or label them as one (if not).
The ‘Supervision Register’ in England, discussed earlier, illustrates most of my points concerning privacy. An aim is for the key-worker to maintain contact with the patient. The register is meant to act as a ‘flagging device’ to staff that there is a significant risk. It is also meant to ensure that identified patients receive the care they need, and that resources are targeted on this group, even though the introduction of supervision registers was not accompanied by any increase in resources. It may be difficult for some patients ever to be discharged from the register as many of the factors suggesting a risk of violence (e.g. a past history of violence) will persist. If the patient defaults from treatment, the key-worker and other members of the team are enjoined to make every effort to re-establish contact. This includes contacting the carer or family, consulting the general practitioner, checking with local accident and emergency departments for any contacts, and checking with neighbouring mental health units. Immediate consideration is also to be given ‘to the need to inform the police service if contact is lost with a patient who could pose a significant risk to him/herself or others’ [17]. The nature of the relationship between clinician and patient may shift from care to supervision, akin to that between a probation officer and client. Is the supervision register more likely to ensure that patients receive the care they need, or to stigmatise them with the label of ‘dangerous’, leading to exclusion from community services or amenities, including housing?
The patient has a right to know that he or she is on the register, unless it would cause ‘serious harm to his or her physical or mental health’. The patient may request removal, but it is up to the psychiatrist in discussion with the multidisciplinary team to decide. If the patient remains dissatisfied, the normal channels for complaint and the right to a clinical second opinion apply.
The status of ‘guidance’, such as that issued concerning supervision registers [16], is not law but, if challenged, the clinician is liable to criticism, for example by an independent inquiry, and possibly to a charge of negligence, if he or she cannot show that account was taken of the guidance. The impact on practice is thus major.
Confidentiality
Confidentiality may be considered an aspect of privacy in which information obtained from a patient on the assumption that it will not be disclosed to others, is disclosed. As the patient is commonly treated by a multidisciplinary team, sharing of information among its members is common. Patients may not know that this is to be expected. More complex is the sharing of information between agencies: health, social, voluntary, housing and so on. Very needy patients' access to benefits and other goods may depend on information about them being revealed to those in a position to supply them. Since information may flow frequently, confidentiality may receive less emphasis. There may be what might be called the ‘cat already out of the bag’ phenomenon, and an attitude that ‘the patient has less to lose by certain breaches of confidentiality than other kinds of patients do’ [26].
Confidentiality may be breached ostensibly in the interests of the patient as above, or for the protection of others. The latter is considered below, including the interests of family and carers.
Coercion
‘Coercion’ is a complex notion. A range of pressures may be exerted by community mental health teams to gain the patient's cooperation with treatment. The least problematic is ‘persuasion’, an appeal to reason. Since the key-worker may have established a relationship with the patient, broader in scope and more intimate than the conventional patient-clinician relationship, opportunities for other kinds of pressure arise. ‘Leverage’ or ‘interpersonal pressure’ may be exercised through the patient's emotional dependency on the key-worker [27]. A vague threat may be implied of something being removed, or the patient may simply want to please someone who has proved helpful. More obvious threats, veiled or direct, may be employed. For example, the patient may be threatened with the loss of a valued aspect of their treatment or of a social benefit if they fail to comply with treatment, especially medication.
When does ‘leverage’ become ‘coercion’? A helpful account by Wertheimer [28] suggests that no simple definition can be offered for the term ‘coercion’. The issue for us in mental health care is whether there are morally relevant distinctions between different kinds of pressure or force on patients to comply with treatment. A helpful view of coercive proposals may be that ‘threats’ coerce, but ‘offers’ do not. The former threatens to make the recipient worse off than they are at present, while declining an offer does not. Whether the subject is worse off depends on what has been termed the ‘moral base line’. Threatening to remove something to which the subject is ‘entitled’ (e.g. a housing benefit determined by statute) makes the subject worse off if he or she does not accede. An offer of something which is not an entitlement but is in the nature of special assistance (e.g. a mental health worker having a connection with a second-hand furniture dealer who gives special discounts) made on condition that the patient complies with the treatment would, if rejected, not make the patient worse off compared with the relevant moral base line: what his or her position would have been if the offer had never been made. One imposes a penalty, the other offers a reward.
On this account, ‘coercion’ can be distinguished from other kinds of pressure (e.g. persuasion or inducement by making offers). It refers to propositions which, under normal circumstances, ought not to be made and which are aimed at motivating a person to act in a way desired by the coercing agent. Thus, conditional access to monetary benefits (statutory entitlements), as occurs when some patients in the United States have a ‘representative payee’ under Supplemental Security Income/Social Security Disability Insurance (SSI/SSDI), who only gives the patients their benefits when they comply with treatment is coercive [27]. Other features of coercion may include a feeling of constraint in the subject, a restriction on autonomy, a reduction of choice, and the mitigation of moral or legal responsibility (including freely given consent).
Deception, less equivocally seen as morally wrong, may be another way of inducing patient compliance. It is possible that some outpatient commitment orders depend for their effectiveness on a patient's misconception concerning the consequences of not complying. The patient may falsely believe that transgression of the order will result in re-hospitalisation or enforced treatment, but it may in fact only permit the conveying of the patient to a treatment facility (as in ‘supervised discharge’ in England, discussed earlier).
The most direct form of force is a compulsory treatment order. With a change of treatment to the community, a number of jurisdictions have introduced outpatient commitment orders carrying varying powers. They may allow commitment to outpatient treatment instead of inpatient treatment as a less restrictive alternative; permit earlier conditional release from inpatient commitment; or allow preventative commitment [27]. The third category is the most controversial since it may allow the compulsory treatment of a patient who is not currently at risk. Based on a proven record of relapse when treatment is discontinued, and of dangers previously demonstrated when relapse has occurred, compulsion is used to avert future risk. Outpatient commitment orders are not dealt with further here.
Increasing financial constraints set limits on what is possible in a mental health service. It has been suggested that some calls for increased coercion on patients to comply with treatment are an inappropriate attempt to compensate for money-starved, poor quality services. Clinicians may find themselves having to discharge patients before they are really ready because of insufficient beds and to consider controlling treatments in the community as a consequence. A patient may prefer to forgo medication after discharge from hospital despite a high risk of relapse and readmission. If admissions are expensive or beds are scarce, clinicians may sometimes be inclined not to adequately consider such a patient's wishes.
Conflict of duty to patient versus others
Risk of harm to others
As previously discussed, the current climate in which community psychiatry is practised often raises the question of the degree to which the psychiatrist has a duty to protect others. If a specific risk to an identified person is established, the clinician's duty to protect that person is clearer. When the risk to others is general, judgements are harder.
An important statistical fact is germane to risk of serious violence to others. Despite the public's fears, acts of serious violence are rare. The prediction of a rare event inherently lacks precision. A predictive instrument, even with high sensitivity and specificity, will seriously err in the direction of false positives. This can be demonstrated by a study involving patients treated by my own community mental health team in Camberwell, South London. Seventy-three of 320 patients seen in 1 month were assessed by clinicians as posing a moderate or severe risk of violence to others over the next 6 months. Using the best available predictive statistics from large scale predictive studies we determined that seven of the 73 patients would be likely to be violent to person in the next 6 months. But so also would 10 of the 245 low risk patients. Indeed, more violence was thus to be expected from the low risk group than the high risk group!This limitation in prediction must be borne in mind. What is the social and moral cost of unnecessarily restricting a large number of patients in order to prevent (probably inefficiently) harm by a few? Prediction is better the higher the base rate of violence in the population of patients who might pose a risk, for example, in forensic patients. Not able to be easily assessed is the possibility that clinicians are better able to predict danger in the very short-term (that is, in the next few hours or days).
The often visible nature of community treatment may change expectations of members of the public. For example, the team may be asked to intervene by members of the public (for example neighbours or shop keepers) when they are disturbed by a patient's behaviour. They may be frightened or simply wish that a nuisance be removed. A further aspect of such situations may be the possibility that if the team does not act, prejudice against the patient will increase and his or her community tenure be threatened. The balance between the clinician's duty of care to the patient and to the local community may be difficult to determine. An intervention by a worker in a public place to allay anxiety may have serious implications for the patient's privacy and confidentiality.
Community psychiatrists are expected to be competent in assessing risk to others as well as to patients themselves. This often requires information from a range of informants, particularly concerning previous incidents of violence and risk factors such as substance abuse. Sometimes, the mere seeking of information from others concerning the patient's past behaviour may reveal that the person is being treated by a psychiatric team. It may even raise unwarranted anxieties in their minds.
Informal carers
Informal carers, usually family, are central to a successful policy of community care. However, the extent to which their own needs should be met is cloaked with uncertainty. Where there is a danger of serious physical harm to the carer, the clinician's responsibility is usually straightforward. Far more common are less serious threats to their wellbeing which nonetheless cause substantial suffering. They may experience difficulty in coping with burdensome behaviours, lack important knowledge about their relative's illness, and may not know to whom to turn for support [29, [30]]. The patient may prohibit the clinician any contact with the family. In these circumstances, it is unclear to what extent the community mental health team owes a duty of care to the family. This question has been largely ignored in the literature.
Approaches to ethical problems in community psychiatry
In this last section, I shall suggest ways of helping clinicians resolve the ethical dilemmas discussed above.
Increasing patient involvement in their care
Patients are likely to feel less coerced the more they play an active role in their treatment. The current trend for increased service users' consultation and involvement in determining the shape and nature of local mental health services may be helpful in this regard. There are also initiatives at the individual patient level worth exploring. In Britain, ‘crisis cards’ originated as a voluntary sector initiative to facilitate access to an advocate and to state a patient's preferences for treatment in an emergency when he or she might be too unwell to express their wishes coherently [31]. They are designed to be carried by the patient and have the potential to record a range of useful information about the patient's treatment plan as well as advance statements. ‘Crisis Cards’ have usually been drawn up by the patient alone, without discussion with the treatment team. There is scope for this idea to be developed in to what we have termed ‘joint crisis plans’. Here the content of the card, while still ultimately determined by the patient, is negotiated with the treatment team. The aim is to reach agreement on the care plan. This occurs when the patient is well enough to make competent judgements about what is in his or her best interests.
We have studied the utility of crisis cards and joint crisis plans in Camberwell. Participation was offered to all patients with a psychotic illness who had had at least two previous admissions to hospital. Forty percent wished to develop a card. They chose to include a wide range of information including diagnosis, current treatment, contact information for carers and professionals, first signs of relapse and the preferred treatment for these, treatment preferences and refusals for an established relapse, indications for admission, and practical requests (e.g. who should ensure that domestic arrangements are not neglected in the case of admission). Independent assessment of the process showed that it was experienced by the patient as non-coercive and that they felt their voice was heard. Follow-up six to 12 months later showed that for patients who had experienced a crisis, the card was consulted in most cases. The information was considered helpful by patients and mental health professionals. Over 50% of the patients reported feeling more involved in their care, more positive about their situation and more in control of their mental health problem. Forty percent of patients reported feeling more likely to continue with their treatment as a result of developing the card. We concluded that a joint crisis plan could serve both a manifest, practical function in a crisis (to provide important information when the patient is too ill to do so), as well as a latent, psychological one (positive effects on the patient's attitudes to themselves, their illness and treatment, and their relationship with the clinical team).
Cards such as these also have the potential for further transformation into ‘psychiatric advance directives’ which would be legally binding. Such a directive, anticipating relapse of a psychosis develops the concept of the ‘living will’. It could reconcile two apparently contradictory themes in the current practice of psychiatry: on the one hand, the call to provide non-consensual treatment in the community; and on the other, the promotion of patient autonomy [32]. They could for some be an alternative to community treatment or outpatient commitment orders. Challenges to their implementation are discussed by Halpern and Szmukler [32] and include the problem of evaluating the patient's competence: when is it adequate to make a directive, sufficiently inadequate to trigger its application, and adequate to revoke a directive; ensuring the absence of coercion in drawing up the directive; mechanisms for appeal in cases of dispute; resource implications; and, clarifying relationships between advance directives and existing mental health legislation? Psychiatric advance directives are possible now in some jurisdictions in the United States [33, [34]], but they have been little taken up presumably because of such difficulties. Despite this, advance directives might still find a place for some groups of seriously mentally ill patients.
Crisis cards and their variants could significantly reduce the number of situations in which clinicians find themselves needing to act against the patient's wishes expressed when he or she is unwell rather than well. Many of the ethical dilemmas discussed above occur at precisely such times of crisis.
Grounds for paternalism
Most invasions of privacy, breaches of confidentiality and uses of coercion rest on a paternalistic justification: ‘intentional non-acquiescence or intervention in another person's preferences, desires or actions with the intention of avoiding harm to or benefiting the person’ [35]. There is debate whether overriding the wishes of a person who is unable to exercise substantial autonomy is paternalistic. For the clearly non-autonomous person acting beneficently may be termed ‘paternal’ [36].
Despite the differences in views concerning the exact nature of paternalism, given the instances in which clinicians in the community may act without the patient's agreement, staff in community mental health teams must develop a working knowledge of acceptable grounds for paternalistic interventions. This may be difficult to achieve: ‘developing a position on issues of paternalism is a matter of appreciating the limits of principles and the need to give them additional content, while attempting to render one's consequent rules and judgements as coherent with other commitments as possible. The problem of medical paternalism is the problem of putting just the right specification and balance of physician beneficence and patient autonomy in the patient-physician relationship. It is a messy complicated problem, and coherence is difficult to achieve. Paternalistic intervention requires persons of good judgement as well as persons with well-developed principles able to confront contingent conflicts’ [35]. We should aim to train such persons.
The approach to paternalism of Culver and Gert [36] is helpful. A person is acting paternalistically to another if his action benefits the other; his action involves violating a moral rule with regard to the other; his action does not have the other's past, present, or immediately forthcoming consent; and the other is competent to give consent. Culver and Gert regard the subject as able to give consent if he is able to understand the information and to appreciate it (i.e. to understand why it is being asked for). Consent here is taken to include simple and possibly ‘valid’ consent. In attempting to justify a paternalistic act, at least five questions can be asked: (i) what are the moral rules which are violated when the clinician acts against the patient's wishes (e.g. deceit, limiting freedom of choice, causing psychological pain) and what are the evils thus to be perpetrated on the patient and for how long will they last; (ii) what is the seriousness of the evils to be avoided through the paternalistic intervention (e.g. death, disability, worsening of the psychiatric disorder), and what is their likelihood; (iii) how does the clinician rank the sets of evils above compared to the patient; (iv) is the patient's preference when comparing the evils to be avoided with the evils to be incurred, irrational (i.e. the patient has no rational reason to prefer an outcome with apparently greater evils); (v) can the clinician advocate publicly for his ranking of the evils to be perpetrated compared to those to be avoided? In other words, would all or most rational people agree that this kind of moral violation should in such circumstances be universally allowed?
Conflicts of duty
Risk of harm to others
There is a distinction to be drawn between the danger to self (leading to paternalistic interventions) and a danger to others. As Culver and Gert [36] point out, in the latter, the question of the patient's ‘rationality’ is barely relevant, only being so in relation to the longer term harms it could be anticipated will be suffered by the patient as a result of his or her violent actions to another. The key question is whether the potentially dangerous person has volitional ability to refrain from acting dangerously (i.e. the degree to which ‘unvoluntariness’ is present). The ethical basis for the supervision register in England seems to confuse these issues. Dickenson [37] points out that if the purpose of the register is to ensure that a key-worker is held responsible, or to protect the community against risk, there is no justifying argument from ‘benevolence’ or paternalism. On what basis then is it proper for a patient to be fit to be discharged from hospital, yet to not be well enough to enjoy the civil rights of those never mentally ill? What level of risk to others justifies such an infringement of autonomy; is a non-specific risk to the community sufficient?
Risk assessment is now at the crux of managing potentially dangerous behaviour. It thus becomes essential that clinicians adopt risk assessment approaches which represent best practice and are based on the best available research evidence. A process of self-interrogation regarding whether a particular assessment is defensible scientifically as well as clinically is desirable. Discussion with colleagues, both in and outside the multidisciplinary team helps. The high proportion of false positives must also be remembered. Factors such as a history of previous violence (and its details), potential disinhibiting factors (especially substance misuse, provocative social settings), treatment compliance, stresses and available social supports, and particular features of the mental state (threat/control override symptoms, emotions related to violence, specific threats) need special consideration [38]. Collateral information from informants is essential. It is important to establish the relationship between dangerousness and the patient's mental illness; psychiatric interventions are only justified when the mental illness significantly increases risk. On the basis of the risk assessment, a clinical management plan can be devised which offers the least restrictive approach to the patient.
Elements of the ‘crisis card’ can also be brought in if possible. Discussion with the recovered patient, previously violent, may lead to agreement between patient and clinician at precisely what point (e.g. recurrence of a specific delusion; an unprovoked outburst of anger to a specified person) it would be appropriate to intervene against the patient's wishes if a relapse were to occur.
At the same time, community psychiatrists should engage in a dialogue with those representing the community about the balance to be struck between risk to others and restrictions on the liberty of mentally ill people. The public must be helped to recognise that risk-taking is fundamental to community care and that inevitably there will be tragedies. It should also be pointed out that despite media perceptions, there is no good evidence of an increase in homicides by the mentally ill as a result of the implementation of community care policies.
Duties to carers
As mentioned above, carers raise important questions in relationship to community care. Szmukler and Bloch [39] have argued for clarification of the ethical basis for working with families of psychotic patients. Much can be achieved in avoiding later dilemmas by an ethically sensitive approach to family engagement at the outset. This essentially involves spelling out through the process of obtaining informed consent from patient and family the basis on which care will proceed. The clinician's position in respecting the relative interests of family members and on confidentiality within the family is made explicit although subject to re-negotiation with the family as treatment proceeds.
Szmukler and Bloch [39] also consider justifications for involving a psychotic patient's family without the patient's consent out of concern for the family's well-being (short of serious, physical danger). Legislation in England as well as Victoria (Australia) suggests that carers' interests are beginning to be formally recognised. This is likely to grow in parallel with the expanding role of informal carers as members of the ‘care team’ in the community. For example, the Victorian Mental Health Act, amended in 1995 (Section 120A), allows the giving of information relating to a patient by a member of staff to a guardian, family member or primary carer of the person to whom the information relates if: (i) the information is reasonably required for the ongoing care of the person, and (ii) the guardian, family member or primary carer will be involved in providing that care.
At least three arguments can be adduced for involving relatives primarily for their own benefit, but against the patient's wishes. The first is to view the family as the ‘unit of treatment’, given that the family has much to contribute to the patient's care and the impact of the illness on them is usually so profound. This would represent a radical departure from conventional notions of medical confidentiality. The second invokes the principle of justice or ‘fairness’ the needs of the family would merit distinctive attention if resources required to meet these needs reside in major part within the treatment team, but not outside it. The third, and probably most compelling argument is to reframe relatives' relationship to the patient as not only familial but also as ‘carer’. As such they should enjoy rights intrinsically attached to all carers, whether relatives or not. These would cover for instance an account of the illness and guidance about how to deal with the patient's problems insofar as they impinge on the carer's life. This might include details about other agencies that offer help. This would bring the family's position into alignment, for example, with that of a hostel support worker. This argument is consistent with the trend in legislation which views ‘carers’ as having special status vis-a-vis the patient.
Whether any of these arguments will be adopted remains to be seen. However, what seems clear is that in this era of community care we expect much from informal carers but this has not been balanced yet by mapping out our duties towards them.
Conclusions
The ethical dilemmas discussed in this paper are not new. However, they present insufficiently different guises to warrant reconsideration in their new context. Diamond and Wikler [26] note that staff working in ACT-like services do not ordinarily view their work as raising significant ethical problems. Their diagnosis of why this is so is in accord with my own observations. They ascribe it to the pragmatism of community mental health team workers:
It is obvious that their work has beneficial effects. Patients who once bounced in and out of hospitals and jails come to live reasonable lives, with apartments and friends and enjoyable activities. It seems reasonable that any clinical intervention that produces such positive effects must be ethically justified. Moreover, these are very difficult patients, and community treatment is very difficult work. Staff are busy trying to prevent mayhem and develop creative treatment plans, develop their clinical skills, give each other needed support, soothe the concerns of the community, and deal with many other social agencies. The constraints of confidentiality, voluntariness, and other moral requirements whose application to the community treatment context is unclear often seem to be issues of bureaucratic nicety, important to rule makers but remote from the real concerns of patients and staff.
This describes the context and sense of beleaguer-ment often encountered in the community. However, for community psychiatric practice to survive and to develop it must rest on an acceptable ethical base. If traditional modes of analysis are wanting, then new ways of tackling the problems must be found which do justice to changing circumstances.
Fear of the mentally ill is omnipresent. Care in the community treads a fine line. If it is not managed successfully ‘there is a danger that progress over recent decades in emphasising the individuality of patients and affirming their rights (and responsibilities) within a therapeutic relationship could be compromised by early resort to unnecessary inpatient supervision and coercive models of care’ [21].