Ethical concerns about methods in psychiatry have been discussed for at least three centuries.1 Methods used for controlling behaviour, such as seclusion and restraint, have been particularly questioned.2 3 A consensus, concerning an optimal theoretical framework for ethical decision-making, has not yet been reached.4 Until the 1960s mental health legislation reflected a paternalistic approach towards involuntary psychiatric treatment, including seclusion, and this was considered to be acceptable in the case of patients who were incompetent to decide.5 In the last 40 years we have witnessed a development towards more patient autonomy in healthcare.6 This has resulted in renewed discussions about the use of seclusion and restraint in psychiatry.7 New legislation, recommendations and professional guidelines to control the use of coercive measures in psychiatry have since then emerged.8 In recent literature many institutions, associations and hospitals have come up with programmes to reduce seclusion and restraint.5 9^–13 The reasons named to substantiate the need for reduction of seclusion and restraint are: respect for autonomy, human dignity and the net negative consequences in the sense of traumatic and harmful experiences during seclusion and restraint. At the same time coercion remained justified in certain circumstances in which patients are a danger to others or to themselves.9 14^–17 Some go further and committed themselves to eliminating the use of seclusion and restraint completely.12 18 The contribution this paper wants to make is to discuss whether we can justify not using coercive measures at all, especially seclusion.

BACKGROUND

Definition and prevalence

There are a number of definitions for seclusion in the international literature. Lendemeijer has formulated the following definition after reviewing the literature on seclusion in psychiatry: “Locking a patient alone in a room for protection of the patient and his environment and in order to control problem behaviour and to enable nursing and treatment”.7 Patients can voluntarily choose to be secluded, however this paper will address involuntarily use of seclusion. The definition above implies that seclusion itself is not a form of treatment but an intervention to make treatment possible. Some people even see the use of seclusion and restraint as treatment failure.11 12 Some articles do, however, describe therapeutic effects of seclusion; patients become calm and gain access to repressed memories during seclusion due to decreased stimulation from sensory overload. Also it has been described that seclusion provides effective means for preventing injury and reducing agitation.19 20 The question remains whether effects should be attributed to the seclusion room itself, to the placement of a patient in a seclusion room, or to all activities around the secluded patient.

Seclusion is still commonly used in psychiatry, however the frequency of use is unclear. There is great variation in the type and frequency of using coercive measures between countries. Published research studies mostly report data from one hospital, which may not be representative of practice in the country of origin.21 Furthermore, studies published use different definitions of seclusion, study different populations and use different endpoints to express the frequency, which makes comparison difficult. Percentages of patients who are involved with seclusion during an admission vary from 1.9–66%, and the average duration also varies greatly between countries and hospitals and study populations.7 20

Patient-related and non-patient-related influences

The associations between seclusion and patient-related factors of age, gender, diagnosis and ethnicity are also unclear; seclusion seems to be associated with younger patients and a diagnosis of schizophrenia, manic-states or personality disorders.5 7 8 12 19 22^–24

Characteristics of the staff, such as their educational level, clinical experience and gender also seem to be factors associated with seclusion.7 9 20 In addition hospital characteristics seem to play a role, for example, architectural aspects and the patient-to-staff ratio. For instance the frequency of seclusion increases when the availability of single rooms decreases.25 Taxis concludes in his article that staff education, patient education, environmental alterations, good communication and administrative and programmatic changes are factors that have contributed to the reduction of seclusion and restraint.9

Another important factor in the decision to seclude is the person who makes that decision. Any given situation of deciding on seclusion will be based on the history of the subjective self of the decision-maker.26 This means that the moral conscience of the caregiver, the culture of the institution in which he works, the legislation, former decisions and the ideas of colleagues, relatives and superiors play a key role in the decision to seclude or not. In order to reduce the use of seclusion and restraint a change of culture is the key.11

We can conclude that the decision to use seclusion not only depends on the behaviour of the patient but also on many complex factors associated with the caregiver and the institution. This causes discussion about the use of seclusion as such. Therefore it is time to discuss the question whether a complete non-seclusion approach can be justified.

ETHICAL ISSUES

At first glance respect for autonomy and the violation of human dignity would seem to support a non-seclusion approach. Likewise, the effects of seclusion—protection from harm and creating a therapeutic climate—would seem to support the use of seclusion.

Autonomy

As mentioned already, the development towards more patient autonomy in healthcare is a trend of the last 40 years. A presumption in favour of individual self-determination has developed, and critique on a paternalistic approach to the healthcare relationship.27 The consequences of this for psychiatry are expressed by P Soloff: “As attitudes toward ‘madness’ become more enlightened, social pressure toward greater freedom and less restrictive care push the psychiatrist to the limits of safety in managing the truly disturbed patient”.28

There have been three kinds of attacks against a central role for autonomy in medical ethics.

Firstly, autonomy is just one of the prima facie principles to take account of; there is no reason why autonomy should have priority over all other moral considerations as, for instance, beneficence. Beauchamp & Childress,29 in their influential book about biomedical ethics, firmly deny that respect for autonomy overrides all other moral considerations. They construct a conception of autonomy that is not excessively individualistic (ie, not neglecting the social nature of individuals and the impact of individual choices and actions on others), and not unduly legalistic (ie, not highlighting legal rights and downplaying social practices).29

Secondly, the concept of autonomy itself is broader; the ideal of autonomy in terms of complete self-sufficiency and independency, free from controlling factors does not exist. It has been argued that autonomy cannot be seen in isolation from other persons and relationships. Instead, it has been stressed that autonomy as a moral capacity can only be developed in relation to others; in other words, autonomy is relational autonomy.30

Thirdly, there are moral values which are more fundamental than autonomy. Campbell,27 following Pellegrino, argues that dependency is a moral principle, and that it is more fundamental than autonomy. He argues that a very independent individual could be lacking in autonomy, in the sense that he cannot consistently exercise any degree of self-direction and cannot function as a member of society. Conversely a dependent individual (eg, a partner in a marriage) can still exercise autonomy. Campbell claims that dependency and autonomy are not polar opposites.27

The consequence of these attacks is that we cannot dismiss seclusion by invoking autonomy only, without paying attention to other moral principles and without paying attention to the relational context of autonomy.

Verkerk agrees with this concept of autonomy and dependency in which the relational context plays an important role and pleas for the possibility of “compassionate interference” in healthcare. “Interventions in care can be shown to be in the interest of patients, that is they can be seen as interventions for attaining autonomy instead of threatening autonomy”.30 So it can be argued that to avoid self-destructive behaviour sometimes coercive measures can be necessary to foster or to regain autonomy. When we see seclusion as an intervention for attaining autonomy instead of threatening autonomy we have lost one reason for eliminating seclusion completely. Obviously, this does not mean that seclusion does not raise many issues related to autonomy.

It seems the natural thing to consider paternalism here as well. Paternalism has been described as a conflict between beneficence and autonomy.29 However, since we will show that there are no scientific grounds for the beneficial effect of seclusion or restraint for the psychiatric patient, there can not be a conflict between violating autonomy and beneficence. Even if we would assume (in the absence of supporting data), that seclusion itself is beneficent using it would amount to a form of weak paternalism because a person’s competency will be compromised in some way by its psychiatric disorder.

Human dignity

Violation of human dignity is often used as an argument against seclusion.18 31 32 With regard to the application of biology and medicine the dignity and identity of all human beings must be protected and respect for their integrity, without discrimination, and other rights and fundamental freedoms must be guaranteed.33

Human dignity is often used as a central concept in our contemporary moral vocabulary, but both the content of the notion and its basis as a moral category are often left unspecified; there is not enough clarity about the meaning of human dignity.34 There are at least two conceptions of dignity: (1) “inherent dignity” of every human person as a universal and inalienable moral quality which can not be earned and can not be taken away; and (2) “individualistic dignity” that is tied to personal goals and social circumstances, which can be either enhanced or diminished depending upon a variety of circumstances. Dignity, in this sense, can be affected by events outside the control of the persons involved, a debilitating disease, for example, might rob one of one’s dignity.35

In the context of seclusion a violation of inherent dignity will mostly be the case. Although this characteristic of humans cannot be taken away from the patient, we can act in a way that is not in accordance with it and as such violate it. In case of “individualistic dignity”, however, things are less clear.

Nordenfelt describes “the dignity of identity” as “the dignity that we attach to ourselves as integrated and autonomous persons, persons with a history and persons with a future with all out relationships to other human beings”.36 This dignity is strongly associated with self-respect and feelings of worth. Nordenfelt argues that other persons can humiliate us and can restrict our autonomy in many ways; this however does not just entail feelings of worthlessness or of humiliation. Intrusion in the private sphere is a violation of a person’s integrity but it also entails a change in the person’s identity and thereby his dignity.

In this sense it can be argued that seclusion does intrude in the private sphere of the patient and entails a change in the patient’s identity and thereby violates the patient dignity. In this way it can be argued that we ought not to use seclusion.

On the other hand a study about patients’ perceptions of the concept of dignity in a psychiatric setting showed that “encountering competent and committed staff”, “being confirmed”, “being looked upon as like anyone else”, “being helped to reduce the shame” and “being understood” are conditions associated with respect for dignity.37 It is unclear if seclusion violates human dignity in this sense, when it is done by competent staff, with confirmation and understanding of the patient and in a way that reduces the shame. It seems that respect for human dignity in this view says nothing about whether we should use seclusion or not but about the way we use it.

Furthermore, it can be argued that patients have already been robbed of their dignity by their disease and the situation that leads to seclusion. Both Pullman and Nordenfelt argue that a disabling disease can rob one of one’s dignity. Seclusion might be necessary to reach a breakthrough in the situation and to regain one’s dignity by regaining self-respect. When we argue in this way, it follows that seclusion remains a problem in one understanding of dignity, but does not necessarily block the use of seclusion in another. That view leads to the conclusion that seclusion must be done with respect for human dignity in a way that is least violating.

Experiences of patients

Another important issue in the moral debate on seclusion are the experiences of patients; do they feel non-autonomous and violated in their dignity? Literature predominantly reports negative seclusion room experiences, but positive experiences have also been reported, especially a considerable time after seclusion.38 Important factors mentioned by patients are the lack of adequate information and education about seclusion.39 Patients report most negative experiences when they do not understand the reason for seclusion and when they do not have the possibilities to discuss it with others.38 Treatment must be individualised treatment which means that it must be discussed with the patient, it must be reviewed and revised regularly, and it must be provided by adequately qualified staff.40

Patients do see beneficial effects of seclusion to themselves. In one study both nursing staff and patients generally believed seclusion to be beneficial to the patient some of the time. However, significant differences were found between nurses’ and patients’ perceptions of the extent of benefit, with only 4% of patients and 60% of nurses believing the procedure to be often beneficial to the secluded patient.41

These studies seem to say something about the way we use seclusion and about the reasons and frequency of seclusion, not about whether we should use it or not (as has been argued for the violation of dignity). In Meehan’s study only 22% of the patients and 2% of the staff were in favour of eliminating seclusion completely.41

Effects of seclusion; promoting wellbeing

The definition of seclusion (locking a patient alone in a room for protection of the patient and his environment, in order to control problem behaviour, and to enable nursing and treatment) implies that it is used with good intentions, namely to protect the patient and his environment and to create a therapeutic situation. This is a view of seclusion based on doing good and avoiding (further) harm. Indications for seclusion (and restraint) are: (1) to prevent imminent harm to the patient or other persons; (2) to prevent serious disruption of the treatment programme or significant damage to the physical environment (to the patient or to the group); (3) to assist in treatment as part of ongoing behaviour therapy; and (4) to decrease the stimulation a patient receives.42 Whether seclusion helps to reach these goals is completely unclear as there is no evidence for its effects. The conclusion of a review by the Cochrane library in the year 2000 is that there are no controlled studies that evaluate the value of seclusion or restraint in those patients with serious mental illness.43 Since then, as far we know, no controlled trials have been published. In the absence of controlled trials, the observations by clinicians remain and they sometimes report positive effects, as mentioned before.41 These observations may be correct in the reported cases but the question remains whether they would be the same in other cases, whether these effects can be attributed to seclusion itself and whether seclusion is the appropriate measure. There is no evidence that seclusion is better than alternatives, such as forced drug treatment in preventing harm and disruption, in assisting in treatment and decreasing overstimulation.

Furthermore negative effects have been published. Fisher concludes in his review that the use of seclusion and restraint can have substantial deleterious physical and psychological effects on both patient and staff.19 The Citizens Commission on Human Rights published Deadly restraints, psychiatry’s “therapeutic” assault, a report that summarised reports of deaths, patient abuse and neglect as a result of seclusion and restraint.44 These reports, however, do not describe controlled trials about the negative effects.

In the absence of any controlled trials in patients with serious mental illness, no recommendation can be made about the effectiveness, benefit or harmfulness of seclusion or restraint. This lack of evidence and reports of serious adverse effects do not mean that seclusion cannot be practical and safe. But when we want to practice evidenced based medicine it is arguable that the use of seclusion should only be continued in the context of randomised trials.43 The design of such a study will not be without problems, but we submit that a study using a cluster randomised design would be feasible and ethically defensible. The lack of evidence of positive effects of seclusion supports a non-seclusion approach. It should stimulate us to find alternative methods of dealing with extreme circumstances of aggressive behaviour in which the safety of the patient or others is threatened.

DISCUSSION

At first we assumed that respect for autonomy and the violation of human dignity would support a complete non-seclusion approach. On basis of the explanation just given we can contest this assumption. Both autonomy and dignity can be used to support seclusion and to reject a non-seclusion approach. In a certain way coercion voids an act of autonomy, on the other hand it can be argued that seclusion in healthcare is sometimes necessary for attaining autonomy, instead of always threatening autonomy.29 30 To a certain extent the same goes for (individualistic) dignity. The ambiguity in these ethical concepts makes clear that an appeal to respect for autonomy and/or human dignity can not be a sufficient reason to reject seclusion. But neither is this ambiguity sufficient reason to close all criticism, and to continue using seclusion.

Because of the deleterious physical and psychological effects described that seclusion can have on both patients and staff, reduction of coercive measurement must be encouraged.19 It is remarkable however, that only 22% of patients who have been secluded support a total elimination of seclusion.41

The assessment of the effectiveness of programmes to prevent seclusion and restraint is hampered by the lack of parallel control groups.8 Variable results have been reported. One study reports a 52% reduction of seclusion, this unfortunately was accompanied by a significant increase of assault on patients.10 Another study, which used data of nine hospitals, reports a total elimination of seclusion and restraint in five of the nine hospitals, without an increase of patient assault.12 Most reduction programmes, however, do not reach a 100% reduction. It seems that there remain circumstances of extreme problem behaviour with danger to the patient or others, which need to be controlled by coercive measures such as seclusion, restraint or forced medication. There is no data to guide clinical decisions as to which combination of seclusion, restraint and/or forced medication would be better in specific patient populations.20 It can be argued, however, that forced drug treatment might even be more harmful than seclusion. Part of this argument is that the medication is given as an intramuscular injection and thus violates the borders of the human body. Also most medication causes a lowering of consciousness which can be seen as violating autonomy. And when the patient is admitted for the first time and has no known diagnosis, the observation of a secluded patient can give the psychiatrist more (diagnostic) information than observing a tranquilised patient. On the other hand one could argue that there is a more direct relation between some states of illness and giving specific medication, for example, in cases of psychosis and anti psychotic drugs. Obviously, this would not be the case when tranquillisers are used as these drugs can be seen as a form of chemical restraint.

An obvious hiatus in the total discussion about coercive measures is the lack of clear definitions and data of its use, and the total lack of trials in which different measures are compared and effects are investigated.

The council of Europe, contrary to the earlier recommendations, states that the use of mechanical restraint is not prohibited: “isolation and mechanical or other means of restraint for prolonged periods should be resorted to only in exceptional cases where there is no other means of remedying the situation”. Furthermore, patients have the right to receive treatment in the least restrictive environment and using the least restrictive means appropriate to the patient’s health and the need to protect the safety of others.40 There is, however, a lack of consensus regarding which method is the least restrictive.20 This absence of data should guide future efficacy and effectiveness studies.

CONCLUSION

In contrast to many reports, we have argued that reasons such as respect for autonomy and the violation of human dignity are not sufficient reasons to eliminate seclusion. Altogether, at least in some interpretations, these norms can be used both to support and to reject a non-seclusion approach.

Reduction programmes do not reach a 100% reduction of coercive measures in psychiatry; there remain circumstances wherein coercive measures seems to be the only option to control problem behaviour.

The total lack of controlled trials about the effects of seclusion in different populations, however, argues against the use of seclusion even in these extreme circumstances in which the safety of the patient or others is threatened. It is very hard to argue that seclusion is the appropriate measure to control these circumstances when there are no data to support this. The use of seclusion should therefore only be continued in the context of randomised trials, in which the effects of seclusion in extreme circumstance most be explored and in which this must be compared with other measures such as forced drug treatment. Designing a controlled trial in which individuals are randomised to intervention may be difficult, but a cluster randomised trial must be possible. We have argued that we can neither rule out nor accept coercive measures relying on autonomy or dignitarian grounds alone. We do need to know whether coercive measures are beneficent to complete the argument, and therefore research is necessary. For as long there is no evidence for positive effects of seclusion the precaution principle of “primum non nocere” should guide our actions.