The Least Restrictive Argument

The substantial ethical issue regarding CTOs, at least from a principalist perspective, is that they constitute a restriction on patient autonomy. That CTOs restrict autonomy is more or less true by definition, or at least this will be true in any case where a CTO is appropriately applied. The institution of a CTO means that the patient must accept psychiatric treatment as designated by the treating physician. If the patient were antecedently inclined to accept this treatment, the CTO would not be required, and therefore no such order would be warranted. This means the only times a CTO is appropriate are exactly when the patient would not accept the treatment of their own volition. This means in turn that any warranted CTO entails that the patient cannot act as they would otherwise choose to act.

Given that CTOs restrict autonomy, a common claim to justify this restriction is that the constraints on a patient's autonomy are less substantial under a CTO than they would be under committal as an inpatient: that restriction in the community is less than that in hospital.^{Reference Hiday and Goodman23} Indeed this has been the default position in law^{Reference Keilitz, Conn and Giampetro24} and the basis of the Western philosophy of Locke, Mill, and others^{Reference Mill and Robson25} focus on maximizing personal freedoms. There is certainly prima facie justification here; it is true that autonomy as an inpatient appears more restricted than autonomy as a patient in the community. However, there are a number of potential concerns with this line of reasoning.

The False Dichotomy in the “Least Restrictive” Argument

From the outset, this “least restrictive” argument, argued first at law almost half a century ago,^{Reference Parker26} runs the risk of employing the fallacy of false dichotomy: it relies on the assumption that were the patient in question not on CTO, they would necessarily be a hospital inpatient, when there is no clinical reason to think this is the case.

Interestingly this problem was discussed in detail in Lake v. Cameron in 1966. In this case a 60 year old woman, with an organic brain syndrome was committed to a mental hospital as she was considered a danger to herself by wandering. Her condition was essentially untreatable and no family could care for her (nor did she have the means to pay for private care). Nonetheless the court found her indefinite detention unlawful if alternatives could be found. CTOs did not exist in Washington in 1966, however it is not hard to imagine this as being presented as a least restrictive alternative. Nonetheless there is no evidence that a CTO in a case like this, or indeed any, would increase her freedom by enabling discharge. The argument is that community based patients without CTOs will not engage with treatment, become unwell and therefore be admitted (or readmitted) anyway, the “revolving door argument.”^{Reference Weiden and Glazer27} CTOs, by implication, prevent this revolving door and the increased restriction being an inpatient involves by making treatment happen in the community. However, the best evidence is that CTOs do not have an impact on readmission,²⁸ this argument is not supported by the randomized controlled evidence.

The only other defensible reason for a clinician to hospitalize a patient who is well enough to live in the community would be the potential risk of a future adverse event. However, very high hospital detention rates would be required to prevent future negative adverse events in this way, as our capacity to predict the risk of violence is poor^{Reference Szmukler29} and these rates would be unacceptable on ethical and pragmatic grounds.^{Reference Newton-Howes30} This means detention as a hospital inpatient is not the most relevant alternative for patients on CTOs. Rather, the most relevant option is that they are managed as out-patients without a CTO, engaging in a discussion with their treating team as to the best pathway forward for them. This is both recovery oriented and in turn nullifies the dichotomy that underlies the “least restrictive” argument. At the very least, although this dichotomy may be true of some patients on CTOs (i.e. were they not on a CTO, they would have to be an inpatient), it will not be true of many. This being the case, the “least restrictive” argument cannot be an argument in favor of the restriction of autonomy under CTOs in general.

The Ethical Opacity of CTOs

If we put this false dichotomy aside and (falsely) assume that CTOs are less restrictive than the relevant alternative(s), although this might be a reason to employ CTOs it would in no way mitigate the fact that patient autonomy under a CTO is still restricted. This has been brought into sharp relief in the light of the CRPD and international commentary to CTOs by the CRPD Committee, who go as far as recommending CTOs be abolished in some jurisdictions because of this opacity (for example the committee on the Rights of Persons with Disabilities. Concluding observations on the initial report of Australia, 10th session. UN Doc CRPD/C/AUS/CO/1, 4 October 2013). That is, even if the use of CTOs were the best option from a list of bad possibilities, this would not mean we shouldn't still be highly cognizant of the ethical and legal risks involved.

Indeed, there may be reasons to think that the ethical risks posed by restriction of autonomy are more problematic in the setting of CTOs as opposed to hospital-based care. This is for at least two reasons. First, when autonomy is restricted in the setting of enforced in-patient care, the need for such measures is usually more clear-cut. The comparative severity of illness and/or risk will mean it is apparent that autonomy needs, on balance, to be reduced. When the patient is well enough to live in the community (as, again, they will be in any case of a CTO), the need for autonomy restriction will be less apparent, and therefore more difficult to defend. These, then, are precisely the cases when the most robust arguments for such restrictions should be made available. Second, the restriction of autonomy in the community setting is particularly tricky with respect to ethical concerns, as the restrictions are less transparent. Again, restrictions on autonomy are prominent in the setting of inpatient management, and so suitable checks and balances are clearly required and maintained.^{Reference Newton-Howes31} However, in the case of CTOs, there are opportunities for clinicians to feel satisfied that their actions are ethically preferable (such as through the argument of least restriction), and therefore do not require equivalent scrutiny. And it is not at all clear that this is the case.

The Appearance of Autonomy versus Autonomy

In any situation where an agent is able to make certain choices, there will be the appearance of autonomy. However, in certain circumstances, this appearance may outstrip the reality. In a series of discussions,^{Reference List and Pettit32} Philip Pettit and collaborators show that freedom comes in a variety of types and extent. It is true that someone on a CTO has relative freedom to make decisions regarding domains outside the clinician's jurisdiction. But the options available to this person are still limited, and even this restricted amount of freedom is quite precarious, as it is curated by another individual, and can therefore be taken away by that individual. The consequence of refusal to act as directed by the treating clinician is a possible summons to hospital, usually with police assistance, followed by enforced treatment. So in at least some strong sense, the administration of physician-dictated treatment to a patient on a CTO is just as inevitable as that of an inpatient. A freedom that can be taken away in this manner is not the type of freedom that one would usually consider valuable. This therefore may give an appearance of autonomy, while the actual autonomy of the patient is in fact more restricted.

The Failure of the Least Restrictive Argument

The above points mean that, although we might be well aware that a patient's autonomy is restricted under a CTO, we can be fooled into thinking this restriction is less ethically and legally substantial than it actually is. This is for two reasons: we are reassured by the fact that a CTO is less restrictive than being hospitalized, and we perceive that the patient retains some real freedoms. Both of these lines of thought, although perfectly understandable, are misleading and may leave clinicians and policy-makers with a false sense of reassurance. This is reflected in the developments in mental health law and the challenges the CRPD presents.^{Reference Gooding33} There can be little doubt that CTOs are necessarily coercive and restrict the autonomy of the patient. They may appear to be the best option out of the two available, but in fact coerced inpatient treatment is not the only — or even the most likely — alternative, and so this reasoning is based on a false dichotomy. CTOs actually may be more ethically challenging than forced inpatient treatment, because the issues are more ethically opaque. CTOs coerce the patient through the presence of threat, while generating an appearance of increased freedom which outstrips the actuality. So even though CTOs do in fact restrict autonomy to a lesser extent than inpatient management, using a “least restrictive argument” is insufficient to ethically justify CTOs and is increasingly recognized as inappropriate by international legal convention.

The Best Interests Argument

The other primary justification for the use of CTOs is the “best interests argument”:^{Reference Burns and Firn34} that on balance CTOs improve a patient's wellbeing and are therefore in that patient's best interests.^{Reference Brophy and Ring35} This argument is a proxy for beneficence. This argument is not commonly made explicit in clinical practice, although it is a standard rationale behind many proactive interventions in medicine, including medico-legal interventions such as CTOs, and reflects the desire of the physician to “do good.”

The Patient or Clinician's Best Interests?

Once again, there are a number of problems with this argument. As discussed above, whenever a CTO is warranted, the treatment will certainly not be seen as “in the patient's best interests” by the patient themselves. This does not necessarily mean the use of CTOs is wrong, ipso facto. However, the fact that a particular treatment is deemed by the clinician to be “in the patient's best interests” is not in itself a sufficient argument for the restriction of autonomy in the face of patient disagreement: by itself this would be simple paternalism, which is no longer seen as legitimate grounds^{Reference Breeze36} and is diametrically opposed to the tenets of the CRPD. It is also worth noting this line of reasoning would also be questioned by the recovery model of care, now the guiding model of care in the USA, New Zealand, Australia, the UK and Canada.^{Reference Le Boutillier, Leamy, Bird, Davidson, Williams and Slade37} Rather, it is necessary to show that this patient's best interests are being considered, and are the primary motivation for intervention. In order to confirm this, two things need to be established: First, that the patient is incapable of making their own decisions regarding their own well-being, and second, that the treating team has sufficient evidence to take it upon themselves to make that decision on the patient's behalf. In respect of the former this is a high bar to set, and if at all possible the patient should be supported in making their own decisions. The first general comment on the CRPD would insist that in all circumstances the patient be supported to make all decisions,^{Reference Arstein-Kerslake and Flynn38} a stance that makes CTOs in any sense indefensible. This view has, however, been seen as radical and challenged.^{Reference Dawson39} Nonetheless it makes clear that significant efforts must be made to support the patient in their decision making prior to abandoning this effort. Second, there needs to be little, if any, doubt that the intervention is the best one to choose. This bar is also high as it requires evidence that is scientific (i.e. that the intervention works) and idiosyncratic (that the patient would choose the effective intervention if they could wholly make such a choice).

Making Capacity Meaningful

In order to meet the first requirement, it must be shown that the patient lacks capacity and cannot be supported to gain this in order to make their own choice. Although CTOs uniformly require the presence of problems with mental functioning and usually require an assessment of risk, many jurisdictions do not require am assessment of capacity.⁴⁰ This means a patient may be detained, suffer a loss of privacy, and be required to take treatment such as injectable medication (with concomitant risk of side effects), despite potentially having an intact capacity to refuse intervention. This is a breach of the principle of freedom and control over one's own person, and is extremely difficult to justify⁴¹ from a beneficence perspective. The only other medical setting where enforced treatment can be instituted despite capacity to make an informed choice is when patients are detained in hospital to prevent an epidemic of an infectious disease. This is a rare occurrence, the detention is in hospital, and it is very obviously in the community's interests. Further the intervention is applied equally, based on diagnosis (i.e. infection), not idiosyncratically, based on a doctor's likely faulty risk assessment.^{Reference Ryan, Nielssen, Paton and Large42} These arguments cannot be made to support the use of CTOs.

Ensuring the Best Interest Even with Restricted Capacity

Regarding the second requirement, in order to act without patient consent it is necessary to have evidence that a treatment maximizes the probability of benefit. At a minimum, the data must suggest the particular intervention is likely to result in a better outcome than the relevant alternatives, including no intervention. However, the best evidence to date shows minimal benefit from CTOs, and concludes specifically that a well-being argument cannot be sustained. There have been three randomized controlled trials of community treatment orders, two in the US^{Reference Swanson, Van Dorn, Monahan and Swartz43} and one in the UK.⁴⁴ None of these trials showed a benefit from CTOs in their primary outcomes. Some secondary outcomes showed benefit, although these have been questioned and are true only for a small sub-population of those involved, again preventing a generalization of this argument. Cohort trials suggest benefits that are not found in the RCT evidence, although these appear, or at least may be, a result of increased psychiatric care. This suggests CTOs may act in some jurisdictions as a gate keeper for best (or at least better) care. This bizarre situation would then require a patient to be held on a CTO to access such “better care,” a situation hard to justify for either the individual patient in question or the larger group of non-detained patients who are receiving lesser care than their coerced colleagues. As such it becomes difficult for a doctor to justify the use of CTOs on the evidence available, as this evidence identifies little benefit and potentially creates unequal service provision.

The Failure of the Best Interests' Argument

To recap: in order to be ethically justified in overriding a patient's autonomy for reasons of enhancing their best interests, one must be able to demonstrate that the patient is incapable of knowing what is in their own best interests, and also show that the treating team does know what is in their best interests, from the literature and understanding of the patient's wishes and preferences. It does not appear that the former requirement is always met and the CRPD questions if this will ever be the case. It is also debatable whether the second requirement is ever met given the status of the current evidence base. Together, these mean that the institution of a CTO is difficult to justify on the basis of “best interest,” failing to adequately assess and consider capacity in many jurisdictions and further failing to have evidence of improvements in care. This puts both the patient and physician at risk on ethical grounds and opens the door to challenge under the CRPD in those jurisdictions where individual appeal is agreed to.

Being Clear about the Failure of Best Interests

First, the concern that the patient will not believe the CTO to be in their best interests should be explicitly addressed in each case prior to implementation. Of course, the interests of the patient would be best served if the clinician and patient reached consensus with regard to treatment, invalidating the need for a CTO. Efforts in shared and supported decision making help,⁴⁶ however, this is not always possible. We suggest that the patient's perspective regarding the treatment regime is assessed and documented in such circumstances. This would mean the patient's views regarding their own best interests and well-being are considered and incorporated into clinical thinking. This requirement is in line with a values-based medicine paradigm^{Reference Fulford47} and the recovery model of care. Further, working toward management that is supported by patient, family, and the clinician should be the explicit aim of ongoing treatment, enabling the CTO to be rescinded as soon as practicable.

Being Clear about the Limits of Capacity

Second, when a CTO is indicated and treatment occurs against the patient's wishes, the threat of ethical misadventure is particularly high if assessment of capacity is not required. Enforced management of a person who is competent to refuse that management is a gross breach of autonomy. The solution to such a risk is clear: the assessment of patient capacity should be required for all CTO applications. If a capacity assessment is required in law, and CTO application only permitted where no or highly limited capacity exists, there would be less concern regarding breaching the autonomy of the patient, assuming efforts have been made to provide a supported decision. A best interest's intervention may be considered appropriate if the patient cannot make informed choices themselves despite support, so a CTO will simply be one of many cases where medical treatment occurs in the absence of capacity. Such a clause also allows for a CTO to be applied flexibly by the courts. For example, if no capacity to consent to pharmacological treatment exists, but capacity to refuse entry into home does, the court could apply a CTO for the former, but not the latter. This would both individualize the use of CTOs and make the limits of their use explicit.

Instituting a requirement to ascertain non-competence may mean that some CTOs do not occur that would otherwise have been enacted; namely those cases where the patient is competent to make the decision him or herself. However, this would presumably be a good outcome, at least from an ethical perspective. In those very cases an agent would have been forced to take psychiatric medication and management against their (possibly competently formed) wishes to the contrary.

Recognizing the Temporal Limits of the Evidence of Benefit

Third, there is the problem of a lack of evidence for the efficacy of CTOs, if not management of specific mental disorder. The remedy in this instance would be to require a detaining doctor to identify the diagnosis that requires treatment, in order to ensure a treatment has an evidence base supporting clinical effectiveness. In cases where diagnostic uncertainty exists, such as the new patient to a community team, identifying the differential diagnosis and their evidence-based management would fulfil the same purpose. In essence this would make the CTO a conduit to evidence based treatment with proven effectiveness, as opposed to a treatment in and of itself.

Distributing Power and Working Towards Parity of Esteem

Fourth, the possible situation where the responsibility for treatment and detention decisions are undertaken by a single clinician can be avoided by requiring that different individuals must be involved in these tasks. The appearance of freedom in CTOs should not obscure the risks of such a concentration of power. This apparent freedom exits as the patient is in the community and is often based on medical impression, as opposed to formal assessment of capacity. Those in mental health should be looking to apply enforced treatments under conditions similar to their physical health counterparts, bringing the foundations of mental health and physical health ethically closer together. The legal remedy to this problem is a “fusion act,”^{Reference Dawson and Szmukler48} bringing together mental health and incapacity legislation to apply to all medical conditions. As described, this now exists in one jurisdiction: Northern Ireland. This appears to be far closer to the principles of the CRPD, albeit still based on a failure of mental capacity as opposed to a supported legal capacity paradigm.