Whom to treat?

Although the risk factors for suicide are well established, predicting who will actually take their own life is not so easy. For example, patients with a history of self-harm are more than four times as likely to die by suicide than those without such a history. However, people who self-harm outnumber those who take their own life, i.e. although most people who die by suicide have self-harmed, only a subgroup of those who self-harm go on to kill themselves (Reference Appleby, Shaw and KapurAppleby 2006). As noted earlier, the use of psychometric risk measures adds little in these cases.

There is also a complication and expense in enforcing mental health legislation. Many who repeatedly self-harm are not seeking help or care; indeed, some actively refuse it or assault those who are trying to prevent them from self-harming (Reference SarkarSarkar 2011a). Reducing their risk of dying by suicide often requires detention under the Mental Health Act 1983, at a time when there are fewer beds for this purpose. Detention may also imply support under Section 117 of the Act, which includes assertive attempts at engaging patients (e.g. telephone calls, home visits or contact with the patient's family) (National Confidential Inquiry into Suicide and Homicide by People with Mental Illness 2009). Following the Wooltorton case (Box 1), capacitous patients may refuse medical treatment after self-poisoning and may also refuse to engage with psychological therapies recommended by clinicians. It seems clinically futile to detain someone under the Mental Health Act to prevent them from harming themselves if they are simultaneously refusing to be helped. However, the legal position is that if there is ‘real and immediate risk to life’ about which the authorities know or ought to know, they have a duty to do ‘all that reasonably could [be] expected’; failing to do so will lead to (legal) consequences that can be devastating, not just for the patient but also for the staff and the service (Savage v South Essex Partnership NHS Foundation Trust 2008).

There is an interesting overlap here in terms of the ethical duties of doctors. Treating every individual who self-harms as though they are at potential risk of suicide will entail a significant increase in clinical resources. It may also result in so many false positives that true positives may be missed. In terms of best use of resources, it might be better to focus on high-risk individuals, and not only on anyone who cuts themselves (Reference Runeson, Tidemalm and DahlinRuneson 2010). However, such an approach means that some individuals will die by suicide because no algorithm is 100% accurate, and cases deemed to be low-risk will turn out to be occult ‘high-risk’.

What to treat?

Might it make more sense clinically to focus on high-risk factors such as diagnosis, not behaviours? This is true to an extent. The two most common single diagnoses among those who die by suicide are affective disorders (46%) and schizophrenia (19%); next, a large proportion (29%) have diagnoses of personality disorder, drug and alcohol misuse, anxiety or adjustment disorders (Reference Appleby, Shaw and KapurAppleby 2006). So even if clinicians targeted the two categories of serious mental illnesses that can be easily treated with pharmacotherapy, there would still be a sizeable group who would need more complex care packages; and (as already mentioned) might refuse to engage. Similarly, the sheer numbers would be a challenge for current resources. For example, if every person with a drug or alcohol problem is assessed for suicide risk, and offered treatment, this would mean offering therapy to thousands more people.

When to treat?

Does knowledge of high-risk periods help the clinician? A study by Reference Gunnell, Hawton and HoGunnell et al (2008) found that the risk of suicide in the month after psychiatric in-patient care was around 100 times greater than that for general population. More than 6.5% of all patients discharged from psychiatric in-patient care were readmitted for an episode of self-harm within 12 months, with a third of these episodes occurring in the 4 weeks after discharge, thereby sharing many of the features of suicide after discharge. The risk of self-harm in this period was higher in females, younger people, those with diagnoses of depression, personality disorder and substance misuse, and those with shorter lengths of stay. However, these data do not help clinicians to easily determine who should be allowed to go home with advice, who should be persuaded to stay voluntarily in hospital and who should be detained against their will. They also do not often tell the clinician who should receive the full force of psychiatric input in terms of monitoring and support.

In summary, assessing and acting on the risk of suicide presents many of the same clinical challenges as managing the risk of violence (Reference Sarkar, Agrawal, Bolton and GaindSarkar 2011b). There are a number of well-established actuarial risk factors (age, gender, diagnosis and past behaviour), but no accurate algorithm for establishing which of the people with these risk factors are actually at high risk of completing a suicidal act. Further, it is in the nature of suicide (as it is in fatal violence to others) that some deaths occur impulsively, in the absence of any known risk factors, or unintentionally. Hindsight bias may make such cases appear ‘obvious’ later, but this is a cognitive illusion (Reference KahnemanKahneman 2011).

Who decides?

Clinicians are sometimes required to make judgements on life and death issues on behalf of their patients. The risks of allowing someone other than the person whose life it is to determine which lives are ‘worth living’ and ‘worth saving’ is a highly controversial matter (Reference GloverGlover 1990). It may be difficult to separate one's own views from those of society about what constitutes a life worth living or a life worth saving. This moral decision is what often determines to which individuals a society will ‘cause death to save other lives’ (Reference GloverGlover 1990), which individuals are ‘allowed’ to take their own lives, which ones are forced to endure lives of tremendous hardship and pain, and which individuals are taken care of by society irrespective of how intensely they wish to die. The eugenics programmes of several nations in the early 20th century and the excesses of Nazi Germany are not too distant reminders of the risks implicit in such a process.

Mental health law is a social response to the profound ethical dilemmas posed by adults who are a risk to themselves or others as a result of mental disorder or distress. The ethical tension is between respect for autonomy and liberty to be left alone and the public duty of society to protect those who are vulnerable. All mental health legislation is a trade-off between the principles of respect for autonomy and liberty on one side and the principle of respect for welfare on the other (Reference Adshead and SarkarAdshead 2005), and on occasion it leads to medical paternalism.

Medical paternalism

For a clinician to decide whether it is ‘appropriate’ to prevent someone from taking their own life is an example of medical paternalism. The Mental Health Act 1983, which governs England and Wales, allows for individuals with a recognised (in international classificatory systems) mental disorder to be detained for appropriate treatment if their illness is of a sufficient nature or degree to require hospital treatment. This ultimately boils down to an assessment as to whether the person, as a result of the mental disorder, poses a significant danger to themselves, their own health or to others. Amendments to the Act in 2007 require a demonstration by the detaining authority that ‘appropriate medical treatment’ exists to treat the condition that the potentially detained patient has. In terms of a suicidal patient, this requires the clinician to consider whether it is appropriate to detain the person simply to prevent suicide or whether appropriate treatment can ameliorate the condition to the extent that the person no longer remains at risk of suicide at the end of treatment. This determination allows that clinician to override the person's wishes, even if they retain mental capacity to make a reasoned decision.

There has been particular concern about the mental legislation of England and Wales because it is strongly paternalistic in that it allows for the detention and involuntary treatment of capacitous, and therefore autonomous, individuals because clinicians consider it is ‘appropriate’ to do so. The consequence of this is a coercive anomaly in that, if one has full mental capacity, one can refuse any medical treatment; but it is not possible to refuse admission to a psychiatric hospital or even physical/pharmacological treatment. An attempt to introduce a capacity-based mental health act (whereby one would have to show that a person lacked capacity before involuntary detention and/or treatment) (Reference RichardsonRichardson 1999) failed under the Labour government of 1999 and the current coalition government has done nothing to change the status quo. This suggests a political will to ensure that those considered to be mentally disordered are treated differently from the rest of the society when it comes to autonomy; a sad commentary on Liberal Democrat principles.

Autonomy and best interests

In psychiatry, respect for autonomy may mean that a clinician will allow a patient to make decisions free of coercion, even if this may appear to go against the patient's best interests – the basis of assessing capacity in statute law. Ethical principles for assessing capacity are now enshrined in law, in the Mental Capacity Act 2005 for England and Wales. They are as follows:

1. • people are assumed to have capacity to make their own decisions unless otherwise proven

2. • everything should be done to help a person arrive at a decision before declaring them incapable of making it

3. • whatever is done for a person lacking capacity should be done in their best interests

4. • the decision maker should choose the least restrictive intervention.

For instance, where someone is found in a state of semi-consciousness following a large overdose and their life is at stake, the Mental Capacity Act would require that immediate action be taken in the person's best interests to save life. It would not be practical or reasonable to go on an exhaustive fact-finding mission before initiating treatment. However, beyond the emergency phase, while a person still lacks capacity, the Act requires a much more concerted assessment of the person's capacity, the views of their carers or appointed attorneys, formal assessment of their best interests and consideration of any advance decisions. Within this Act, an advance decision is specifically a decision to refuse treatment made in advance and in writing. If it was prepared at a time when the patient had capacity, an advance decision is legally binding unless there are legitimate questions as to its validity or applicability (see the Wooltorton case outlined in Box 1). Further guidance is contained in the Code of Practice accompanying the Mental Capacity Act (Department for Constitutional Affairs 2007).

Fluctuating capacity and the tripartite model

What is a clinician to do when a patient's capacity changes from day to day and week to week? Severe disturbances of affect and arousal regulation cause rapid oscillations in a person's perceptions about themselves, their world and their anticipated future (Reference Sarkar and AdsheadSarkar 2006); and equally oscillating decisions about accepting or refusing life-saving care and treatment (Reference SarkarSarkar 2008). These are states of mind in which there is fluctuating capacity, a phenomenon that has been largely ignored in capacity judgments to date (Reference Sarkar and BeeleySarkar 2011c). Despite a slew of legislation that can bewilder clinicians, fluctuating capacity raises the problem that theoretical principles do not necessarily address real, complex clinical issues (Reference Mogg and BartlettMogg 2005). The bar in assessing capacity must be set high when clinicians make best interests decisions when someone's life is at stake, as in the case of suicidal patients. This approach was recognised by the Court of Appeal in Re MB (An Adult: Medical Treatment) [1997] when it was stated that ‘the graver the consequences of the decision, the commensurately greater the level of competence required to make that decision’.

Case law on capacity suggests that temporary factors such as confusion, shock, fatigue, pain, drugs or panic can completely erode capacity (Royal College of Psychiatrists 2004). Clinicians will add to this list other factors, such as dissociation and problems of self-identity, rapidly fluctuating mood and behaviours, and concomitant risk of harm to others (Reference Gallagher and SheldonGallagher 2010; Reference SarkarSarkar 2011a). This highlights the real clinical challenge of working out a person's real intent and choices and decisions when they are in great distress. I argue that it is imperative to consider a third element in capacity decisions (Reference SarkarSarkar 2008). This is the determination of how ‘stable’ or ‘settled’ a decision judged to be capacitous actually is. If it is affected by or a product of underlying mental disorder, it may change with the illness over time. Therefore, a written decision cannot simply be accepted as the final and enduring decision of the individual. English capacity legislation requires that the person's past and present wishes and feelings – expressed verbally, in writing or through behaviour or habits – beliefs and values (e.g. religious, cultural, moral or political) that would be likely to influence the decision in question are taken into consideration when best interests judgements are made by others.

It is critical that clinicians take a longitudinal view of the capacity of all patients. In other words, just because a person appears to be capacitous at the time of assessment (a cross-sectional view) does not mean that they will be capacitous the next week, the next day or even the next hour. A useful strategy is to go beyond the current approach to capacity assessment based on the Mental Capacity Act, which is often limited to two specific elements: that the decision be informed and that it be clearly communicated to others.

Reference Eastman and HopeEastman & Hope (1988) state that capacity decisions are not binary (present or absent) but graded, i.e. there are degrees of capacity. Different decisions about treatment that a patient is required to make would demand different levels of capacity. They suggest combining the complexity of the decision and the significance of outcome (based on the decision made) in what they call ‘the balance model’. However, this approach has been criticised as being potentially too paternalistic because it can have a tendency to focus too much on the desirability of outcome (Reference Mogg and BartlettMogg 2005).

I propose an ‘enhanced’ approach to the balance model of capacity judgements if there are relatively rapid changes in the person's mental state, leading to rapid changes in capacity and risks. The ‘enhanced’ tripartite model requires three elements to be assessed: the rationale or logic behind the decision made; the conviction with which the decision is held and is aligned with the patient's life narrative, beliefs and values; and the clarity with which this decision is communicated to others (Table 1). It is suggested that these three elements, when assessed together, are likely to address the confounding effect that fluctuating capacity has on issues of consent (Reference Mogg and BartlettMogg 2005), which can often lead to reluctance of clinicians to treat in difficult cases (Reference McCall SmithMcCall Smith 2001). By trying to ensure that the decision is stable, enduring and consistent with the patient's frameworks of value and meaning, this approach provides the clinician with a strategy to limit paternalism as much as possible. The clinician's tasks therefore include:

1. • judging whether the decision (e.g. to self-harm, discharge from voluntary admission, go on leave from hospital) is informed and balanced

   TABLE 1 A tripartite model of capacity judgement in relation to a wish to die

   

2. • judging whether it is communicated clearly and unambiguously

3. • judging the conviction with which it has been made by assessing its stability and endurance over time: the more important the decision and the more final its consequences, the longer the time frame should be over which its stability is assessed.