Beneficence and non-maleficence: a diffuse risk, and barely effective interventions translate into adverse effects

Risk factors for suicide are well established but algorithms to predict suicide in individual patients are blunt. Many who die by suicide would be categorised as having a low risk, and the overwhelming majority of those who would be categorised as having a high risk will not die by suicide.^{Reference Large and Ryan10,Reference Large, Kaneson, Myles, Myles, Gunaratne and Ryan11} Thus, even if directed at comparatively high-risk patients, preventive efforts need to target many, most of whom would not die by suicide. Still, even if effective, such efforts would not be sufficient to achieve even near zero unless comparative lower-risk individuals were also included.

Some efforts are benign or beneficial to the many people who are targeted by the measure regardless of their effect on suicide. Increased access to high-quality mental health services and to primary care, strengthening of social safety nets and efforts to destigmatise sexual orientations and mental health conditions are typically cases in point, and raise little question.⁷ When measures are intrusive, restrictive or risk harm, trade-offs are more complicated. This applies to measures that target entire populations, subpopulations or individuals.

At a population level, one of the best-established methods for suicide prevention is reducing access to lethal means for suicide. The positive effects of restricting access to some means through changes of composition of household gas, use of blister packs for pain relief and reduced access to hazardous pesticides typically outweigh minor burdens on end users.^{Reference Duberstein, Marnin and Nock12,Reference Zalsman, Hawton and Wasserman13} Restricting other means for suicide is more complicated and may have unintended effects.^{Reference Florentine and Crane14} Drugs used in suicide by overdose are often the same drugs used for treatment of depression, anxiety and pain conditions. If we aim at removing causes of suicide, effective treatments need to be accessible for as many affected as possible. Moreover, not all means can be removed, at least short of population-wide restrictions and monitoring.

Likewise, some more targeted measures raise few or no ethical questions. Consider improved treatment for psychiatric illnesses and for pain conditions, or effective follow-up after psychiatric in-patient care. These measures are benign, and desirable for reasons beyond suicide prevention. Other measures are less benign. Take involuntary hospital admission: a brief period of hospital admission to save a person's life may seem like a uniformly justifiable trade-off. However, if the vast majority of apparent high-risk patients do not die by suicide and adverse effects are significant, the trade-off is more problematic.^{Reference Borecky, Thomsen and Dubov15} Unfortunately, evidence of effective targeted efforts is scarce.^{Reference Fox, Huang and Guzmán16,Reference Huang, Harris, Funsch, Fox and Ribeiro17} Even if we could effectively single out people with a high risk of suicide, there is no simple suicide-reducing effort we could easily offer to all.

Finally, anti-suicide measures may expose an entire population to universal primary suicide prevention. As a thought experiment, take the unconventional idea of adding trace lithium to a population's drinking water for suicide reduction.^{Reference Vita, De Peri and Sacchetti18} Even if such a measure were highly effective, for example by halving the suicide rate, in a standard population preventing a single death would require exposing about 17 000 people to lithium.^{Reference Ng, Sjöstrand and Eyal19} Even if adverse effects from the exposure were rare and mild, cumulative harm might be substantial.

Personal autonomy: the moral permissibility of suicidal acts

Absent reference to religious foundations, it is difficult to ground a blanket objection to all suicide. Specifically, it is far from clear that ending one's life can never be a person's autonomous choice, or in their own considered interests. Narratives of patients contemplating physician-assisted suicide (PAS) often involve intense prior deliberation, terminal illness and incurable suffering and dependence.^{Reference Chapple, Ziebland, McPherson and Herxheimer20}

The most important secular argument against all suicide comes from 18th-century philosopher Immanuel Kant. For Kant, suicide for the sake of relieving suffering trades our dignified personhood for mere pain relief irrationally, and immorally, even when it disadvantages no other person.^{Reference Johnson and Cureton21} However, from a Kantian perspective, suicide can also serve personal will or prevent indignities.^{Reference Kerstein22} Importantly, contemporary bioethics determines whether decisions are autonomous and dignifying procedurally, rather than by their content.^{Reference O'Neill23} Admittedly, many suicides are seemingly impulsive, non-deliberated acts, associated with mental disorders, substance misuse and poor problem-solving abilities, all factors that can undermine capacity for autonomous decision-making.^{Reference Hegedűs, Gál, Szkaliczki, Andó, Janka and Álmos24,Reference Jollant, Bellivier and Leboyer25} However, the factors leading to a person's suicide are too variegated for all to lack personal autonomy and rationality. Decision-making capacity may be retained even in severe psychiatric disorders.^{Reference Curley, Watson and Kelly26} The global suicide rate is higher in later ages than in adolescence and correlated to physical ailments and functional impairments,^{Reference Fazel and Runeson1} and even when a death by suicide appears impulsive to external observers, it may well have been pre-meditated and considered for a long period of time.^{Reference Anestis, Soberay, Gutierrez, Hernández and Joiner27}

Whether or not people have a reason or a moral right to die by suicide, it might be argued that society has a duty to prevent suicides. However, liberal societies typically endorse the view that authorities should not interfere with people's capacitated decisions as long as these do not infringe on the rights of others. Contemporary bioethics assigns competent adults particularly strong rights to decide autonomously on matters central to their own life and death.

Although in rare instances contemporary bioethics may allow either the person's good or the community's good to supersede the person's will, such exemptions are unlikely to rule out all suicide.

To say all that is not to endorse the other philosophical extreme – a certain libertarian position according to which suicide is nearly always an individual's right, such that coercive suicide prevention is nearly always a wrongful interference with individual freedom.^{Reference Szasz28} Moreover, suicide is often a serious harm to others,^{Reference Cerel, Jordan and Duberstein29} and that should also figure in the balance.

These considerations also play out in the international arena. In Sweden, the zero goal has been repeatedly invoked as an argument for retaining a ban on PAS and euthanasia.^{Reference Holm30} Therefore, whether or not one agrees that PAS is a permissible practice, international health organisations should arguably not commit to rejecting them altogether. Perhaps inadvertently, international efforts to prevent any suicide risk doing precisely that.

Fighting injustice: attacking the symptom instead of the malaise

Suicides are disproportionally common among those who are less privileged: people with mental illness, people who are socially marginalised and people with limited economic options.^{Reference Spencer-Thomas31} Much of the suffering that drives suicides comes from unjust national and global socioeconomic arrangements, stigmas and deficient medical and social services. We should fight these injustices much harder than we currently do, both for all the independent reasons to fight injustice and because they instigate suicide. However, the duty to fight injustice does not equate to a duty to fight directly all of its results. If workers who are abused and exploited die by suicide because of, for example, terrible work conditions, the solution is not simply to ban suicide or restrict their access to means for suicide. The underlying causes are what must be altered. An obligation to reduce a cause of something (such as bad work conditions) is not the same thing as an obligation to reduce that thing (for example workers’ suicide rates) irrespective of the means to achieve it.

The Swedish vision zero states that ‘promoting good life opportunities for less privileged groups’ is one of its main strategies.^{Reference Wasserman, Tadić, Bec, Edvardsson Björnberg, Belin, Hansson and Tingvall3} Yet a goal stated in terms of reducing suicide rates risks diverting from promoting just opportunities – the good experiences that may drive such reductions. We should fight the ‘fire’ of social injustice instead of the ‘smoke’ of suicide. Indeed, blowing away the ‘smoke’ may fan the ‘fire’: unjust stigma against an ethnic minority or LGBTQA (lesbian, gay, bisexual, transgender, queer/questioning, intersex, asexual) youth increases suicide, but targeted, exclusive monitoring in these communities for signs of intent to suicide could be perceived as discriminatory and distrustful, exacerbating stigma and injustice.

Transparent communication: a goal known to be unrealistic may undermine trust and exacerbate frustration and stigma

Suppose that public health authorities announced a goal of ‘zero myocardial infarctions’. That clearly unrealistic goal would only make authorities look detached. The inevitable failure to meet this overambitious goal could sow frustration, distrust and finger-pointing. It would obfuscate realistic goals for reduction of the number of myocardial infarctions, and therefore their proper ranking compared with other realistic public health goals, such as reducing cancer. Similar dangers lie in the clearly unrealistic zero-suicide goal. While there are methods to reduce suicide in a population, there is no credible strategy to reach zero, making failure unavoidable. Ethically, if we cannot reach an outcome, we cannot have a duty to reach it. By implying that every suicide is preventable, a zero goal risks demoralisation,^{Reference Mokkenstorm, Kerkhof, Smit and Beekman32} and exacerbating feelings of guilt among staff and among relatives bereaved by suicide.

Advocates of zero-suicide goals may respond that they should be seen as ‘aspirational’ and are not literally meant to be achieved, or that aiming for zero is instrumental to lowering suicide rates. That, and some of the statements quoted above, can render these goals too vague to give action guidance.

We agree that unrealistic goal setting can be ethical when likely to lead to major improvements. However, such improvements require a clear path: the implementation of specific measures and specified subtargets that can credibly lead to the underlying goal.^{Reference Eyal and Sjöstrand33} For the goal of zero suicide, however, no such path has been presented. A recent audit concluded that it was difficult to evaluate the effects of the Swedish suicide-prevention strategy as it did not involve measurable goals or indicators.³⁴ Perhaps relatedly, Swedish suicide rates have remained largely unchanged since the overarching zero goal was adopted, whereas suicide rates decreased in comparable European countries.^35,36

Even if unintended, a myopic focus on suicide prevention may distract from patient-centred care practices towards defensive medicine and, at worst, create attempts at prevention that are ineffective and counterproductive.^{Reference Karlsson, Helgesson, Titelman, Sjostrand and Juth37} A more honest and more modest approach would curb these exigencies.

A middle path

How, then, should national and global public health address suicide? If zero suicide is not the goal, what should the goal be? When it comes to suicide, is there any acceptable goal other than zero?

In between the extremes of religious opposition and Kant on the one hand and dogmatic libertarianism on the other lies a middle path for public health. This middle path mirrors predominant positions on decision-making at the end of life in clinical bioethics. It takes seriously the patient's interests and autonomous will, as well as the pragmatic epidemiological considerations we relayed. It also takes seriously the roots of much suicide in impulse or in social injustice and suicide's typical devastating externalities in bereaved families and communities. None of the path's proposed elements is entirely novel, but their combination with our justification, which readily connects to widely accepted principles of healthcare, is surprisingly absent from current literature.

The proposed path recognises suicide as a central public health problem but also the ethical conflicts of suicide prevention. It rejects suicide-prevention efforts when they backfire and when they harm, burden or cost others too much in expectation. It accepts the possibility that in some situations, death can be the least bad option for an individual, and that a person's choice to die is sometimes autonomous and rational. Many patients who currently secure permission for PAS are cases in point. The middle path need not embrace a general legalisation of PAS but, importantly, it does not reject and work against it either. All this leaves wide room for appropriate suicide prevention under four constraints.

First, as we do not know who will die by suicide, suicide prevention should normally employ only benign methods or ones that improve individual and population health and living conditions, and remove conditions that lead up to suicide. Strengthening social and economic safety nets without which people find themselves in crises, and extending universal access to healthcare and high-quality mental health services would often restore quality of life, health and well-being.^{Reference Patel, Saxena and Lund38} Preventing suicide may be a by-product, albeit a highly desirable one. At the individual level, identifying psychosocial needs and offering person-centred care for psychiatric disorders and medical conditions, while strengthening individual autonomy, would help in more ways than one.^{Reference Duberstein, Marnin and Nock12}

Second, the middle path also accepts coercive interventions when truly necessary in order to prevent a likely decisionally incapacitated suicide, especially when the non-voluntary intervention makes their ensuing decisions more autonomous, for instance by providing treatment to restore decisional capacity, and by offering psychosocial interventions to improve living conditions. Such ‘soft paternalism’ aims to protect individuals from harm when they are unable to make autonomous decisions on the matter.^{Reference Miller and Wertheimer39} At the population level, the middle path accepts non-voluntary interventions on the basis of ‘soft group paternalism’,^{Reference Miller and Wertheimer39} which permits limited paternalistic policies towards everyone, to protect non-autonomous individuals from suicide. Soft group paternalism allows laws restricting access to means for impulsive suicide, including guns and certain pesticides. Many such efforts will also have other positive public health effects beyond suicide prevention, for example to reduce accidental gun injuries and interpersonal violence.

Third, interventions need not always make means for suicide completely unavailable (typically a decision with steep social costs), just not readily available; the resulting ‘lag time’ enables further deliberation and consultation through helplines and other clinical resources. The primary aim here should be to promote autonomy by enabling rational deliberation and by helping vulnerable individuals to get the help they need. Limited additional restrictions on means for suicide could rest on concern about negative externalities from even autonomous suicide.

Fourth, the middle path rejects goals known in advance to be unachievable as hypocritical, misleading and dangerous for public trust and public health campaigns. Treating suicide as a never-event risks exacerbating feelings of guilt, blame and potential burnout among practitioners and patients’ families, and, worse, stigma and shame among patients mulling the idea. Warranted suicide prevention is one of many important goals in healthcare and public health, but it is best served without a zero-suicide goal.