Most countries that permit assisted suicide do so only for patients who are terminally ill. In Switzerland, Belgium, the Netherlands and Luxemburg, however, the law is more permissive: assisted suicide is permitted independent of life expectancy, which means those patients suffering from severe mental illnesses such as treatment-refractory major depressive disorder, treatment-resistant chronic schizophrenia or severe persistent anorexia nervosa can also avail themselves of assisted suicide (or euthanasia, except in Switzerland). This more liberal legislation does, however, require that patients be of sound mind and have the decision-making capacity to consent to or request assisted suicide.

Four criteria for medical decision-making capacity are widely accepted: the ability to understand the relevant information, the ability to appreciate the disorder and the medical consequences of the situation, the ability to reason about treatment choices and the ability to communicate a choice.Reference Grisso and Appelbaum¹

The controversy surrounding decision-making capacity evaluations in patients seeking assisted suicide is because of the controversial nature of assisted suicide itself. Because it is unusual for doctors to be involved in deliberately ending a patients a life and because of the irreversibility of the decision, some believe that a higher standard of competence should be demonstrated to ascribe decision-making capacity to the patient.Reference Buchanan and Brock² Thus, the stringency of decision-making capacity requirements should increase in proportion to the potential risk to the patient.Reference Parker³ Some critics have even suggested that a request for assisted suicide is itself suggestive or even indicative of a lack of decision-making capacity.Reference Werth, Benjamin and Farrenkopf⁴ Given these fundamental disagreements, it is hardly surprising that there is considerable variation in experts' opinions regarding proposed threshold of decision-making capacity, the required time frame to prove that a request for assisted suicide is stable and settled, and whether/how many different doctors must certify decision-making capacity. In a 2000 survey, 58% of psychiatrists claimed that major depressive disorder means that a patient is de facto incompetent.Reference Ganzini, Leong, Fenn, Silva and Weinstock⁵ This is unfortunate, as mental illness or cognitive dysfunction do not in themselves indicate that a patient lacks decision-making capacity.Reference Grisso and Appelbaum¹ Many mental and cognitive conditions are compatible with having decision-making capacity, and a patient can have depressive disorder without lacking decision-making capacity.Reference Okai, Owen, McGuire, Singh, Churchill and Hotopf⁶ A recent systematic review concluded that depression can influence DMC, but not that it always does so.Reference Hindmarch, Hotopf and Owen⁷

It appears that the DMC evaluation in AS is often affected by a sort of ethical contamination, with relatively simple cases of patients with a terminal illness requesting who request assisted suicide being conflated with more challenging cases of assisted suicide requests from patients who are not terminally ill; in addition, patients in both these categories may also have a psychiatric disease, making four categories in total. It is important to bear in mind that all four categories could include both patients who are mentally competent and those who not. Indeed, one of the reasons why standards are often set so high for those who are terminally ill but have no mental health issues is that concerns about decision-making capacity and lack of terminal illness can contaminate and infect discussions of even ‘simple’ requests for assisted suicide. One reason for this might be that it is not always obvious which category a patient will fall into.

We should prevent patients who are mentally incompetent from harming themselves; in this case, the harm would consist in helping them end their lives when they are not fit to make such a decision. The principle of respect for autonomy tells us that we should not prevent patients who can make autonomous decisions from accessing assisted suicide if they wish to do so. Similarly, the principle of non-maleficence indicates that we should not inflict harm upon patients who are mentally competent by insisting that they remain alive and suffering. More generally, the principle of justice prohibits unfair discrimination, and a decision-making capacity evaluation that stops patients who are mentally competent from accessing assisted suicide would be unjust. As Schuklenk and Van de Vathorst have stated, ‘Erring on the side of caution… would imply also to let a fair number of patients suffer.’Reference Schuklenk and Van de Vathorst⁸

All of these arguments are particularly applicable to the context of patients who are terminally ill and wish to access assisted suicide, for two main reasons. First, their autonomy should be respected; most such patients are no more likely to be incompetent than other patients of a similar age (in contrast, those who seek suicide in the absence of terminal illness raise more concerns regarding autonomy). Second, the potential harm of lost time alive to them is less than for patients who are not terminally ill; they are relatively close to death and simply want to avoid the potentially immense suffering in the last few months of their life. A contrasting argument might be that patients who are terminally ill will not suffer for very long, and that caution is therefore justified. However, the counter-objection is that caution is more important in cases where the patient is not terminally ill, and the consequent loss of life will be greater. In the case of patients who are terminally ill, they will not lose much life if they are granted assisted suicide, but they will avoid great suffering. The harms of unjustified assisted suicide are also great; patients without decision-making capacity should not be able to access it, in case their decision would be different if they had decision-making capacity. But the threshold for justification should not be unreasonably high.

One way in which the bar is sometimes set too high for patients who are trying to access assisted suicide is by broadening the scope of decision-making capacity evaluations too far beyond issues of capacity. One example of this is using an inappropriately high standard of ‘stability’ of attitudes, in line with the Swiss Academy of Medical Sciences (SAMS) criteria, which mention that a patient's wish must persist over time.⁹ In principle, a patient can be competent and still be ambivalent about a wish. This can result in unstable and changing requests over time in a patient with competency. However, unstable attitudes can also be the consequence of fluctuating cognitive and autonomous mental capacities.Reference Trachsel, Hermann and Biller-Andorno¹⁰ This shows that it is important to be more specific about decision-making capacity evaluations in patients who are terminally ill and to explain in more detail what should be evaluated. Otherwise, for example, doctors opposed to assisted suicide could insist that a long period of time passes before it can be provided to a terminal patient with early-stage dementia. If the period is too long, the patient will either have already died or will no longer pass a decision-making capacity evaluation at the end of it because their disease will have progressed. Because of these considerations, the degree of required stability should be reasonably proportional to the length of time before death, and should take into account the previous expressed attitudes of a patient. If someone has always been strongly autonomous and previously mentioned the intention to avoid suffering at the end of life, there is no reason to wait for another 3 weeks of stable wishes after a request for assisted suicide; 1 week could be enough. But in cases where the patient had not expressed strong pro-autonomy views or had previously changed their mind about assisted suicide, it would be prudent to re-evaluate stableness more often and for a longer period.

Another issue is coercion. To find an example of a case where concerns about voluntariness and coercion contaminated decision-making capacity evaluations, one need look no further than Local Authority v Z.,¹¹ where an English welfare authority attempted to prevent a woman travelling to Switzerland to seek assisted suicide. In this case, she was found competent, but it was stated that she might have been found incompetent if she could not ‘assimilate the issues, or fully appreciate the consequences’ or if she was ‘unduly influenced by the views of others or by undue concern for the burden that her condition imposed on others’.Reference Trachsel, Hermann and Biller-Andorno¹⁰ The first quote does concern decision-making capacity, but the second does not; undue influence concerns coercion, and the burden argument also relates to questions about voluntariness. It would be a mistake if a patient were diagnosed incompetent purely on these grounds. Of course, if coercion is indeed present, assisted suicide should not be provided, but this example illustrates how other ‘excuses’ can creep into what should be a narrow evaluation of decision-making capacity.

In some cases, of course, concerns about coercion are very closely linked to concerns about decision-making capacity, and about stability; for example, someone who technically meets decision-making capacity criteria might nonetheless feel under some pressure because of the financial strain they are placing on the family; if combined with a relatively unstable wish for assisted suicide, doctors might be correct to be concerned about decision-making capacity, all things considered. The interrelatedness of these concepts makes it quite easy for doctors opposed to assisted suicide to make excuses for not granting requests, and care must be taken to keep these concepts separate whenever possible.

Any doctor who attempts to prevent a patient who is mentally competent from accessing assisted suicide is adopting an over-paternalistic stance that is contrary to the more general emphasis on autonomy in biomedical ethics. One might never choose assisted suicide for oneself or might think that the practice itself is deeply unethical, but to impose those values on one's patients is deeply unethical and unprofessional.

In terms of specific rules, we recommend that physicians use the approach summarised in Table 1.

Table 1 Recommendations for assessment of decision-making capacity in assisted suicide requests.

Decision-making capacity evaluation in patients requesting assisted suicide can be complex, but our analysis shows that doctors may be exaggerating the complexity of it in some cases. For patients at the end of life, decision-making capacity evaluations should be relatively straightforward, even if they have mental health issues. For those who are not yet at the end of life, evaluating capacity can be more difficult and a higher standard may be justified, but care must be taken to avoid letting other considerations contaminate the decision-making capacity evaluation. Most importantly, doctors should not let any personal qualms about assisted suicide to infect the objectivity of the decision-making capacity evaluation.