Equality in the Informed Consent Process: Competence to Consent, Substitute Decision-Making, and Discrimination of Persons with Mental Disorders

Abstract

According to what we propose to call “the competence model,” competence is a necessary condition for valid informed consent. If a person is not competent to make a treatment decision, the decision must be made by a substitute decision-maker on her behalf. Recent reports of various United Nations human rights bodies claim that article 12 of the Convention on the Rights of Persons with Disabilities involves a wholesale rejection of this model, regardless of whether the model is based on a status, outcome, or functional approach to competence. The alleged rationale of this rejection is that denying persons the right to make their own treatment decisions based on an assessment of competence necessarily discriminates against persons with mental disorders. Based on a philosophical account of the nature of discrimination, we argue that a version of the competence model that combines supported decision-making with a functional approach to competence does not discriminate against persons with mental disorders. Furthermore, we argue that status- and outcome-based versions of the competence model are discriminatory.

I. INTRODUCTION

It is widely accepted that competence is a necessary condition for informed consent (Meisel, Roth, and Lidz, 1977; Beauchamp and Childress, 2013). In this view, if a person is not competent to make a specific treatment decision, the treatment decision must be based on an advance directive or made by a substitute decision-maker based on the substituted judgment standard or the best interest standard (Buchanan and Brock, 1990; Grisso and Appelbaum, 1998; Vollmann, 2000; Kim, 2010). We will call this the “competence model.”

The nature of the competence model varies, depending on how the criteria for competence are spelled out.¹ This article is intended as an intervention in the debate among disability scholars about the discriminatory nature of the competence model. For this reason, we adopt the distinction, familiar in this debate, between a status, an outcome, and a functional approach to competence (Dhanda, 2007; McSherry, 2012; Flynn and Arstein-Kerslake, 2014a; Committee on the Rights of Persons with Disabilities, 2014, 2018).² These labels are broad characterizations of approaches to competence found in the law and medical practice. On the status approach, incompetence is understood in terms of the presence of a mental disorder. On the outcome approach, incompetence is understood in terms of the substantive irrationality of the treatment decision. Finally, on the functional approach, incompetence is understood in terms of a substantial impairment of functional psychological capacities related to decision-making. We will refer to the compound of these capacities by the term “decision-making capacity” (DMC).³

The United Nations (UN) Convention on the Rights of Persons with Disabilities (CRPD) presents a major challenge to the competence model. According to a recent report on mental health and human rights prepared by the UN High Commissioner of Human Rights (High Commissioner), article 12 of the CRPD should be interpreted as implying a wholesale rejection of the competence model. The report claims that to achieve compliance with article 12, “states should repeal legal frameworks allowing substitute decision-makers to provide consent on behalf of persons with disabilities” (2017, par. 26).⁴ The report adopts the interpretation of CRPD article 12 proposed by the UN Committee on the Rights of Persons with Disabilities (the Committee). In its highly influential General Comment No. 1 on article 12, the Committee similarly claims that “states parties have an obligation not to permit substitute decision-makers to provide consent on behalf of persons with disabilities” (2014, par. 41).⁵

The second paragraph of article 12 is crucial here. It states that “states parties shall recognize that persons with disabilities enjoy legal capacity on an equal basis with others in all aspects of life.” Since we pursue systematic aims in this paper, this is not the place for lengthy legal commentary. We follow the Committee’s interpretation of article 12 to avoid turning a substantive disagreement into an interpretative quarrel. On this interpretation, paragraph 2 renders it impermissible to make the recognition of legal capacity conditional on whether a person’s DMC meets a certain threshold: “All persons, regardless of disability or decision-making skills, inherently possess legal capacity” (2014, par. 25). Since the act of making a treatment decision involves an exercise of legal capacity, the Committee’s interpretation implies that persons with disabilities may not be deprived of the right to make their own treatment decisions on the basis of impaired DMC. As the Special Rapporteur on the rights of persons with disabilities puts it, “all forms of substitute decision-making are prohibited under the Convention, including those based on the assessment of mental capacity skills” (2018, par. 26). We call this the “strict interpretation” of article 12.⁶

Unsurprisingly, the strict interpretation of CRPD article 12 has provoked an intense debate. Whereas disability theorists and advocates see the strict interpretation of the article as a vehicle for the emancipation of persons with mental disorders (Dhanda, 2007; Minkowitz, 2007; Quinn, 2010; Arstein-Kerslake and Flynn, 2017; Degener, 2017), other scholars have argued that it makes these persons worse off (Ward, 2014; Freeman et al., 2015; Appelbaum, 2016). Although we strongly endorse the CRPD’s principles and aims, we find ourselves on the latter side of the debate. Elsewhere, we have identified six adverse consequences of the strict interpretation of CRPD article 12 for persons with psychosocial disabilities or mental disorders (Scholten and Gather, 2018).⁷

Some proponents of the strict interpretation may want to bite the bullet and accept the potential adverse effects of their proposed policy on persons with mental disorders. After all, as any Kantian will tell, sometimes one must stick to principle no matter the consequences. The Committee claims that the strict interpretation of article 12 is premised on the CRPD’s principle of nondiscrimination (2014, par. 4, par. 32; 2018, par. 7, par. 47). According to CRPD (2006) article 4(b), states parties have the obligation “to take all appropriate measures, including legislation, to modify or abolish existing laws, regulations, customs and practices that constitute discrimination against persons with disabilities.” The Committee holds that the competence model discriminates against persons with mental disorders regardless of whether it is based on a status, outcome, or functional approach (2014, par. 15). This is made explicit in the Committee’s recently published General Comment No. 6 on equality and nondiscrimination:

Discrimination through denial of legal capacity may be present in different ways, including status-based, functional and outcome-based systems. Denial of decision-making on the basis of disability through any of these systems is discriminatory . . . States parties should reform existing legislation to prohibit discriminatory denial of legal capacity, premised on status-based, functional or outcome-based models. (2018, par. 47–49)

We wholeheartedly embrace the CRPD’s principle of nondiscrimination. Working toward compliance with this principle is a matter of urgency because persons with mental disorders continue to experience a wide range of human right violations (Drew et al., 2011), as well as discrimination in the context of family, friendship, intimate relationship, and work (Thornicroft, 2006; Thornicroft et al., 2009; Lasalvia et al., 2013; Sayce, 2016). However, we contest that the principle of nondiscrimination entails the strict interpretation of article 12. Agreeing that all discriminatory denials of legal capacity must be abolished, in this article we argue that a competence model based on the functional approach does not discriminate against persons with mental disorders. That is, we contend that it is not discriminatory to deny persons the right to make treatment decisions on the basis of substantially impaired DMC.⁸ We substantiate this claim with an account of the nature of discrimination, drawing from recent debates in philosophy. Based on this account, we furthermore show that status- and outcome-based versions of the competence model do discriminate against persons with mental disorders.

This article is structured as follows. In the following section, we reconstruct the argument that underlies the strict interpretation of article 12. We develop an account of discrimination and propose a refinement of the CRPD’s definition in Section 3. Subsequently, we use this account of discrimination to assess whether the various approaches to competence are discriminatory. In Sections 4 and 5, we contend that, respectively, the status and the outcome approach discriminate against persons with mental disorders. In Section 6, we argue that the functional approach does not.

II. THE ARGUMENT FROM DISCRIMINATION

According to the Committee (2014, par. 4), the strict interpretation of article 12 derives from the CRPD’s general principles, notably the principles of equality of opportunity and nondiscrimination. Here we reconstruct this derivation in detail. The first two paragraphs of CRPD (2006) article 5 assert an explicit prohibition of discrimination on the basis of disability:

1. States Parties recognize that all persons are equal before and under the law and are entitled without any discrimination to the equal protection and equal benefit of the law.

2. States Parties shall prohibit all discrimination on the basis of disability and guarantee to persons with disabilities equal and effective legal protection against discrimination on all grounds.

Scholars have noted that many human rights documents that prohibit discrimination fail to define the concept (Altman, 2015). The International Covenant on Civil and Political Rights and the European Convention on Human Rights provide notable examples. The CRPD fares better in this respect. Article 2 contains the following definition of discrimination on the basis of disability:

“Discrimination on the basis of disability” means any distinction, exclusion or restriction on the basis of disability which has the purpose or effect of impairing or nullifying the recognition, enjoyment or exercise, on an equal basis with others, of all human rights and fundamental freedoms in the political, economic, social, cultural, civil or any other field. (CRPD, 2006)

This definition ties discrimination conceptually to impairments of human rights and fundamental freedoms. The question accordingly arises whether the enjoyment of legal capacity is a human right. Interestingly, the CRPD itself is silent on this issue. The Committee’s understanding, however, is that this idea is an integral part of the CRPD. It claims that “legal capacity is a universal attribute inherent in all persons by virtue of their humanity” (2014, par. 8), that “legal capacity is an inherent right accorded to all people, including persons with disabilities” and, finally, that “all people, including persons with disabilities, have legal standing and legal agency simply by virtue of being human” (par. 14). This premise is widely endorsed by disability scholars (Dhanda, 2007, 2012; Minkowitz, 2007; Bach and Kerzner, 2010; Quinn and Arstein-Kerslake, 2012; Flynn and Arstein-Kerslake, 2014a).

Legal capacity comprises both legal standing and legal agency, where legal standing refers to the capacity to have rights and legal agency to the capacity to make legal transactions (McSherry, 2012; Committee, 2014, par. 12–13). Since the act of giving informed consent involves an exercise of legal agency (Faden and Beauchamp, 1986), we will focus only on legal agency. The rationale underlying the second paragraph of article 12 can be reconstructed as follows:

• (P1) It is discriminatory to make a distinction on the basis of disability having the purpose or effect of impairing or nullifying the recognition, enjoyment, or exercise of human rights or fundamental freedoms. (article 2)

• (P2) Legal agency is a human right.

• (C1) It is discriminatory to deny legal agency on the basis of disability. (from P1 and P2)

• (P3) Discrimination is impermissible. (article 5)

• (C2) It is impermissible to deny legal agency on the basis of disability. (from C1 and P3)

We will refer to this argument as the Argument from Discrimination. The gist of the argument is that “states parties must abolish denials of legal capacity that are discriminatory on the basis of disability in purpose or effect” (Committee, 2014, par. 25).

Our formalization of the Argument from Discrimination makes explicit how the prohibition on denying legal agency can be derived from the CRPD’s prohibition of discrimination. The Committee explains in a footnote that the prohibition on denying legal agency articulated in article 12 follows from article 2, in conjunction with article 5 (2014, par. 25n.). The Committee applies a similar reasoning when it explains the relationship between article 12 and the prohibition of discrimination expressed in article 5 (2014, par. 32). However, our reconstruction reveals that this is incorrect and that (P2) must be added to make the argument complete.

Responding to the challenge posed by CRPD article 12, proponents of the functional approach to competence have argued that the construct of DMC is essentially disability neutral (Szmukler, Daw, and Callard, 2014; Dawson, 2015; Szmukler, 2019). If so, denying legal agency on the basis of impaired DMC does not pass for discrimination according to the CRPD’s definition. The Committee denounces this strategy decisively, however, claiming that it is based on a conflation of mental capacity with legal capacity (2014, par. 15) and reasserting that “all persons, regardless of disability or decision-making skills, inherently possess legal capacity” (par. 25). As Richardson puts it emphatically, the point of the strict interpretation is precisely that “legal capacity should not depend on mental capacity” (2012, 345).

The Committee’s response shows that one can substitute “impaired DMC” for “disability” in the Argument from Discrimination. Its conclusions then run as follows:

• (C1) It is discriminatory to deny legal agency on the basis of impaired DMC.

• (C2) It is not permitted to deny legal agency on the basis of impaired DMC.

We do not challenge the strict interpretation of article 12 to avoid turning a substantial disagreement into a philological dispute. Yet we do challenge the Argument from Discrimination. In fact, we reject both (C1’) and (C2’). Since the argument is logically valid, this forces us to reject at least one of its premises.

We reject (P2) because the premise is clearly false. Although we agree that legal standing is a human right, we contest that legal agency (i.e., the right to make legal transactions) is a universal attribute inherent in all persons simply by virtue of their humanity. If the enjoyment of legal agency were such a human right, banks would commit human rights violations by refusing mortgages to wealthy 8-year-olds. We take it that this reductio suffices to show that to be granted the right to make legal transactions requires more than just being human.

Some might want to resist our counterexample and hold on to the idea that legal capacity is a universal human right. To accommodate the intuition that banks do not commit human rights violations by refusing mortgages to wealthy 8-year-olds, one might contend that 8-year-olds do have the right to make legal transactions, but that they are merely not allowed to exercise this right until they reach a mature age. One might proceed to explain that the same holds true for persons with substantially impaired DMC: these persons possess the right to make legal transactions, and it is only that they are not allowed to exercise this right until they regain DMC. This dispute is merely verbal. That is, we agree on the subject matter and merely have a different understanding of what it means to have a right. For conceptual clarity, we understand it that one has a (liberty) right to do x only if one is allowed to do x.⁹

If the enjoyment of legal agency is not a human right, it follows that no denial of legal agency can possibly count as discriminatory according to the Argument from Discrimination. The reason is that, according to the CRDP’s definition, an action, practice, or policy is discriminatory only if it has the aim or effect of impairing human rights and fundamental freedoms. We believe this condition renders the definition too restrictive and, therefore, we propose a refinement in the following section. On this refined definition, actions, practices, and policies can count as discriminatory, even if they do not have the aim or effect of impairing human rights and fundamental freedoms. If this is correct, the question whether (P2) is true is irrelevant to whether an action, practice, or policy is appropriately called discriminatory. On the refined definition of discrimination, then, the Argument of Discrimination can do without this implausible premise, which in turn implies that denials of legal agency can potentially be discriminatory.

III. THE NATURE OF DISCRIMINATION

Scholars have by now started to explore the issue of whether and when denying legal capacity to persons with mental disorders is discriminatory (Weller, 2008; Callard et al., 2012; Burch et al., 2014; Nilsson, 2014; Dawson, 2015). In this literature, CRPD article 12 is interpreted in light of other articles and principles of the convention, as well as international discrimination law. What is missing is an in-depth conceptual analysis of discrimination. We aim to give such an analysis in the current section.

We draw on recent philosophical debates to formulate an account of discrimination. These debates are complex and multifaceted, and there are few distinctions that are accepted by all participants of the debate. We must thus start from a definition of discrimination that contains only the elements that can count on a reasonably broad consensus. Altman provides us with such a definition:

Discrimination consists of acts, practices or policies that impose a relative disadvantage on persons based on their membership in a salient social group. (Altman, 2015)

This definition of discrimination incorporates the key elements of the account proposed by Lippert-Rasmussen (2014). The definition contains two elements: first, persons who are discriminated against must incur a relative disadvantage, and second, they must incur this relative disadvantage based on their membership in a salient social group.

The first element has two aspects. To say that discrimination involves a disadvantage is to say that discrimination is more than merely differential treatment. Thus, on the proposed account, having separate but equally equipped toilets for male and female clients is not discriminatory against men or women, although of course it may well be discriminatory against persons who are transgender. To say that the disadvantage is relative is to say that discrimination is an essentially comparative concept. On the proposed account, persons are thus discriminated against only if they are treated disadvantageously compared to the way members of other social groups are treated. This implies that one can dismiss a charge of discrimination against the members of one salient social group by saying that one treats the members of other social groups equally as badly.

Now look at the second element. According to Lippert-Rasmussen, a group is socially salient “if perceived membership of it is important to the structure of social interactions across a wide range of social contexts” (2014, 30). Groups defined by attributes such as race, sex, sexual orientation, and religion are typical candidates for social salience. In contrast to trivial properties, such as one’s shoe size or arm length, the perceived possession of the former type of attributes is important to the structure of interactions between people across a wide range of contexts. Thus, it does not involve discrimination when the owner of a soccer club refuses to hire players who do not have the number 8 in their birth date.¹⁰

One could object to the idea that membership in a salient social group is a prerequisite for discrimination by arguing that the order of explanation is the other way around: it is not that disadvantageous treatment of certain persons counts as discrimination because these persons form a salient social group; these persons form a salient social group because they are discriminated against. The following example suggests that the account we propose gets the order of things right. Suppose that in a given society women are not treated disadvantageously in comparison to men and hence are not discriminated against. It seems perfectly consistent to suppose that in this society being female is nevertheless considered important to the structure of interactions between people across a wide range of contexts: when people go over to greet women, they kiss rather than shake hands, men ask them to dance on a night out and help them carrying heavy luggage, and so on. Now, if a single hiring committee in this society chooses to hire a male candidate even if a female candidate was better qualified for the job, it would seem that they discriminate against the female candidate. The most plausible explanation for this is that they treat her disadvantageously based on her membership in a salient social group.

Although it is not our intention to defend the proposed account of discrimination in the context of this article, it is helpful to contrast it with two accounts of discrimination that have intuitive appeal, but turn out to be inadequate on reflection. First, consider what might be called “the involuntary traits account.” On this account, to discriminate against a person is to treat that person disadvantageously on the basis of a trait that is beyond her voluntary control. Since persons typically cannot help being disabled, discrimination on the basis of disability may appear to be a subclass of this broader category. Bernard Boxill has argued convincingly that this conception of discrimination is inadequate. On the one hand, it would seem discriminatory to deny persons of a certain skin color access to a bar, even if there were a pill available that could change a person’s skin color; on the other hand, it does not seem discriminatory to deny a blind person a driver’s license or a person with little athletic ability a place on the basketball team (Boxill, 1992). This shows that treating a person disadvantageously on the basis of a trait that is beyond her voluntary control is neither necessary nor sufficient for discrimination.

Then consider what might be called “the irrelevance account” (Halldenius, 2018). On this account, to discriminate against a person is to treat that person disadvantageously on the basis of an irrelevant property. A well-known counterexample to this account is as follows (see Lippert-Rasmussen, 2014, 23). Suppose that an academic hiring committee hires a less-qualified applicant for a publicly advertised job because the dean favors members of his own academic circle and the chosen applicant is one of his academic friends. Even if the members of the committee treat the other applicants disadvantageously based on an irrelevant property (i.e., the property of not being a friend of the dean), most will think that this is a case of nepotism rather than one of discrimination. This shows that disadvantageous treatment based on an irrelevant property is not sufficient for discrimination.

Proponents of the irrelevance account could respond by claiming that all cases of nepotism are cases of discrimination, but this claim seems highly implausible. Even if they were to succeed in establishing this claim, however, it is possible to construct additional counterexamples that do not involve nepotism. Suppose a hiring committee has no time to consider the applications and therefore chooses simply to select the applicant whose surname comes first in alphabetical order (Scanlon, 2008, 70). Again, even if the other participants are treated disadvantageously based on an irrelevant property (i.e., not having a surname starting with the letter “a”), few people would think that raising discrimination charges would be appropriate in a case like this.

Another familiar counterexample against the irrelevance account of discrimination involves so-called “reaction qualifications,” that is, job qualifications based on how other people are thought to react on employees (Lippert-Rasmussen, 2014, 24). Suppose a shop owner chooses not to hire a member of a minority group as her new shop assistant because she believes—justifiably so, we stipulate for the sake of the argument—that customers prefer to be served by a shop assistant from the dominant social group. Even if under the stipulated circumstances the property of being a member of a minority group is relevant to the shop’s legitimate aim of profit maximization, most will say that the applicant is nonetheless discriminated against. This shows that disadvantageous treatment based on an irrelevant property is not necessary for discrimination either.

Let us now proceed and assess whether mental disorder is a candidate for social salience. It seems to be a particularly clear candidate. To give two striking examples, people tend to give a negative answer when they are asked if they would be willing to live next door to a person with a psychiatric condition (Angermeyer and Matschinger, 1997), and employers tend to find persons with mental disorders less acceptable as potential employees than ex-convicts (Brand and Clairborn, 1976). Recent research shows that, notwithstanding various antistigma campaigns, the reorganization of psychiatric practice and the growing recognition of mental disorder as a brain disorder, the attitudes of the general public toward persons with mental disorders have not changed for the better (Mehta et al., 2009; Pescosolido et al., 2010: Angermeyer, Matschinger, and Schomerus, 2013). Persons with mental disorders can thus be said to form a salient social group.

In closing this section, let us compare the proposed account of discrimination with the definition in the CRPD. To recapitulate, CRPD (2006) article 2 defines discrimination on the basis of ability as “any distinction, exclusion or restriction on the basis of disability which has the purpose or effect of impairing or nullifying the recognition, enjoyment or exercise, on an equal basis with others, of all human rights and fundamental freedoms.” Although this definition of discrimination is not explicitly comparative and does not explicitly mention the notion of a relative disadvantage, it is plausible to read the clause “on an equal basis with others” as introducing such a comparative element. Based on this reading, a government that illegitimately appropriates the possessions of both people with disabilities and people without disabilities does not discriminate against persons with disabilities, although it no doubt violates human rights.

Whereas our proposed account defines discrimination in terms of a relative disadvantage, the CRPD ties discrimination to the violation of human rights and fundamental freedoms. The CRPD’s definition of discrimination seems too restrictive on account of this. After all, there seem to be many cases of discrimination that do not involve the impairment of human rights or fundamental freedoms. To give but one example, employers who give their employees without disabilities but not their employees with disabilities special privileges clearly discriminate against their disabled employees, yet they do not violate any human rights or fundamental freedoms (i.e., apart from the right to nondiscrimination itself).¹¹ Since our account broadens the scope of discrimination and lowers the threshold for valid discrimination claims, we expect that supporters of the CRPD will embrace the proposed refinement.

Turn, finally, to the second element of discrimination. CRPD article 2 defines a subclass of discrimination, namely, discrimination on the basis of disability. According to the definition, an act, practice, or policy must impose a “distinction, exclusion or restriction on the basis of disability” to fall under this subclass (CRPD, 2006). This requirement can be plausibly read as a particular version of the more general requirement that a distinction must be made on the basis of membership of a salient social group in order to count as discriminatory. The notion of disability has been markedly absent in the lists of attributes defining salient social groups in the most important human rights documents predating the CRPD (Degener, 2000). Article 14 of the 1950 European Convention on Human Rights and article 26 of the 1966 International Covenant on Civil and Political Rights, for example, provide the following list: “race, color, sex, language, religion, political or other opinion, national or social origin, property, birth or other status.” The CRPD is the first human rights document that mentions disability explicitly as a property defining a salient social group. Adding disability to the list of attributes that can give rise to discrimination claims can be considered a major achievement of the CRPD.

IV. THE STATUS APPROACH

Based on the proposed account of discrimination, we can investigate whether it is discriminatory to deny persons who are assessed as not competent to consent the right to make treatment decisions. We noted that the nature of the competence model varies depending on how the criteria for competence are fleshed out. Recall that the Committee rejects all three approaches to competence (i.e., the status, the outcome, and the functional approach) as discriminatory (2018, par. 47). By contrast, we argue that a competence model based on the functional approach is not discriminatory. In the current and following section, however, we first show that status- and outcome-based versions of the competence model discriminate against persons with mental disorders. According to the proposed account of discrimination, the question we must address is whether the relevant version of the competence model imposes a relative disadvantage on persons with mental disorders based on their membership in a salient social group.

On the status approach, incompetence is understood in terms of a psychiatric diagnosis or presence of a mental disorder. More specifically, to regard a psychiatric diagnosis or the presence of a mental disorder as either necessary or sufficient for incompetence is to adopt a status approach to competence. If a person’s diagnostic status is understood to be only a necessary condition for incompetence, it is typically also considered a necessary condition that treatment choice under consideration is a consequence of the mental disorder. Despite the latter qualification, incompetence is likely to be understood as permanent and global on this approach (Roth, Meisel, and Lidz, 1977). To be more precise, it is likely to be seen as having the same duration as the diagnosis and as holding for all or at least a range of decisions.

The status approach to competence has attracted heavy criticism in the last few decades (Campbell and Heginbotham, 1991). Notwithstanding this criticism, there is empirical evidence showing that healthcare practitioners still frequently infer incompetence from a psychiatric diagnosis (Markson et al., 1994), thereby implicitly taking the presence of a mental disorder not merely as a necessary, but even as a sufficient condition for incompetence. In a study conducted by Ganzini et al. (2003), for example, 66 percent of 395 surveyed consultant liaison psychiatrists and psychologists responded that it is common or very common for practitioners to believe that if a person has a diagnosis of Alzheimer’s disease or another dementia, even if mild, the person lacks competence in relation to all medical decisions; similarly, 70 percent of respondents rated it as common or very common for practitioners to believe that if a person has a mental disorder such as schizophrenia, the person lacks competence to make any medical decisions. The status approach is thus by no means a fiction from the past.

The law often considers a diagnosis of a mental disorder as a necessary condition for overriding treatment decisions. Two legal reviews revealed that mental health laws in many countries still deny legal agency to persons on the basis of the presence of a “mental disorder” or “mental illness” in combination with a perceived dangerousness to self or others (Salize, Dressing, and Peitz, 2002; Fistein et al., 2009). This approach also attracted criticism and it has been argued that conventional mental health laws discriminate against persons with mental disorders (Dawson and Szmukler, 2006; Szmukler, Daw, and Dawson, 2010; Szmukler, Daw, and Callard, 2014).

Is the status approach discriminatory on our proposed account? We would argue that it is. First, insofar as the status approach defines incompetence at least partly (i.e., as a necessary condition) in terms of a diagnosis of mental disorder, it treats persons differently based on their membership in a salient social group. Second, insofar as it limits the potential denial of legal capacity to persons with mental disorders, it imposes a relative disadvantage on persons with mental disorders in the sense that persons with mental disorders can be deprived of the right to make treatment decisions, whereas the same does not hold for persons without mental disorder.

In defense of the status approach, one might argue that this relative disadvantage is outweighed by the benefit of protection that persons with mental disorders enjoy under a regime of substitute decision-making. This argument does not hold ground. A large body of research shows that although conditions such as dementia, schizophrenia, mania, and depression are risk factors for impaired DMC, these patient groups are in themselves highly heterogeneous with respect to DMC (Vollmann et al., 2003; Kim, 2010). In view of this, it is plausible to assume that many members of these groups are in a better position to make decisions regarding their own treatment than the most accomplished substitute decision-maker. By being deprived of the right to make treatment decisions, these persons thus incur a relative disadvantage without receiving any benefit in return. A competence model based on the status approach therefore imposes a relative disadvantage on persons based on their membership in a salient social group. In other words, a status-based competence model is discriminatory.

V. THE OUTCOME APPROACH

On the outcome approach, incompetence is understood in terms of the substantive irrationality of treatment decisions. More specifically, to adopt an outcome approach to competence is to regard a substantively irrational treatment decision as either necessary or sufficient for incompetence.

To understand the outcome approach, it is important to distinguish substantive rationality from formal rationality. A treatment decision is substantively rational or irrational, analogous to the way in which a conclusion of an argument is true or false, whereas a decision-making process is formally rational or irrational, analogous to the way in which an argument is valid or invalid. On the outcome approach, the rationality of treatment decisions is thus not measured against a subjective standard (e.g., consistency with the decision-maker’s values and starting premises), but rather against an objective standard. In the clinical context, a person’s treatment decision is often deemed substantively irrational when the decision is not in her “medical best interest.”

Although only a few jurisdictions adopt an outcome approach to competence (Fistein et al., 2009), the outcome approach is still widespread in clinical practice. In the aforementioned survey study among 395 consultation-liaison psychiatrists and psychologists carried out by Ganzini et al. (2003), 89 percent of respondents rated it as common or very common that healthcare professionals fail to consider the possibility that a patient may be incompetent, as long as the patient agrees with the recommended treatment. In the same study, 88 percent of respondents rated it as common or very common for clinicians to give greater weight to the content of the patient’s decision than to the decision-making process by which the decision came about.

Let us first assess what may be called the pure outcome approach. The pure outcome approach defines incompetence fully in terms of the substantive irrationality of treatment decisions. It can readily be seen that this approach is strongly in tension with liberal values. It is a cherished liberal principle that persons may choose as they please as long as they do not violate the rights of others. Since the pure outcome approach allows doctors to override their patients’ treatment choices whenever they are allegedly not in their best interest, liberals will see the approach as unduly paternalistic.

Although clearly in conflict with liberal values, the pure outcome approach does not seem to discriminate against persons with mental disorders: persons are competent as long as they make substantively rational decisions, and incompetent whenever they make a substantively irrational decision, and this holds regardless of whether they have a mental disorder.

Given its illiberal nature, however, few people would be willing to endorse the pure outcome approach on reflection. One strategy to address the liberal worry is to modify the approach by saying that persons are not competent to make a decision if and only if their decision is strongly against their medical best interest. This adjustment is not enough to take away the worry, since debates between paternalists and liberals typically revolve around cases where patients’ decisions are strongly against their medical best interest. The refusal of blood transfusion by a Jehovah’s Witness is a notable example. Another strategy is to introduce what is often called a “diagnostic threshold” for incompetence such that persons who make a substantively irrational treatment decision can be declared incompetent only if they do so on account of a mental disorder. This qualified outcome approach grants most persons the right to make substantively irrational treatment choices, and this might suffice to reassure liberals that persons will not be declared incompetent simply because they make treatment decisions that allegedly are not in their best interest.

Note, however, that the qualified outcome approach denies persons with mental disorders the right to make substantively irrational treatment decisions based on their membership in a salient social group. Indeed, by applying a diagnostic threshold, proponents of the outcome approach make the presence of a mental disorder a necessary condition for incompetence. Proponents of the outcome approach face a difficult dilemma on account of this: either their approach conflicts with liberal values or it collapses into a status approach.

Proponents of the qualified outcome approach could respond in the following way. True, they may admit, the qualified outcome approach treats persons with mental disorders differently based on their membership in a salient social group. Yet, the only thing it denies to persons with mental disorders is the right to make substantively irrational treatment decisions; and one cannot seriously claim that persons with mental disorders incur a relative disadvantage by being denied the right to do irrational things. Consequently, the qualified outcome approach does not discriminate against persons with mental disorders.

Although this argument may seem convincing at first blush, an example can illustrate that it underestimates the extent to which modern society is characterized by value pluralism. Recall that the outcome approach employs an objective standard of substantive rationality. Now assume that a person’s values and conception of the good life are radically different from the values and conception of the good life of the treatment team. Such situations are quite likely to occur in modern society. Assume, furthermore, that by means of a formally rational decision-making process the person arrives at a treatment decision that is consistent with her values and conception of the good. The decision happens to be a refusal of the treatment that was strongly recommended by the treatment team. In such situations, it is likely that the members of the treatment team agree among each other that the person’s treatment refusal is substantively irrational. The outcome approach would then allow them to override the person’s treatment refusal in the person’s alleged best interest.

Objections against this type of paternalistic intervention are strong and widespread in modern society. What is more, we grant people not merely the right to make objectively irrational decisions, but also the right to make decisions that are irrational by their own lights. That is why we do not go about snatching cigarettes from the mouths of unwilling smokers. On reflection, then, we highly value the right to make mistakes, and this implies that persons with mental disorders incur a relative disadvantage when they are denied this right. Moreover, since the diversity of lifestyles is arguably greater in psychiatry than in general health care, disagreements about the substantive rationality of treatment decisions are more likely to occur in mental health care. There is thus reason to think that the right to make substantively irrational treatment decisions has special importance for persons with mental disorders.

Based on this analysis, we can conclude that the version of the outcome approach that is likely to be used in liberal contexts (i.e., the qualified outcome approach) imposes a relative disadvantage on persons with mental disorders based on their membership in a salient social group. A competence model based on the qualified outcome approach is thus discriminatory, though less straightforwardly so than a status-based competence model.

VI. THE FUNCTIONAL APPROACH

The functional approach defines incompetence in terms of a substantial impairment of a person’s psychological and functional capacities related to decision-making (Grisso, 2003). On this approach, substantially impaired DMC is both necessary and sufficient for incompetence.

The formulation of the relevant abilities varies somewhat across the available models. According to the model developed by Buchanan and Brock (1990), persons are competent to make a particular treatment decision if, and only if, they are sufficiently able to understand the disclosure information, communicate a treatment decision, reason and deliberate about the consequences of the various treatment options, and evaluate these consequences in the light of a set of personal values. According to the model developed by Grisso and Appelbaum (1998), persons are competent to make a particular treatment decision if, and only if, they are sufficiently able to express a treatment choice, understand the disclosure information, appreciate that this information applies to their condition and process the information in a rational way (see also Appelbaum, 2007; Kim, 2010).

The MacArthur Competence Assessment Tool for Treatment (MacCAT-T) is a clinical tool that operationalizes the criteria of the latter model. Studies have shown that the use of this tool results in an extremely high level of interrater agreement among capacity evaluators (Cairns et al., 2005). This suggests that modest schooling can ensure that healthcare professionals apply the criteria for competence in a nonarbitrary way (Raymont et al., 2007).

Like the Committee, various disability theorists and advocates claim that the functional approach to competence is discriminatory. Bach and Kerzner, for example, assert that a competence model based on the functional approach “systematically discriminates against people with intellectual, cognitive, mental and communication disabilities” (2010, 66–67). Similarly, the International Disability Alliance (2010) writes that “functional testing for legal capacity constitutes disability-based discrimination.” Other scholars use more cautious formulations, but draw the same conclusion (Dhanda, 2007; Flynn and Arstein-Kerslake, 2014a, 2014b). Proponents of the functional approach, on the other hand, deny that it is discriminatory to deny persons legal capacity on the basis of impaired DMC (Burch et al., 2014; Szmukler, Daw, and Callard, 2014; Dawson, 2015; Dute, 2015; Appelbaum, 2016; Szmukler, 2019). Thus far, little progress has been made in this debate because the arguments on either side fail to find resonance with the opponents.

Let us briefly consider the argument developed by Dawson (2015) to illustrate this. Without giving a conceptual analysis of discrimination, Dawson starts from the assumption that wrongful discrimination involves treating people differently based on an irrelevant property. Subsequently, he claims that having impaired DMC is relevant to whether one should have the right to make one’s own treatment decisions. If DMC is such a relevant property, it follows that it is not wrongfully discriminatory to deny persons the right to make treatment decisions based on impaired DMC. Other scholars have endorsed this argument (Appelbaum, 2016).

This strategy has proven ineffective. The reasons are clear: since Dawson’s opponents insist on a strict distinction between mental and legal capacity, they precisely deny that a person’s DMC is a relevant property. Based on the same account of discrimination, these scholars thus conclude that a competence model based on the functional approach is discriminatory because it denies persons the right to make treatment decisions on the basis of what they take to be an irrelevant property. In this way, the debate reaches an impasse.

Our proposed account of discrimination promises a way out. Recall that in Section 3, we showed that disadvantageous treatment based on an irrelevant property is neither necessary nor sufficient for discrimination. If so, the debate about whether DMC is a relevant or irrelevant property is beside the point. The question we must address instead is whether a competence model based on the functional approach imposes a relative disadvantage on persons with mental disorders based on their membership in a salient social group.

Based on Membership in a Salient Social Group?

It is helpful to start addressing this question by briefly surveying the Committee’s reasons for rejecting the competence model. The Committee notes that “a person’s status as a person with a disability or the existence of an impairment (including a physical or sensory impairment) must never be grounds for denying legal capacity” (2014, par. 9). Furthermore, it writes that “‘unsoundness of mind’ and other discriminatory labels are not legitimate reasons for the denial of legal capacity” (par. 13) and, finally, that “denial of legal capacity must not be based on a personal trait such as gender, race, or disability” (par. 32). We agree on all points. In contrast to the Committee, however, we do not think that this requires a rejection of a version of the competence model based on the functional approach.

The functional approach defines incompetence completely in terms of a substantial impairment of a person’s DMC with respect to a specific decision at a specific point of time. On this approach, then, the aforementioned status attributes are not sufficient for incompetence. Psychiatric diagnoses, psychotic symptoms, and low scores on cognitive tests are only indirectly relevant on this approach, namely, only insofar as they compromise a person’s DMC, which of course they need not do (Kim, 2010). Furthermore, on the functional approach, having a diagnosis of a mental disorder is not a necessary condition for incompetence either. A cross-sectional study that operationalized functional criteria found that incompetence is common among acute medical inpatients without mental disorder (Raymont et al., 2004). On the functional approach, the criteria for competence thus indiscriminately apply to everyone, regardless of whether one has a mental disorder. It follows that a competence model based on the functional approach does not deny persons the right to make treatment decisions based on their membership in a salient social group.

The proposed functional approach must be distinguished from quasi-functional approaches typically found in the law. The Mental Capacity Act, 2005 (MCA) in England and Wales is a case in point. Section 2, paragraph 1, of this act pronounces that “a person lacks capacity in relation to a matter if at the material time he is unable to make a decision for himself in relation to the matter because of an impairment of, or a disturbance in the functioning of, the mind or brain.” Accordingly, a person is incompetent to make a decision according to MCA criteria only if her inability to make the decision is explained by “an impairment of, or a disturbance in the functioning of, the mind or brain,” where the latter phrase is usually interpreted as referring to the presence of a mental disorder. Unlike the proposed functional approach, the MCA employs a diagnostic threshold for incompetence and hence treats persons with mental disorders differently based on their membership in a salient social group.¹²

Some might not be convinced by our proposal yet. Granting that incompetence is not defined in terms of any status attributes on the proposed functional approach, they might argue that one nonetheless becomes a member of the salient social group of “incompetent” persons by being assessed as not competent to consent. This is, first of all, problematic because incompetency labeling potentially produces serious adverse psychological effects (Winick, 1995). Moreover, if the functional approach implies that one will be denied the right to make treatment decisions based on one’s membership in the salient social group of “incompetent” persons, the functional approach is eventually still discriminatory.

To remove this objection, we must show that incompetence itself is not a status attribute on the functional approach. It is not a status attribute for two reasons. First, there is a wide consensus among proponents of the functional approach that competence is task specific and not global. This is because different treatment decisions have different levels of complexity and hence require different levels of DMC.¹³ If competence is task specific, the fact that a person is incompetent to consent to a complex medical intervention does not entail that the person is incompetent to consent to the use of a particular type of medication.

Second, there is a wide consensus among proponents of the functional approach that incompetence is time indexed and not permanent. Our DMC evidently fluctuates over time. Since the functional approach defines competence in terms of a relevant threshold of DMC, it follows that our competence to consent may change over time as well. A person’s DMC, for example, may fall below the relevant threshold upon admission to hospital and cross the threshold once the person has adapted to the new surroundings. Consequently, a determination of incompetence regarding a particular treatment decision does not transfer to later decisions of the same kind.

In various jurisdictions, the law likewise defines competence as task specific and time indexed. If competence is task specific and time indexed, incompetence is not a status attribute. On the functional approach, then, one does not become a member of a salient social group by being deemed incompetent to make a particular treatment decision. The objection is thus removed. An important implication of the time- and task-indexed nature of competence is that plenary and permanent guardianship is unacceptable on the functional approach.

A Relative Disadvantage?

Whether persons incur a relative disadvantage by being denied the right to make a decision for which they have been found incompetent remains to be determined. We hold that healthcare professionals have a positive obligation to provide decision-support before they take recourse to substituted decision-making. This obligation derives from the demand for reasonable accommodation. While CRPD article 2 already stipulates that discrimination includes denial of reasonable accommodation, the demand itself is expressed in the third paragraph of CRPD (2006) article 5: “In order to promote equality and eliminate discrimination, States Parties shall take all appropriate steps to ensure that reasonable accommodation is provided.” Accordingly, withholding decision-support to persons with impaired DMC involves discrimination.

Here, we see a need to distinguish between what we propose calling “input” and “process” support. Input support involves enhancing persons’ DMC by influencing situational factors that are negatively correlated with DMC. Process support, on the other hand, involves interpreting the person’s current wishes and preferences, carrying out the intellectual processing required to translate these into actual decisions, and communicating these decisions to others (Bach and Kerzner, 2010, 86–90; Flynn and Arstein-Kerslake, 2014a, 94–98). The evidence base for the latter type of support is very meager, and serious ethical concerns have been raised with regard to it (Kohn, Blumenthal, and Campbell, 2013; Kohn and Blumenthal, 2014; Craigie, 2015; Burch, 2017). The prospects for input support are much better. Various studies have shown that relatively simple interventions can significantly enhance the DMC of persons with, for example, schizophrenia, bipolar disorder, or mild forms of dementia (Nishimura et al., 2013). Input support can thus be considered an evidence-based intervention.

We hold that healthcare professionals have an obligation to provide input decision-support to persons with impaired DMC and that states have the obligation to provide the necessary resources and make the necessary legal and institutional arrangements. Up to now, only a few countries have made explicit provisions for supported decision-making (Pathare and Shields, 2012).

The aim of input decision-support is to enhance the DMC of persons, ideally up to a point where substitute decision-making becomes superfluous. There will be cases, however, in which decision-support is not sufficient for a person’s DMC to cross the relevant threshold. If a person’s DMC remains below the relevant threshold despite the provision of support, the person is incompetent to make the treatment decision, and recourse to substitute decision-making must be taken.

Persons have an interest in making their own decisions, even when their DMC is substantially impaired. When treatment choices of persons with impaired DMC are overridden, this interest is obviously frustrated. Since this does not hold for other persons, it may appear as though the competence model based on the functional approach imposes a relative disadvantage on persons with impaired DMC. But we have to consider other interests as well. Buchanan and Brock (1990) note that competence is a social construct enabling us to strike an appropriate balance between the values of autonomy and well-being. This implies that the disadvantage incurred by persons with impaired DMC due to being denied the right to make their own treatment decisions should be set off against the disadvantage these persons may incur due to possible harmful consequences of their decisions.

In cases where a person’s DMC remains below the threshold of competence despite the provision of support, the person may not be able to determine which of the available treatment alternatives is in keeping with her life plans and will promote her well-being. For example, persons who experience a severe psychotic episode may remain unable, despite the provision of support, to understand that hospital admission can promote their medical and social well-being and prevent their life plans from being disrupted. If healthcare professionals were to abide by their current preferences under such conditions, these persons’ interests in their well-being and their ability to live a life according to their own plan would in many cases be set back. Under the assumed circumstances, these persons would thus incur a greater overall disadvantage if their current preferences were not overridden. Consequently, under the circumstances, these persons do not incur an overall disadvantage by being denied the right to make the treatment decision.

Moreover, many proponents of the competence model hold that competence is risk relative (Drane, 1985; Buchanan and Brock, 1990; Grisso and Appelbaum, 1998; Kim, 2010). Risk relativity says that the less favorable the ratio of expected benefits and burdens of a treatment option as compared to the treatment alternatives, the higher the threshold of DMC that is required to be deemed competent to choose the option. This implies that persons are less likely to be found incompetent to make a treatment decision if the stakes in making the decision are relatively low. In relatively low-risk scenarios, persons with mild to moderate impairment of DMC will thus typically be allowed to make their own treatment decisions on the functional approach.

A Proposal for a Decision Procedure

Admittedly, the competence model faces difficulties even when it is based on the functional approach. A serious problem is that substitute decisions are often inaccurate. A review of empirical studies showed that substitute decision-makers predict the treatment preferences of patients accurately only in 68 percent of cases (Shalowitz, Garrett-Mayer, and Wendler, 2006). It might accordingly be thought that in one-third of cases, the frustration of the person’s interest in making his or her own decision is not compensated by any benefit yielded by the protection of autonomy and well-being.

But we must be careful here. First of all, most studies included in the review used hypothetical scenarios, and it is to be expected that substitute decision-makers will achieve higher accuracy in real-life situations (Kim, 2010). Moreover, the disadvantage incurred due to an inaccurate substitute decision should be weighed against the disadvantage that these persons would incur if healthcare professionals were to abide by their treatment preferences in the situation where their DMC is substantially impaired; and it would seem that their autonomy and well-being will often be negatively affected to a much greater extent in the latter case. That means that the disadvantage incurred due to an inaccurate substitute decision will typically not be a disadvantage in comparison to the available alternatives.

That said, it is clearly not for us to make the assessment of the relative weight of benefits and burdens attached to substitute decision-making. The maxim “nothing about us, without us” is a case in point. We propose that this assessment should be made based on a democratic decision-making process in which all persons with mental disorders are represented, whether directly or indirectly.

Setting up such a process will certainly not be easy. At the moment, we can only make a cautious estimation of the outcome of the process based on the available evidence. Looking at the Committee’s General Comment on article 12, one could be inclined to think that persons with mental disorders will see substitute decision-making as inevitably worse than any possible alternative. But this is too quick. We applaud the fact that persons with mental disorders were consulted during the drafting process of the CRPD. However, there are clear indications that the persons involved in drafting the CRPD and the General Comment on article 12 only represent a small minority of persons with mental disorders (Freeman et al., 2015; Appelbaum, 2016). In our view, a more representative sample of persons with mental disorders should have been included.

Contrary to what one would expect based on the Committee’s General Comment on article 12, the empirical evidence suggests that most persons with mental disorders endorse substitute decision-making. One large study, for example, showed that 83 percent of persons whose treatment preferences were overridden when their DMC was substantially impaired approved of the substitute treatment decision after they regained DMC (Owen et al., 2009). Various studies on patients’ attitudes toward involuntary hospitalization and involuntary treatment give a comparable, though less positive picture. A large international prospective study found that an average of 63 percent of patients admitted involuntarily approved of the admission after a period of 3 months (Priebe et al., 2010). Other studies confirm this picture (Schwartz, Vingiano, and Perez, 1988; Katsakou and Priebe, 2006; O’Donoghue et al., 2010; Katsakou et al., 2012).

The results of the studies on involuntary hospitalization and treatment should be treated with caution in this context. First, these studies report cases in which hospitalization or treatment was in conflict with the person’s actual preferences. Considering that a substantial share of substitute decisions does not override patients’ actual preferences, the endorsement rates for substitute decisions will expectedly be higher than these studies suggest (Burch, 2017). Second, these studies do not report on whether the people who were involuntarily hospitalized or treated were competent or incompetent with regard respectively to hospitalization and treatment. Based on other empirical studies (Okai et al., 2007; Owen et al., 2008), it can be expected that a substantial part of the sample of the studies was competent in either of these respects.

Based on the empirical evidence, we surmise that the proposed democratic decision procedure will result in the judgment that the benefits of substitute decision-making outweigh its burdens. That means that, on balance, persons with mental disorders do not consider being denied the right to make a treatment decision based on impaired DMC as a relative disadvantage. Should future research prove otherwise, we will modify our conclusions accordingly.

An important qualification must be made here. The empirical evidence suggests that a sizeable minority of persons with mental disorders will make a negative assessment of the benefits and burdens of substitute decision-making. An accommodation must be made for this group, and advance directives can be useful here. Advance directives are written documents that enable persons to state their treatment preferences for a future mental health crisis. By completing such a document, one can thus forestall decisions by substitute decision-makers. We suggest that any jurisdiction that allows for substitute decision-making should also have provisions that make advance directives legally binding (Scholten et al., 2019).

To summarize, in this section we have argued that a competence model based on the functional approach does not deny persons the right to make a treatment decision based on their membership of a salient social group and that persons who are found incompetent on the functional approach do not incur a relative disadvantage by being denied the right to make a treatment decision. On the proposed account of discrimination, it follows that it is not discriminatory to deny persons the right to make a treatment decision on this basis.

VII. CONCLUSION

It is a principle aim of the CRPD to counteract discrimination of persons with mental disorders. The Committee and various disability scholars see the abolition of the competence model and the practice of substitute decision-making as a necessary means to that end. We have argued that status- and outcome-based versions of the competence model are indeed discriminatory, but that a competence model that combines supported decision-making with a functional assessment of competence does not discriminate against persons with mental disorders.

The upshot of our argument is that current regimes of substitute decision-making are in urgent need of reform. Only a few countries have provisions for supported decision-making and even fewer countries make the necessary resources available. Guardianship laws of many countries are not in accordance with the functional approach, inasmuch as they deny persons the right to make treatment decisions based on the presence of a mental disorder, specify a diagnostic threshold for incompetence, or allow for plenary and permanent rather than time- and task-indexed forms of substitute decision-making. Even in jurisdictions where the laws are in accord with the functional approach, healthcare professionals often fail to apply the appropriate criteria for competence. Discrimination against persons with mental disorders in the informed consent process is a global reality, and there is still a long way to go in the process of eliminating this form of discrimination.

If our argument is correct, the key to eliminating discrimination of persons with mental disorders in the informed consent process does not lie in the abolition of all substitute decision-making arrangements, but in the radical reform of existing guardianship laws and the education of healthcare professionals regarding methods for supported decision-making and nondiscriminatory criteria for competence assessment.