Surrogate decision making for unrepresented patients: Proposing a harm reduction interpretation of the best interest standard

Introduction

R.G. is a 74-year-old male with a diagnosis of schizophrenia that has not been treated for many years, who has been diagnosed with stage 3 rectal and colon carcinoma. He is refusing treatment, has been declared to lack decisional capacity, and has no family or friends to help make decisions for him. Patient R.G. is part of a growing population of “unrepresented” individuals in need of medical care, who are deemed to lack decision making capacity, do not have a family or friends to act as a surrogate decision maker, and have not completed an advance directive regarding their care preferences. ¹ It is estimated that in the United States (US), unrepresented patients currently make up 3–4% of long-term care residents, ² 16% of patients admitted to the ICU, ³ and 5% of ICU deaths. ⁴ The majority of them are male, over 65 years of age, and have multiple chronic conditions.^3,4 Many of them are homeless and/or are diagnosed with a mental disability or mental disorder.^3,5 Of unrepresented patients admitted to the ICU, the majority will not regain capacity during their stay in the hospital. ⁴ When asked, three-quarters of physicians report having made a medical decision for an unrepresented patient within the past month. ⁶ While the number of unrepresented patients in need of medical care is increasing, ³ there has been a lack of public attention to the needs of this vulnerable population.^7,8

Inevitably, additional risks arise in medical settings when patients have no voice and no advocate to speak up for them; these risks involve overtreatment, undertreatment, and treatment that does not reflect their values or promote their wellbeing.^9–11 Data indicate that unrepresented patients are rarely sent to hospice even if they are eligible, and regrettably, that surrogates are very often found when more thorough searches are done by health-care teams. ¹² In response to concerns related to this population, many have raised the question of who ought to act as the surrogate decision maker for unrepresented patients. While the vast majority of decision making for these patients is done by physicians, ⁹ concerns have been raised about the lack of oversight that exists for such decisions, and the role that conflicts of interest and biases might play in physicians’ decision making.^13,14 Others defend the ability of physicians to decide what is best for those who are unrepresented, maintaining that this context is not significantly different from others in which we allow physicians to act as decision makers for patients.^5,9

In 2005, England and Wales responded to growing concerns surrounding the treatment of unrepresented patients with the Mental Capacity Act, which mandates that Independent Mental Capacity Advocates are appointed to represent and support the interests of those lacking capacity and a surrogate. ¹³ In most institutions in the US, however, no formal policy for supporting those who are unrepresented exists. This has led many providers, institutions, and districts to develop their own ad hoc procedures to deal with these cases. ¹⁵ Some of these responses include the creation of “health fiduciaries,” trained professionals who act as surrogate decision makers ³ ; the development of algorithms to predict unrepresented patients’ preferences ¹⁶ ; a state-wide ombudsman office to oversee such decisions^11,17; and The Wishard Volunteer Advocates Program, which trains volunteers to support unrepresented older adults in their medical decision making, and has led to a significant decrease in ER visits and a significant increase in Medicaid reimbursements for the hospital. ¹⁸

While the bioethical literature contains several thoughtful responses to the question of who ought to decide for the unrepresented, there has been no sustained discussion of how to decide for these patients. In this paper, we address the question of how to make decisions for unrepresented patients by formulating a new interpretation of an established surrogate decision making principle that, we argue, would be better suited for this population. In particular, we aim to improve decision making for unrepresented patients who are deemed to lack capacity as a result of a mental disorder, but are still able to express preferences about their care. We begin with an overview of the standard approaches to surrogate decision making and then show how applications of these models often come up short in situations involving unrepresented patients dealing with sustained mental distress. We propose the harm reduction interpretation of the best interest standard as an approach to decision making for this unique population, and argue that, if adopted, this approach would better support the well-being of these vulnerable patients.

Standard models of surrogate decision making

In the case of R.G., for which one of the authors (NG) offered clinical ethics support, the team decided that the best course of action was to seek the appointment of a legal guardian, a process which, in the U.S., can take from several weeks to a month. In the interim, the team would have to settle on the goals of care and possibly begin treatment. Two standard models of surrogate decision making—the substituted judgment standard and the best interest standard—are typically used to guide the surrogate decision making process for patients. We will show, however, that standard interpretations of these models are not always aligned with the goal of promoting the well-being of unrepresented patients diagnosed with a mental disorder, who are in important ways a unique population.

Surrogate decisions apply in circumstances when patients are not able to make medical decisions for themselves. This can happen when a patient is not competent or when they lack decisional capacity. It is important to distinguish between competence and capacity, as they are two related but distinct concepts. ¹⁹ Decisional capacity is judged by physicians; these assessments are task and context specific, and are usually performed to determine whether the patient is able to accept or refuse medical treatment. Competence refers to the global ability of the patient to make any decisions, including medical decisions. An assessment of global incompetence is made by a judge. Incompetent patients are usually appointed a legal guardian. Unrepresented patients usually have not been declared incompetent by the legal system and they do not have a guardian. All unrepresented patients, however, have been determined by their medical team to lack capacity to make medical decisions.

Some patients who lose capacity will have already specified their care preferences and values, and these specifications can offer guidance for medical decision making. These preferences and values can come in the form of advance directives (e.g. a signed DNR/DNI (do not resuscitate/do not intubate), an organ donor card), a living will, or wishes that the patient expressed in the past. Unrepresented patients most often do not have any type of advance directive. ⁵

The substituted judgment principle

When the patient has no advance directives, appropriate surrogates are called upon to use the substituted judgment principle in order to make decisions on behalf of the patient. Based on this principle, the surrogate decision maker should take into account previously expressed preferences or a history of values articulated by the patient and apply them to the current medical situation. ²⁰ In addition, the substituted judgment standard allows surrogates to make decisions for the patient based on general values expressed by the individual (e.g., about quality of life over length of life) and use those to determine whether a patient would have wanted a medical intervention that may be indicated (e.g., artificial nutrition and hydration). This standard is applicable in situations when individuals temporarily lose their ability to make decisions due to a treatable medical conditions, including delirium, fever, or stroke, and in situations where patients become permanently incapacitated due to a currently irreversible disease such as Alzheimer’s. What is important to note about this principle is that it is designed to respect the currently incapacitated patient’s previously expressed autonomous preferences and values, although some commentators have questioned whether the principle is effective in doing so. ²¹

This emphasis on preserving autonomy through surrogate decision making makes the substituted judgment standard difficult to apply to unrepresented patients with long-term diagnosis of mental disorders. In part, this is because they do not have friends and family to provide a holistic picture of their values and preferences; what is known about these patients’ wishes is based solely on their interactions with medical professionals. While some unrepresented patients may have a history of such interactions, and in some cases information can be gathered from the patient’s primary physician, this is not the case for the majority of unrepresented patients. ⁴ Moreover, the application of the substituted judgment standard presumes that the preferences and values taken into account by the surrogate are those the patient expressed at a time when they had decisional capacity. However, unrepresented patients, especially those who have been diagnosed with a mental illness or disability, are often deemed to lack decisional capacity in the long term and their preferences and values are often discounted because they are not judged to be autonomous. For example, R.G., who has had a long-term diagnosis of schizophrenia, has had a history of interactions with medical professionals, but in all of them he was determined to lack decision making capacity. Although R.G. consistently has refused treatment, his expressed wishes were often not taken into account by the health-care team. Thus, most of his interactions with the health-care team required some degree of compulsory treatment or treatment not in line with his expressed wishes.

A further problem with the substituted judgment standard is the distance between what patients would have chosen and the judgments made by surrogate decision makers, partly as a result of the common tendency to project one’s own preferences onto others.^22,23 Physicians, in particular, have been found to be poorer than family members at predicting the care that a patient would have wanted. ²⁴ Within the context of intensive care, physician’s assessments of patient’s preferences have been found to correlate strongly with their personal preferences regarding intensive care, but very little with their patient’s actual preferences. ²⁵ When asked about their recent experiences with surrogate decision making, the majority of physicians (73%) report that patient preferences ought to be the most important factor within surrogate decision making, although only 29% report that patient preference was the most important factor in the most recent surrogate decision they made for a patient. ²⁶ This evidence indicates that in practice, implementation of the substituted judgment standard often fails to capture the preferences of the incapacitated patient, particularly when there are no family members present to recall the patient’s previously expressed wishes. Because physicians are most likely to act as surrogate decision makers in cases involving unrepresented patients, the likelihood that their decisions may not reflect the wishes of the patient is high. ⁹

The best interest standard

In cases when very little or nothing is known about the patient’s preferences or values, surrogate decision makers tend to apply the best interest standard in order to set goals of care for a patient. The best interest standard guides surrogates to weigh the risks and benefits of each relevant medical intervention, including the risks and benefits of not pursuing treatment, and make a decision that involves the most favorable risk and benefit ratio for the patient. ²⁷ Surrogates utilizing the best interest standard are called upon to take into account the quality of life that can be anticipated as a result of each, or no, intervention, as part of the process of weighing risks and benefits. ²⁸ While in principle, the best interest standard is an ideal model of surrogate decision making, in practice, and particularly with unrepresented patients, the application of the standard often leads health-care professionals into a false dichotomy between overtreatment and undertreatment. This occurs when too much treatment or too little treatment interferes with the aim of patients achieving the best quality of life available to them. Practitioners tend to focus either on curative treatment, which may lead to overtreatment, or comfort/palliative care, which can lead to undertreatment, ignoring options in the middle that involve partial treatment or palliative care combined with treatment.^12,29,30

When making the decision to pursue treatment, the best interest standard may lead to a focus on curative treatment, which can lead to overtreatment, for a number of reasons. For one, judgments about acceptable quality of life are subjective and difficult to make without knowing the values of the patient, ³¹ so healthcare professionals often interpret the best interest standard as erring on the side of life-preservation or as mandating treatment, even in cases where medical intervention might lower a patient’s overall quality of life. This general tendency toward curative measures has been increasingly recognized across a number of different clinical contexts, including those involving unrepresented patients.^32,33 Interestingly, patients are often less interested in potentially curative treatment than surrogates are. In a meta-analysis of the accuracy of surrogate decision making, the most common error involved surrogates choosing treatments that patients did not want. ³⁴ This appears to be particularly common in older populations, both when physicians or family members are acting as surrogate decision makers.^35,36

An additional reason why the application of the best interest standard may lead to overtreatment is the common interpretation of the best interest standard as requiring the implementation of standard of care. ³⁷ When clinicians or clinical ethics committees are comparing available treatment alternatives for unrepresented patients, their deliberations often involve a risk and benefit analysis which compares distinct standard of care recommendations for each treatment alternative, i.e., the surgical standard of care for a particular condition versus the standard of care for the medical management of that same condition. The boundaries of the standard of care, however, are often set by the perceived threat of litigation which compels physicians to err on the side of overtreatment. ³⁸ In addition to being perceived as legally required, the implementation of the standard of care is also interpreted by physicians as being ethically required. This increases the risk that unrepresented patients will experience more treatment than they might want or have wanted, and may partly explain why these patients are not referred to hospice care as often as they should be. ¹²

Unrepresented patients that are deemed to lack capacity due to a long-term diagnosis of a mental disorder, like R.G., often will have had experiences of treatment given without their consent or against their wishes. Research related to forced treatment within psychiatric settings suggests that harms resulting from treating someone against their will can be severe, as individuals can experience compulsory treatment as traumatic, and some develop symptoms of post-traumatic stress disorder in its aftermath.^28,39,40 Unsurprisingly, individual’s ratings of perceived coercion have been found to negatively correlate with positive responses to health-care professionals,^41–43 and when asked about their experiences after the fact, many maintain that the necessity or beneficence that was thought to justify the use of forced treatment such as seclusion or restraints was not in fact present. ⁴⁴ While it is not clear that the harms of forced treatment will always extend from psychiatric settings to medical ones, it is clear that caution should be used in treating unrepresented patients against their expressed wishes, particularly those who may have had negative experiences with compulsory treatment in the past.

There is also a risk of unrepresented patients receiving too little treatment when the best interest standard is followed. This represents the significant gap between curative and palliative measures in health-care that many practitioners are calling for increased attention to.^12,29,45 With unrepresented patients, undertreatment can occur when the standard of care seems too difficult to implement; this is likely in cases when patients are unlikely to be compliant with treatment or do not have a support system that can help them recover from a procedure. This is often the case for unrepresented patients with long-term diagnosis of mental disorders, who have no family or friends to help them recover, often have many comorbidities, and sometimes lack stable housing.^3,5 In such cases, applying the best interest standard can lead to undertreatment. This occurs when a surrogate decision maker judges that treatment is overly burdensome for the patient or impractical, and opts for solely palliative care instead. Evidence suggests that physicians tend to underestimate the quality of life of many of their patients, particularly in elderly patients with chronic illness, which may increase their likelihood of withdrawing treatment. ⁴⁶ In one sample of unrepresented patients admitted to the ICU, treatment was eventually withdrawn for 16% of patients. ⁴ In the same sample, physicians considered writing DNR orders for 36% of the unrepresented patients in the ICU and quality of life judgments were central to their decisions in 61% of these patients. ⁴

Taken together, this suggests that applying the best interest standard in cases of surrogate decision making for unrepresented patient can lead to both overtreatment and undertreatment. On the one hand, associations between the best interest standard, life-preserving treatments, and the standard of care could lead healthcare professionals to opt for treatment even when it may not be the best option for a patient’s quality of life or in line with what the patient would have chosen. On the other hand, if the requirements of the standard of care are impossible or difficult to implement, unrepresented patients risk missing out on treatment altogether.

Harm reduction and the best interest standard

To better accommodate the needs of unrepresented patients living with diagnosis of mental disorders in the US, we propose an interpretation of the best interest standard that that takes into account the complexity of capacity judgments, the risks of forced treatment, as well as the potential benefits of treatment. Our approach extends the model of harm reduction used in public health initiatives to setting goals of care for unrepresented patients who are able to express treatment preferences.

Harm reduction practices are most commonly found in public health initiatives involving marginalized, and often stigmatized, groups, ⁴⁷ and often, at least implicitly, recognize the ways in which discrimination can contribute to an unjust distribution of harms in these groups.^48,49 Common practices that fall under the domain of harm reduction include “opioid replacement therapy (e.g., methadone maintenance programs), needle and syringe distribution, emphasizing routes other than injection for drug administration, counselling, naloxone distribution, sexually transmitted infection services, testing, wound care, vaccinations, social assistance, reducing injury or violence, and peer support.” ⁴⁸ At the heart of the approach is the aim of minimizing harm and reducing the risks that a given population might face, rather than exclusively using resources or practices that push toward what might be thought of as the ideal goal (e.g., abstinence from substance use).

Four basic principles of harm reduction have been characterized by Marlatt and Witkiewitz ⁴⁹ : (1) harm reduction offers a public health alternative to moral and disease models of substance use; (2) harm reduction recognizes that abstinence is ideal but alternatives that minimize harm are important; (3) harm reduction is a grass roots approach based on substance use advocacy rather than top down policies developed by professionals; and, (4) harm reduction promotes “low threshold access” to services. We recommend that clinicians, ethics committees, and guardians who are acting as surrogate decision makers for unrepresented patients with mental health diagnosis could use the harm reduction model to guide their application of the best interest standard. The model we propose consists of three components: taking into account expressed preferences, focusing on harm reduction as well as beneficence, and promoting trust. ⁵⁰

The first two aspects of the harm reduction model for surrogate decision making stand in contrast to two ideals that have shaped the standard approaches to surrogate decision making discussed above. The first ideal is the emphasis on previously expressed autonomous preferences and values when applying the substituted judgment standard. Approaches to surrogate decision making that rely on the expressed preferences of an autonomous agent might work for individuals who lose capacity temporarily, but do not work well for unrepresented individuals with long-term diagnosis of mental disorder, for whom there is rarely a recorded history of values and preferences that are considered autonomously expressed. The requirement that only autonomous preferences should be used for substituted judgments makes little sense for patients who might have never been considered capacitous and able to establish a baseline of autonomous decision making. For this population, we should consider the expressed preferences of individuals, regardless of their capacity status, and take them into account when making medical decisions about them. This does not mean that refusals of medical treatment by unrepresented patients with diagnosis of mental disorders should always be respected, but it does mean that if a patient consistently refuses medical treatment, imposing treatment should be considered a significant harm that would have to be outweighed by a significant benefit in order to be justified.

The second ideal which can ultimately compromise the care of unrepresented patients is the ideal of the best possible medical outcome or the curative treatment ideal. In many cases, providing a patient who is very ill with treatment in line with the standard of care requires the cooperation of the patient or the reliance on social support, since care beyond the intervention is often necessary for positive outcomes (e.g., taking medicines regularly, attending required outpatient appointments). However, in many cases involving long-term compulsory treatment with unrepresented patients diagnosed with a mental disorder, significant follow-up and adherence is not possible, because the patient has no support network to assist in aftercare and may be unwilling to return to the place where he or she underwent treatment. In such cases, a better approach might be to offer partial treatment. The harm reduction approach would prevent healthcare professionals from falling into the false dilemma of either having to treat according to the standard of care or to refrain from treatment all together. This would require a shift away from focusing on beneficence toward a focus on reducing harm. Although these two principles are not mutually exclusive and one can often conceive of harm reduction as beneficence, ⁵¹ the latter principle is more often taken to indicate that a physician has the obligation to pursue some maximally (medically) beneficial course of treatment that might not be achievable in the context of all unrepresented patients with long-term psychiatric diagnosis. ⁵²

Finally, another important aspect of the harm reduction approach to surrogate decision making is fostering the doctor–patient relationship and promoting trust. Although some unrepresented patients with diagnosis of mental disorders are brought to the hospital because of serious illness and have not made the decision to seek care, they should be treated by healthcare professionals in a way that would encourage them to continue seeking medical care. This involves treating the patient with non-judgmental regard, connecting them to additional support services, and maintaining a low threshold for care, thereby acting in a way that promotes trust within the relationship. As was described earlier, forced treatment can cause a great deal of distress and patients who have experienced forced treatment in the past might avoid seeking medical care as a result of such experiences, which can adversely impact their health. Therefore, meeting patients where they are, listening to their expressed preferences, and minimizing forced treatment might reduce overall harm to patient’s health by enhancing trust in health-care professionals.

For patients like R.G., a harm reduction interpretation of the best interest standard could provide a useful decision making framework for those involved in his care, as least until a guardian is appointed. Taking his expressed preferences into account, despite his negative capacity assessment, would alert the team to his refusal of care, and how it may be situated for him in a history of treatment that he did not agree to. Focusing on a reduction of harm may lead the team to consider just how burdensome a compulsory long-term course of chemotherapy could be for R.G., and how painful and traumatic such treatment could be for a patient who is unable to understand the reasons for treatment or to endorse the potential long-term benefits of a difficult treatment course. Treatment through surgery, which requires less time and interaction with the health-care system, may be more plausible. Importantly, given R.G.’s long and difficult history of interactions with health-care professionals, working toward building a relationship of trust with him is crucial. Acknowledging his refusal of care, but not closing the door on the possibility of treatment in the future, allwhile treating R.G. with non-judgmental regard, could go a long way toward improving his relationship with the care team and the system of health-care more generally.

Conclusion

Traditional interpretations of surrogate decision making are not well suited to unrepresented patients who have been diagnosed with a mental disorder and are able to express preferences for care, a population that is presenting at hospitals in the U.S. more and more often. Rather than relying on the substituted judgement standard, which requires previously expressed autonomous preferences to be known, or the best interest standard, which can lead to overtreatment or undertreatment, we suggest that a harm reduction interpretation of the best interest standard is more appropriate for decision making for the unrepresented patient population with psychiatric diagnosis. This model consists of three components: taking into account the expressed preferences of the patient, focusing on the reduction of harm and not just the delivery of clinical benefits, and promoting trust.