NEUROREHABILITATION is an honorable and stressful endeavor. Persons with brain injuries present many clinical and ethical challenges. The clinical science of brain injury rehabilitation has gained considerable momentum over the past 20 years. Nonetheless, rehabilitation professionals continue to face a number of ethical dilemmas, many of which have remained essentially unchanged despite the progress in clinical neurorehabilitation. For the most part, ethical problem solving and decision making has remained the province of biomedical ethicists and neurorehabilitation practitioners trained in biomedical ethics. A central objective of the current column is to raise sensitivity to ethical issues faced by neurorehabilitation clinicians on a day-to-day basis. Consequently, in the following discussion, the thorny issue of beneficence versus respect for autonomy is addressed. Balancing respect for autonomy and beneficence is one of the most common ethical problems rehabilitation professionals face on a daily basis.
Rehabilitation professionals can use various ethical models or theoretical approaches to guide ethical thought and decision making. One of the most respected and applied models is principlism described by Beauchamp and Childress.1 Principlism identifies 4 basic, but complicated, principles that guide all ethical analysis and decision making. These principles include beneficence, nonmaleficence, respect for autonomy, and justice. By using these principles, clinicians can systematically examine the ethical implications of clinical decisions and interventions.
As mentioned, one of the most common ethical dilemmas rehabilitation professionals must resolve is the tension between beneficence and respect for autonomy. Beneficence refers to a basic obligation to help others, but more importantly, beneficence requires an obligation to "further [another's] important and legitimate interests."1(p194) Beneficence is more than simply doing "good" in the generic sense. One must consider what is "good" in a particular person's context. Sometimes, clinicians believe they are acting in a person's "best interest," but acting in a person's "presumed best interest" may not necessarily be a beneficent act. For instance, persons who have sustained brain injuries are routinely enrolled in therapies that they do not specifically request or for that matter want. When a person with a brain injury refuses therapy, we typically encourage him or her to participate because we assume, either correctly or incorrectly, that he or she would benefit from the therapy. In fact, we often assume that such persons do not know what is "best" simply because they are "brain injured." Ironically, in some cases, refusing therapy is interpreted as prima facie evidence that the treatment being refused is really needed.
There are a number of other scenarios in which rehabilitation professionals provide therapies, participate in disposition planning, and make treatment recommendations on the basis of presumed beneficence. In such situations, clinicians should have a clear understanding of what beneficence entails. Beachaump and Childress1 cogently argued that beneficence is the "promotion of health, as defined in part by the patients own values."(p196) Consequently, when we presume that we are promoting the welfare of persons who have sustained brain injuries, we need to be certain that our perception of doing "good" is related in some direct way to the "patients' own values."
For instance, if a person finds cognitive rehabilitation therapy unrewarding and unproductive, should we insist she or he continue in therapy or should we propose an alternative intervention? If we insist that he or she continues therapy, do we do so because we have scientific evidence that cognitive rehabilitation will ultimately prove "beneficial?" If such evidence exists, beneficence is considerably more tangible. If no scientifically credible evidence documenting the efficacy of cognitive rehabilitation actually exists, then we have a different problem. In other words, in some cases, beneficence may be presumed when we simply want patients to comply with traditional treatment protocols without evidence of specific benefits. In such cases, presumed beneficence may conflict with the need to respect the person's right to make decisions that govern his/her life, including treatment decisions.
Respect for autonomy is a core ethical construct and is often equated with a number of other concepts, such as self-governance, free will, and choice.2 In practical terms, respect for autonomy implies that we value an individual's right to determine his or her own destiny. Sometimes, ethicists draw a distinction between autonomous persons and autonomous choices or actions. Although this distinction is theoretically important, rehabilitation professionals should ultimately be most interested in whether persons with neurological injuries act "intentionally, with understanding and without coercion."1(p69) One can argue that it is often difficult to meet this standard in general medical settings, let alone in neurorehabilitation, where cognitive and behavioral functioning is almost always compromised. Nonetheless, the issue is not one of ideal or absolute autonomy, but whether a person's decision is "substantially autonomous."1(p69) As expected, determining whether someone can act in a "substantially autonomous" manner can be a difficult clinical decision.
A major factor in establishing whether a person can behave in a "substantially autonomous" manner is the extent to which she or he is capable or competent. In the most liberal sense, capacity refers to a person's ability to simply make a choice. However, standards for capacity vary. The strictest test of capacity requires that a person with a neurological injury demonstrate complete understanding of the situation in question. He or she would make a decision on the basis of rational reasons, and of course, he or she would fully understand and weigh the potential outcomes.3 Obviously, this test is extremely strict and rarely satisfied. A more liberal, but still defensible test of capacity requires that persons make decisions on the basis of rational reasons. In any case, assessment of capacity is a clinical process that enables rehabilitation professionals to partially objectify the ability of persons with brain injuries to function autonomously. Unfortunately, determining capacity is not always a straightforward process. Capacity determination can be complicated by neurocognitve deficits related to brain injury, as well as the clinician's understanding of what constitutes a "rational reason." Nonetheless, clinicians cannot presume that brain-injured persons' capacity is diminished or impaired for all decisions simply because they are "brain injured."
Categorical challenges to autonomy (eg, "of course, he is brain injured") are common depending upon the situation and circumstances. For instance, refusing medication, therapy, involvement in outpatient programs, and the like are often perceived to be the product of an "injured brain" rather than a rational choice or preference. The most common beneficent act rehabilitation professionals engage in is prevention of harm. Safety is the Holy Grail of rehabilitation, and rehabilitation professionals often constrain autonomy to prevent real or perceived harm. We constrain autonomy to "protect" persons with brain injury from themselves. For instance, we may prevent persons who have sustained brain injuries from driving, or we may decline to support their return to work or school on the basis of the belief that they may have an accident or fail to successfully return to work or academics.
In an earlier article,4 I argued that it was intrinsically unethical to constrain autonomy in the name of beneficence if there was no convincing empirical evidence that the constraint was scientifically warranted. In other words, if we are going to act paternalistically/maternalistically, we should be as certain as possible that our predictions are valid. Quite simply, if our predictions are erroneous, we are not acting beneficently. Unfortunately, errors in clinical judgment are quite common. Nonetheless, clinicians appear to be quite confident when making clinical predictions, even when they are wrong.5 As clinicians, we may believe that "clinical experience" can prevent errors in clinical judgment. Unfortunately, research on clinical experience has not confirmed this belief. Numerous studies have found that experience does not inoculate clinicians against prognostic errors. Experienced clinicians make the same errors in prognosis and diagnosis as inexperienced clinicians.6,7 Ironically, even if we recognize the problems associated with clinical judgment, we are frequently asked to make judgments of risk and harm on the basis of clinical information with no actuarial data. In fact, in neurorehabilitation, we often do not have empirical data to assist in predicting risk and harm.
Rehabilitation professionals can almost always assume that there will be tension between beneficence and respect for autonomy on a daily basis. The amount of tension depends on the specific circumstances. When capacity is intact, persons with brain injuries can act autonomously. In some cases, persons with brain injuries do not agree with our therapy goals or recommendations. Nonetheless, we are ethically bound to let capable persons make decisions we would not make, even if those decisions appear detrimental. Capable persons can refuse all kinds of medical interventions including life-sustaining treatment, so why would a brain injury preclude treatment refusal or refusal of other autonomy-constraining behavior? When a person's capacity is compromised and the medical consequences are significant, paternalism/maternalism (beneficence) may be justified; but if a person is capable and the medical consequences are unknown or undetermined, ethical issues become more prominent.2 In such circumstances, following a specific model of assessment may make clinical decision making less stressful, erroneous, and ethically problematic.
A functional model of assessment includes analysis of the particular situation with respect to the patient's capacity to make relevant decisions. When assessing capacity, clinicians must discuss preferences, so simultaneous consideration must be given to establishing the patient's preferences. Preferences must be detailed and logically connected to the issues at hand. For example, if a nursing home placement is being considered, preferences toward a nursing home placement must be elicited. Once patient capacity and preferences have been established, a third step is analysis of the harm or risk involved. Risk and harm can be difficult to quantify, so clinicians must consult the literature or their own data to determine whether a mathematical risk estimate can be obtained. If risk and harm cannot be quantified, then clinicians must determine whether they have sufficient clinical information to merit a beneficent approach. When doing so, we must be aware that simply worrying about risk, or being concerned about harm, is not sufficient to overrule a capable patient's decision.
Fourth, patients' preferences merit respect, but sometimes patients can be helped to reconsider decisions, especially if clinical concerns are clearly delineated and objectified, rather than presented in a global or coercive manner. In addition, unless prohibited from doing so, family members should be consulted. In many cases, family members may be able to communicate risks, especially unquantified risks, in a more effective manner than clinicians do. Finally, all decisions resulting in overriding patients' wishes should cause considerable consternation and self-reflection. The basis for doing so should be explicit and available for peer review.
In conclusion, ethicists and clinicians have long struggled with the delicate balance between beneficence and autonomy. Some time ago, Banja8 urged clinicians to reflect on the "admittedly vague, complex, frequently unscientific, but morally compelling issue of risk assessment."(p71) Unfortunately, questions regarding the necessary standards of evidence and degrees of clinician confidence when acting beneficently remain unanswered. In general, doing "good" is much more complicated than most clinicians appreciate. As rehabilitation professionals, we are ethically obliged to provide a scientifically sound rational when constraining autonomy.
REFERENCES