Groundhog Day for Medical Artificial Intelligence

Following a boom in investment and overinflated expectations in the 1980s, artificial intelligence entered a period of retrenchment known as the “AI winter.” With advances in the field of machine learning and the availability of large datasets for training various types of artificial neural networks, AI is in another cycle of halcyon days. Although medicine is particularly recalcitrant to change, applications of AI in health care have professionals in fields like radiology worried about the future of their careers and have the public tittering about the prospect of soulless machines making life‐and‐death decisions. Medicine thus appears to be at an inflection point—a kind of Groundhog Day on which either AI will bring a springtime of improved diagnostic and predictive practices or the shadow of public and professional fear will lead to six more metaphorical weeks of winter in medical AI.     

Conflating Capacity & Authority: Why We're Asking the Wrong Question in the Adolescent Decision‐Making Debate

Whether adolescents should be allowed to make their own medical decisions has been a topic of discussion in bioethics for at least two decades now. Are adolescents sufficiently capacitated to make their own medical decisions? Is the mature‐minor doctrine, an uncommon legal exception to the rule of parental decision‐making authority, something we should expand or eliminate? Bioethicists have dealt with the curious liminality of adolescents—their being neither children nor adults—in a variety of ways. However, recently there has been a trend to rely heavily, and often exclusively, on emerging neuroscientific and psychological data to answer these questions. Using data from magnetic resonance imaging and functional MRI studies on the adolescent brain, authors have argued both that the adolescent brain isn't sufficiently mature to broadly confer capacity on this population and that the adolescent brain is sufficiently mature to assume adolescent capacity. Scholars then accept these data as sufficient for concluding that adolescents should or should not have decision‐making authority. Two critical mistakes are being made here. The first is the expectation that neuroscience or psychology is or will be able to answer all our questions about capacity. The second, and more concerning, mistake is the conflation of decision‐making capacity with decision‐making authority.      

A Good Death

A good death is hard to find. Family members tell us that loved ones die in the wrong place—the hospital—and do not receive high‐quality care at the end of life. This issue of the Hastings Center Report offers two articles from authors who strive to provide good end‐of‐life care and to prevent needless suffering. We agree with their goals, but we have substantial reservations about the approaches they recommend. Respect for the decisions of patients and their surrogates is a relatively new and still vulnerable aspect of medical care. For thousands of years, patients and surrogates had no say in medical decision‐making. Today, standards support shared decision‐making, but these articles both carve out exceptions to those standards, limiting the rights of patients and families in decisions about specific end‐of‐life treatments. As bioethics consultants in an acute care setting, we frequently confront conflicts similar to those described by Jeffrey Berger and by Ellen Robinson and colleagues. In such cases, our service emphasizes redoubled efforts at communication and mediation. Focusing on goals and values, rather than interventions, produces the best possible collaboration in health care decision‐making. Cases in which we would overturn a surrogate's recommendations regarding palliative sedation or do‐not‐resuscitate orders are rare and require careful processes and clear evidence that the surrogate's choice is contrary to patient values.     

Justifying Clinical Nudges

The shift away from paternalistic decision‐making and toward patient‐centered, shared decision‐making has stemmed from the recognition that in order to practice medicine ethically, health care professionals must take seriously the values and preferences of their patients. At the same time, there is growing recognition that minor and seemingly irrelevant features of how choices are presented can substantially influence the decisions people make. Behavioral economists have identified striking ways in which trivial differences in the presentation of options can powerfully and predictably affect people's choices. Choice‐affecting features of the decision environment that do not restrict the range of choices or significantly alter the incentives have come to be known as “nudges.” Although some have criticized conscious efforts to influence choice, we believe that clinical nudges may often be morally justified. The most straightforward justification for nudge interventions is that they help people bypass their cognitive limitations—for example, the tendency to choose the first option presented even when that option is not the best for them—thereby allowing people to make choices that best align with their rational preferences or deeply held values. However, we argue that this justification is problematic. We argue that, if physicians wish to use nudges to shape their patients’ choices, the justification for doing so must appeal to an ethical and professional standard, not to patients’ preferences. We demonstrate how a standard with which clinicians and bioethicists already are quite familiar—the best‐interest standard—offers a robust justification for the use of nudges.     

Actions and Uncertainty: How Prenatally Diagnosed Variants of Uncertain Significance Become Actionable

The development of genomic technologies has seemed almost magical. Excitement about it, both in medicine and among the public, stems from the belief that genomic techniques will illuminate the causes of health and disease, will lead to effective interventions for both rare and common genetic conditions, and will inform reproductive decision‐making. Novel diagnostic tools, however, are often deployed before targeted therapies are developed, tested, or available and before their psychosocial implications are explored. Newer technologies such as prenatal whole exome screening are seen as offering “decisional autonomy” to expectant parents, although such technologies identify information about genetic sequencing that may not have clear meaning. The “therapeutic gap” between the ability to conduct genetic sequencing and the ability to fully understand what the test results mean, much less what treatments to offer, leaves families with complex and unclear information they cannot act upon with confidence during pregnancy. In this essay, we will consider the psychosocial and ethical implications of such assumptions—and of the uncertain information produced by these technologies—for individuals and families and for societal aspects such as medical service usage and demographic inequities.     

The Different Moral Bases of Patient and Surrogate Decision‐Making

My topic is a problem with our practice of surrogate decision‐making in health care, namely, the problem of the surrogate who is not doing her job—the surrogate who cannot be reached or the surrogate who seems to refuse to understand or to be unable to understand the clinical situation. The analysis raises a question about the surrogate who simply disagrees with the medical team. One might think that such a surrogate is doing her job—the team just doesn't like how she is doing it. My analysis raises the question of whether (or perhaps when) she should be overridden. In approaching this problem, I focus not on the range of difficulties in practice but on the underlying moral conceptual issue. My concern will be to show that the moral values that underpin patient decision‐making are fundamentally different from those that underpin surrogate decision‐making. Identifying the distinctions will set parameters for any successful solution to the “Who should decide?” question. A patient has a specific kind of moral right to make her own medical decisions. A surrogate has no analogous moral right to decide for someone else. We want the surrogate to make the decision because we believe that she has a relevant epistemological advantage over anyone else on the scene. If and when she has no such advantage or if she refuses or is unable to use it, then there might not be sufficient reason to let her be the decision‐maker.     

You Can't Always Get (or Give) What You Want: Preferences and Their Limits

People who lack decision‐making capacity may be able to communicate preferences, which can and should inform surrogate decision‐making on their behalf. It is unclear whether making a further distinction about “capacity for preferences,” as Jason Wasserman and Mark Navin propose in this issue of the Hastings Center Report , would improve the process of surrogate decision‐making. Anyone who is regularly involved in surrogate decision‐making or who has worked to articulate decision‐making standards and processes can think of cases in which a patient's voice was ignored or his or her preferences and ability to communicate them overlooked. However, we can also think of cases in which it was unclear whether information provided by a patient, recently or in the past, should have been characterized as a preference relevant to a medical decision that the patient lacked the capacity to make. We should also recognize that clear preferences of a patient who lacks decision‐making capacity may conflict with the limits of another person or of a family's resources such that these preferences cannot guide a surrogate's decision. This patient will not get what he wants. A situation in which preferences are known and make sense but cannot be applied in a feasible way demonstrates the limits of the medical decision‐making framework concerning care that is not covered by medical insurance.     

Decision‐Making for an Incapacitated Pregnant Patient

Decisions about continuing or terminating a pregnancy touch on profound, individualized questions about bodily integrity, reproductive autonomy, deeply held values regarding one's capacity for parenthood, and, in the case of a high‐risk pregnancy, the risks one is willing to take to have a baby. So far as possible, reproductive decisions are made between a patient, in some cases her partner, and her medical provider. However, this standard framework cannot be applied if the patient lacks decision‐making capacity. In this essay, we discuss one such case that came before our clinical ethics team. We describe the challenges of respecting a patient's reproductive preferences when the patient cannot share what those preferences are, and we argue that decisions regarding reproductive health care should not be treated with exceptionalism. Rather, they should proceed under the normal processes of surrogate decision‐making, including the application of substituted judgment. This approach enables us to take the patient's values into account when considering the questions implicated in reproductive health care, just as we do for other kinds of health care decisions in which a patient's deeply held values are salient.     

Learning from Me Too

For the past year, many around the world have said, “Me Too.” Women, trans folks, and increasingly men have shared their experiences of sexual assault, harassment, and unhealthy power dynamics in their personal and professional relationships. These experiences were not a surprise to people in bioethics, especially as they relate to the control of women's bodies and behaviors that we have observed in medicine, law, and society. In The Hastings Center's work in reproductive contexts, I have learned about women whose reproductive decisions or choices are negatively affected by external forces—either by restrictive laws and policies, explicit counsel from their physicians, or implicit pressures placed on them by society. Any person with a uterus can be forced to take contraceptives if they want access to other medications that could harm a fetus, even when they are not sexually active with men. And some women of color are criminalized for miscarriages. How can those of us working in reproductive ethics be even more attentive to these experiences and to the impact of gendered power dynamics on reproductive decision‐making?     

Capacity for Preferences: Respecting Patients with Compromised Decision‐Making

When a patient lacks decision‐making capacity, then according to standard clinical ethics practice in the United States, the health care team should seek guidance from a surrogate decision‐maker, either previously selected by the patient or appointed by the courts. If there are no surrogates willing or able to exercise substituted judgment, then the team is to choose interventions that promote a patient's best interests. We argue that, even when there is input from a surrogate, patient preferences should be an additional source of guidance for decisions about patients who lack decision‐making capacity. Our proposal builds on other efforts to help patients who lack decision‐making capacity provide input into decisions about their care. For example, “supported,” “assisted,” or “guided” decision‐making models reflect a commitment to humanistic patient engagement and create a more supportive process for patients, families, and health care teams. But often, they are supportive processes for guiding a patient toward a decision that the surrogate or team believes to be in the patient's medical best interests. Another approach holds that taking seriously the preferences of such a patient can help surrogates develop a better account of what the patient's treatment choices would have been if the patient had retained decision‐making capacity; the surrogate then must try to integrate features of the patient's formerly rational self with the preferences of the patient's currently compromised self. Patients who lack decision‐making capacity are well served by these efforts to solicit and use their preferences to promote best interests or to craft would‐be autonomous patient images for use by surrogates. However, we go further: the moral reasons for valuing the preferences of patients without decision‐making capacity are not reducible to either best‐interests or (surrogate) autonomy considerations but can be grounded in the values of liberty and respect for persons. This has important consequences for treatment decisions involving these vulnerable patients.     

It's All Relative

In this issue of the Hastings Center Report, Daniel Brudney suggests that clinicians have an overly deferential attitude toward their patients’ surrogate decision‐makers that is rooted in a wrongful investment of moral authority. He maintains that surrogate decision‐makers have no moral right to decide for their loved ones and that their value in the decision‐making process is limited to their knowledge of their loved one's preferences. If operationalized, Brudney's framework would ease the way for clinicians to remove a surrogate who cannot provide information relevant to the patient's preferences and to resort to a paternalistic model of decision‐making. Brudney fails to consider that the value of the surrogate does not flow from the surrogate, but rather from the patient's moral claim to have decisions made for him or her by a loved one. This claim recognizes that surrogates have intrinsic value through their relationship to and knowledge of the patient. Bioethics consultation services can assist clinical teams in engaging with the human crowd that surrounds each patient and navigating conflicting values and goals. This relational approach embraces the “mire and blood” that is endemic to relationships and clinical encounters.     

The evolution of integration: innovations in clinical skills and ethics in first year medicine

Critical self-reflection, medical ethics and clinical skills are each important components of medical education but are seldom linked in curriculum development. We developed a curriculum that builds on the existing integration of ethics education into the clinical skills course to more explicitly link these three skills. The curriculum builds on the existing integration of clinical skills and ethics in first year medicine. It refines the integration through scheduling changes; adds case studies that emphasise the social, economic and political context of our province's patient population; and introduces reflection on the "culture of medicine" as a way to have students articulate and understand their own values and moral decision making frameworks. This structured Clinical Skills course is a model for successfully integrating critical self-reflection, reflection on the political, economic and cultural contexts shaping health and healthcare, and moral decision making into clinical skills training.     

Evaluation of a worksheet to structure teaching and learning outpatient internal medicine

The Worksheet for Ambulatory Medicine (WAM) is an educational tool designed to enhance teaching and learning outpatient internal medicine. It was developed to identify student learning needs, focus teaching, and structure educational and patient care activities in a clinic setting. The purpose of the study was to assess the feasibility and educational value of using the WAM with medical students and preceptors. Sixty-five third- and fourth-year medical students and 12 supervising faculty at two university-based general medicine outpatient clinics used the WAM during required internal medicine clerkships. Students and faculty completed written evaluations. Results are reported as percentages of respondents agreeing or disagreeing with a variety of statements, and mean rating scores for several questions designed to assess the feasibility and educational value of using the worksheet. Student response rate was 89%; 83% found the WAM easy to use; 65% found it too structured. Half said the worksheet helped diagnostic decision making and note writing, and two-thirds thought it promoted careful thinking about differential diagnosis and aided in identifying learning issues. Some 56% said using the WAM motivated outside reading. Most students found it helpful for identifying patient agendas and focusing case presentations (61% and 67, respectively). Only 36% said the WAM helped with time management. Most preceptors thought the WAM helped identify earning issues, focus case presentations and clarify student expectations. There was less agreement among preceptors that it allowed them to demonstrate clinical reasoning or provide students with more autonomy in decision making. Nearly half the preceptors did not find it helpful with time management. Both students and preceptors rated the overall value and usefulness of the WAM as good to very good, and a majority recommended that others use it. Using the Worksheet for Ambulatory Medicine was feasible and educationally valuable for many third- and fourth-year medical students and their preceptor in a required ambulatory internal medicine clerkship.     

Tales publicly allowed: competence, capacity, and religious belief

What should we make of someone whose beliefs prevent her from accurately understanding her medical needs and care? Should that person still make her own health care decisions? In fact, she probably lacks decision‐making capacity. But that does not mean she is not competent.     

Glossed Over

I open the glossy cover of my alma mater's alumni magazine. Images of new frontiers in medicine abound—state‐of‐the‐art research buildings and the latest in high‐tech hospital innovations. Nestled among the articles on cutting edge medicine is a feel‐good story about medical students. The article profiles volunteer medical students and their uninsured patients with chronic health conditions. At one site, they are touted as providing “the only medical care.” Indeed. I think back to my medical student days and cringe.     

The outsider

On occasion, I am appointed by our local probate court to help make health care decisions for patients who are no longer capable of making these decisions for themselves. I'll visit my client in the hospital and review his or her chart, and I'll have numerous opportunities to talk about the illness and treatment options with attending physicians, residents, and nurses. My phone calls to them will be returned promptly. If doctors arrive at the conclusion that nothing further can be done for the patient, we'll discuss how best to proceed. This end‐of‐life talk never comes as a surprise to me because I will have been included in the team's decision‐making process, in the manner of professionals interacting with another professional. Last fall, I became a family surrogate decision‐maker after my mother was hospitalized over the course of a few months, for a number of reasons, in three hospitals in New York. I assumed I would manage her care as I did my clients’, with a sense of control and involvement in the process, but I quickly learned that I had in fact entered a kind of alternate universe as far as decision‐making was concerned.     

Lessons from Susan Sontag's Death

The standard model for end‐of‐life decision‐making gives roles to two parties—the physician, who explains the medical options, and the patient, who selects from among those options. The model can be harmful not only for individuals but also for the state, if the patient's right to control her own choices is understood as a positive right of access to whatever is available.     

Learning objects in medical education

A learning object (LO) is a grouping of instructional materials structured to meet a specified educational objective. Digital LOs, which can be stored electronically, allow a new approach to instructional activity, making medical education more efficient, and potentially more cost-effective. They are reusable and can incorporate text, graphics, animations, audio, and video to support and enhance learning. A learning object can stand alone or be aggregated with additional objects to create larger forms of educational content meeting multiple educational objectives. Digital learning objects located in online repositories can be accessed by many computers and are easily handled by an array of learning management systems for delivery to learners at any time. Integrating digital learning objects with traditional educational methods in a blended learning approach assists medical educators in meeting the challenges of competing priorities. Multimedia LOs enable learners to tailor their experience to their preferred learning style. Through the use of learning objects, learners' reactions, their acquisition of knowledge, skills and attitudes, and their behavioral changes become readily measurable. Learning objects provide multiple research opportunities, such as their use in adaptive learning, their added value in preclinical versus clinical education, and their impact as part of a blended learning strategy.     

The Influenza Controversy: Should Limits Be Placed on Science?

Should government have the power to place limits on a scientific pursuit that holds the potential for both good and harm—on what is called “dual‐use research”? That is the highly charged question surrounding research to genetically modify influenza A (H5N1) to render it more easily transmissible from human to human. There is seldom a “right” answer to dual‐use research, but a fair, inclusive, and transparent process—building on the NSABB model—should improve decision‐making. A local institutional panel should evaluate dual‐use research based on the following structured ethical framework.     

Critical care ethics in Hong Kong: cross-cultural conflicts as east meets west

The practice of critical care medicine has long been a difficult task for most critical care physicians in the densely populated city of Hong Kong, where we face limited resources and a limited number of intensive care beds. Our triage decisions are largely based on the potential of functional reversibility of the patients. Provision of graded care beds may help to relieve some of the demands on the intensive care beds. Decisions to forego futile medical treatment are frequently physician-guided family-based decisions, which is quite contrary to the Western focus on patient autonomy. However, as people acquire knowledge about health care and they become more aware of individual rights, our critical care doctors will be able to narrow the gaps between the different concepts of medical ethics among our professionals as well as in our society. An open and caring attitude from our intensivists will be important in minimizing the cross-cultural conflict on the complex issue of medical futility.