This piece was originally published on Bill of Health, the blog of Petrie-Flom Center at Harvard Law School.

There has been too little evaluation of ethics courses in medical education in part because there is not consensus on what these courses should be trying to achieve. Recently, I argued that medical school ethics courses should help trainees to make more ethical decisions. I also reviewed evidence suggesting that we do not know whether these courses improve decision making in clinical practice. Here, I consider ways to assess the impact of ethics education on real-world decision making and the implications these assessments might have for ethics education.

The Association of American Medical Colleges (AAMC) includes “adherence to ethical principles” among the “clinical activities that residents are expected to perform on day one of residency.” Notably, the AAMC does not say graduates should merely understand ethical principles; rather, they should be able to abide by them. This means that if ethics classes impart knowledge and skills — say, an understanding of ethical principles or improved moral reasoning — but don’t prepare trainees to behave ethically in practice, they have failed to accomplish their overriding objective. Indeed, a 2022 review on the impact of ethics education concludes that there is a “moral obligation” to show that ethics curricula affect clinical practice. Unfortunately, we have little sense of whether ethics courses improve physicians’ ethical decision-making in practice.

Ideally, assessments of ethics curricula should focus on outcomes that are clinically relevant, ethically important, and measurable. Identifying such outcomes is hard, primarily because many of the goals of ethics curricula cannot be easily measured. For instance, ethics curricula may improve ethical decision-making by increasing clinicians’ awareness of the ethical issues they encounter, enabling them to either directly address these dilemmas or seek help. Unfortunately, this skill cannot be readily assessed in clinical settings. But other real-world outcomes are more measurable. Consider the following example:

Physicians regularly make decisions about which patients have decision-making capacity (“capacity”). This determination matters both clinically and ethically, as it establishes whether patients can make medical decisions for themselves. (Notably, capacity is not a binary: patients can retain capacity to make some decisions but not others, or can retain capacity to make decisions with the support of a surrogate.)

Incorrectly determining that a patient has or lacks capacity can strip them of fundamental rights or put them at risk of receiving care they do not want. It is thus important that clinicians correctly determine which patients possess capacity and which do not. However, although a large percentage of hospitalized patients lack capacity, physicians often do not feel confident in their ability to assess capacity, fail to recognize that most patients who do not have capacity lack it, and often disagree on which patients have capacity. Finally, although capacity is challenging to assess, there are relatively clear and widely agreed upon criteria for assessing it, and evaluation tools with high interrater reliability. Given this, it would be both possible and worthwhile to determine whether medical trainees’ ability to assess capacity in clinical settings is enhanced by ethics education.

Here are two potential approaches to evaluating this: first, medical students might perform observed capacity assessments on their psychiatry rotations, just as they perform observed neurological exams on their neurology rotations. Students’ capacity assessments could be compared to a “gold standard,” or the assessments of physicians who have substantial training and experience in evaluating capacity using structured interviewing tools. Second, residents who consult psychiatry for capacity assessments could be asked to first determine whether they think a patient has capacity and why. This determination could be compared with the psychiatrist’s subsequent assessment. Programs could then randomize trainees to ethics training — or to a given type of ethics training — to determine the effect of ethics education on the quality and accuracy of trainees’ capacity assessments.

Of course, ethics curricula should do much more than make trainees good at assessing capacity. But measuring one clinically and ethically significant endpoint could provide insight into other aspects of ethics education in two important ways. First, if researchers were to determine that trainees do a poor job of assessing capacity because they have too little time, or cannot remember the right questions to ask, or fail to check capacity in the first place, this would point to different solutions — some of which education could help with, and others of which it likely would not. Second, if researchers were to determine that trainees generally do a poor job of assessing capacity because of a given barrier, this could have implications for other kinds of ethical decisions. For instance, if researchers were to find that trainees fail to perform thorough capacity assessments primarily because of time constraints, other ethical decisions would likely be impacted as well. Moreover, this insight could be used to improve ethics curricula. After all, ethics classes should teach clinicians how to respond to the challenges they most often face.

Not all (or perhaps even most) aspects of clinicians’ ethical decision-making are amenable to these kinds of evaluations in clinical settings, meaning other types of evaluations will play an important role as well. But many routine practices — assessing capacity, acquiring informed consent, advance care planning, and allocating resources, for instance — are. And given the importance of these endpoints, it is worth determining whether ethics education improves clinicians’ decision making across these domains.

This piece was originally published on Bill of Health, the blog of Petrie-Flom Center at Harvard Law School.
​
Medical students spend a lot of time learning about conditions they will likely never treat. This weak relationship between what students are taught and what they will treat has negative implications for patient care.

Recently, I looked into discrepancies between U.S. disease burden in 2016 and how often conditions are mentioned in the 2020 edition of First Aid for the USMLE Step 1, an 832 page book sometimes referred to as the medical student’s bible.

The content of First Aid provides insight into the material emphasized on Step 1 — the first licensing exam medical students take, and one that is famous for testing doctors on Googleable minutia. This test shapes medical curricula and students’ independent studying efforts — before Step 1 became pass-fail, students would typically study for it for 70 hours a week for seven weeks, in addition to all the time they spent studying before this dedicated period.

My review identified broad discrepancies between disease burden and the relative frequency with which conditions were mentioned in First Aid.

For example, pheochromocytoma, a rare tumor that occurs in about one out of every 150,000 people per year — is mentioned 16 times in First Aid. By contrast, low back pain — the fifth leading cause of disability-adjusted life years, or DALYs, in the U.S., and a condition that has affected one in four Americans in the last three months — is mentioned only nine times. (Disease burden is commonly measured in DALYs, which combine morbidity and mortality into one metric. The leading causes of DALYs in the U.S. include major contributors to mortality, like ischemic heart disease and lung cancer, as well as major causes of morbidity, like low back pain.)

Similarly, neck pain, which is the eleventh leading cause of DALYs, is mentioned just twice. Both neck and back pain are also often mentioned as symptoms of other conditions (e.g., multiple sclerosis and prostatitis), rather than as issues in and of themselves. Opioid use disorder, the seventh leading cause of DALYs in 2016 and a condition that killed more than 75,000 Americans last year, is mentioned only three times. Motor vehicle accidents are mentioned only four times, despite being the fifteenth leading cause of DALYs.

There are some good reasons why Step 1 content is not closely tied to disease burden. The purpose of the exam is to assess students’ understanding and application of basic science principles to clinical practice. This means that several public health problems that cause significant disease burden — like motor vehicle accidents or gun violence — are barely tested. But it is not clear Step 2, an exam meant to “emphasize health promotion and disease prevention,” does much better. Indeed, in First Aid for the USMLE Step 2, back pain again is mentioned fewer times than pheochromocytoma. Similarly, despite dietary risks posing the greatest health threat to Americans (including smoking), First Aid for the USMLE Step 2 says next to nothing about how to reduce these risks.

More broadly, there may also be good reasons why medical curricula should not perfectly align with disease burden. First, more time should be devoted to topics that are challenging to understand or that teach broader physiologic lessons. Just as researchers can gain insights about common diseases by studying rare ones, students can learn broader lessons by studying diseases that cause relatively little disease burden. Second, after students begin their clinical training, their educations will be more closely tied to disease burden. When completing a primary care rotation, students will meet plenty of patients with back and neck pain.

But the reasons some diseases are emphasized and taught about more than others often may be indefensible. Medical curricula seem to be greatly influenced by how well understood different conditions are, meaning curricula can wind up reflecting research funding disparities. For instance, although eating disorders cause substantial morbidity and mortality, research into them has been underfunded. As a result, no highly effective treatments targeting anorexia or bulimia nervosa have emerged, and remission rates are relatively low. Medical schools may not want to emphasize the limitations of medicine or devote resources to teaching about conditions that are multifactorial and resist neat packaging, meaning these disorders are often barely mentioned.

But, although eating disorders are not well understood, thousands of papers have been written about them, meaning devoting a few hours to teaching medical students about them would still barely scratch the surface. And even when a condition is understudied or not well understood, it is worth explaining why. For instance, if heart failure with reduced ejection fraction is discussed more than heart failure with preserved ejection fraction, students may wrongly conclude this has to do with the relative seriousness of these conditions, rather than with the inherent challenge of conducting clinical trials with the latter population (because their condition is less amenable to objective inclusion criteria).

Other reasons for curricular disparities may be even more insidious: for instance, the lack of attention to certain diseases may reflect the medical community’s perceived importance of these conditions, or whether they tend to affect more empowered or marginalized populations.

The weak link between medical training and disease burden matters: if medical students are not taught about certain conditions, they will be less equipped to treat these conditions. They may also be less inclined to specialize in treating them or to conduct research on them. Thus, although students will encounter patients with back pain or who face dietary risks, if they and the physicians supervising them have not been taught much about caring for these patients, these patients likely will not receive optimal treatment. And indeed, there is substantial evidence that physicians feel poorly prepared to counsel patients on nutrition, despite this being one of the most common topics patients inquire about. If the lack of curricular attention reflects research and health disparities, failing to emphasize certain conditions may also compound these disparities.

Addressing this problem requires understanding it. Researchers could start by assessing the link between disease burden and Step exam questions, curricular time, and other resources medical students rely on (like the UWorld Step exam question banks). Organizations that influence medical curricula — like the Association of American Medical Colleges and the Liaison Committee on Medical Education—should do the same. Medical schools should also incorporate outside resources to cover topics their curricula do not explore in depth, as several medical schools have done with nutrition education. But continuing to ignore the relationship between disease burden and curricular time does a disservice to medical students and to the patients they will one day care for.

This piece was originally published on Bill of Health, the blog of Petrie-Flom Center at Harvard Law School.

I recently argued that we need to evaluate medical school ethics curricula. Here, I explore how ethics courses became a key component of medical education and what we do know about them.

Although ethics had been a recognized component of medical practice since Hippocrates’ time, ethics education is a more recent innovation. In the 1970s, the medical community was shaken by several high-profile lawsuits alleging unethical behavior by physicians. As medical care advanced — and categories like “brain death” emerged — doctors found themselves facing challenging new dilemmas and old ones more often. In response to this, in 1977, The Johns Hopkins University School of Medicine became the first medical school to incorporate ethics education into its curriculum. Throughout the 1980s and 1990s, medical schools increasingly began to incorporate ethics education into their curricula. By 2002, approximately 79 percent of U.S. medical schools offered a formal ethics course. Today, the Association of American Medical Colleges (AAMC) includes “adherence to ethical principles” among the competencies required of medical school graduates. As a result, all U.S. medical schools — and many medical schools around the world — require ethics training.

There is some consensus on the content ethics courses should cover. The AAMC requires medical school graduates to “demonstrate a commitment to ethical principles pertaining to provision or withholding of care, confidentiality, informed consent.” Correspondingly, most medical school ethics courses review issues related to consent, end-of-life care, and confidentiality. But beyond this, the scope of these courses varies immensely (in part because many combine teaching in ethics and professionalism, and there is little consensus on what “professionalism” means).

The format and design of medical school ethics courses also varies. A wide array of pedagogical approaches are employed: most rely on some combination of lectures, case-based learning, and small group discussions. But others employ readings, debates, or simulations with standardized patients. These courses also receive differing degrees of emphasis within medical curricula, with some schools spending less than a dozen hours on ethics education and others spending hundreds. (Notably, much of the research on the state of ethics education in U.S. medical schools is nearly twenty years old, though there is little reason to suspect that ethics education has converged during that time, given that medical curricula have in many ways become more diverse.)

Finally, what can seem like consensus in approaches to ethics education can mask underlying differences. For instance, although many medical schools describe their ethics courses as “integrated,” schools mean different things by this (e.g., in some cases “integrated” means “interdisciplinary,” and in other cases it means “incorporated into other parts of the curriculum”).

A study from this year reviewed evidence on interventions aimed at improving ethical decision-making in clinical practice. The authors identified eight studies of medical students. Of these, five used written tools to evaluate students’ ethical reasoning and decision-making, while three assessed students’ interactions with standardized patients or used objective structured clinical examinations (OSCEs). Three of these eight studies assessed U.S. students, the most recent of which was published in 1998. These studies found mixed results. One study found that an ethics course led recipients to engage in more thorough — but not necessarily better — reasoning, while another found that evaluators disagreed so often that it was nearly impossible to achieve consensus about students’ performances.

The authors of a 2017 review assessing the effectiveness of ethics education note that it is hard to draw conclusions from the existing data, describing the studies as “vastly heterogeneous,” and bearing “a definite lack of consistency in teaching methods and curriculum,” The authors conclude, “With such an array, the true effectiveness of these methods of ethics teaching cannot currently be well assessed especially with a lack of replication studies.”

The literature on ethics education thus has several gaps. First, many of the studies assessing ethics education in the U.S. are decades old. This matters because medical education has changed significantly during the 21st century. (For instance, many medical schools have substantially restructured their curricula and many students do not regularly attend class in person.) These changes may have implications for the efficacy of ethics curricula. Second, there are very few head-to-head comparisons of ethics education interventions. This is notable because ethics curricula are diverse. Finally, and most importantly, there is almost no evidence that these curricula lead to better decision-making in clinical settings — where it matters.