Empowerment Failure: How Shortcomings in Physician Communication Unwittingly Undermine Patient Autonomy

Abstract

Many health care decisions depend not only upon medical facts, but also on value judgments—patient goals and preferences. Until recent decades, patients relied on doctors to tell them what to do. Then ethicists and others convinced clinicians to adopt a paradigm shift in medical practice, to recognize patient autonomy, by orienting decision making toward the unique goals of individual patients. Unfortunately, current medical practice often falls short of empowering patients. In this article, we reflect on whether the current state of medical decision making effectively promotes patients’ health care goals. We base our reflections, in part, on research in which we observed physicians making earnest efforts to partner with patients in making treatment decisions, but still struggling to empower patients—failing to communicate clearly to patients about decision-relevant information, overwhelming patients with irrelevant information, overlooking when patients’ emotions made it hard to engage in choices, and making recommendations before discussing patients’ goals.

Several decades ago, medical practice in the United States experienced a paradigm shift in medical decision making. Prominent legal cases established standards of informed consent that required clinicians to notify patients about the potential harms and benefits of medical interventions before patients decided whether to receive them (Katz 1984). End-of-life conflicts were resolved by courts that established the rights of patients and families to decide to withhold or withdraw life-prolonging treatment, even against their doctors’ desires (Quinlan 1977). Medical ethics experts convincingly laid out the importance of respect for patient autonomy. Clinical experts began promoting decision aids and other tools as mechanisms for promoting patient autonomy and shared decision making (Wennberg 2010). The goal of all these efforts—legal, ethical, and clinical—was to make sure that health care decisions would more often reflect patients’ individual preferences (Ubel 2012).

These groups all recognized that many health care decisions depend not only upon medical facts, but also on value judgments, specifically upon patient goals and preferences (Ubel 2015). A paradigmatic example of such a “preference-sensitive decision” is localized prostate cancer (Flood et al. 1996). A man with this diagnosis faces a choice between several medically viable treatment approaches. He can choose an active treatment, such as surgery or radiation therapy, or opt for surveillance instead. These alternatives have important pros and cons. Surgery and radiation eliminate localized cancer cells but at the cost of exposing patients to both the short-term misery of the treatments—the ardors of surgical recovery and the burdens of a six-week course of radiation therapy—and the longer term likelihood of experiencing erectile dysfunction and/or urinary dysfunction (Wilt et al. 2012). By contrast, watchful waiting and active surveillance (closely related approaches, but with the latter involving more frequent and aggressive follow-up testing) do not lead to these side effects, but do expose patients to the anxiety of living with untreated cancers inside their bodies, and the burdens of regular follow-up tests.

Prior to the paradigm shift, decision making around topics like localized prostate cancer was relatively straightforward. Patients relied on doctors to tell them what to do. In an extreme example of this “doctor knows best” practice, approximately 90% of U.S. physicians surveyed in the 1950’s stated that they routinely withheld cancer diagnoses from patients who they thought would be harmed by the information (Fitts and Ravdin 1953). Before lawyers, ethicists, and a handful of enlightened clinicians promoted the importance of patient autonomy, the majority of physicians viewed themselves as the primary medical decision makers, believing that their clinical training placed them in a position to determine what was best for any given patient.

Practice has changed dramatically since that time. Few physicians routinely withhold cancer diagnoses from their patients anymore (Holland et al. 1987). Indeed, U.S. medical schools now educate students about the importance of respecting patient values, with most schools including sizable doses of ethics and communication training in their curricula (Dubois and Burkemper 2002).

And yet, current medical practice too often fails to promote patient autonomy. The extent of this failure was made clear when we conducted a study of decision making among patients with localized prostate cancer. We analyzed more than 250 decision making clinical encounters, audio-recorded at four academic VA hospitals across the country, often from the moment the patients learned from their urologists—most of them fellows in training—that they had cancer. We provided decision aids to all the patients in the study before they met with their urologists, in an effort not only to help them understand their treatment alternatives but also to explain the importance that their goals play in determining what treatment is best for them. Nevertheless, in studying those clinical encounters, we witnessed what we characterize as empowerment failure. The urologists we studied made earnest efforts to partner with their patients in making treatment decisions. But most failed.

Physicians in the United States have been told to inform patients about their health care alternatives without being taught how to do so effectively. Physicians have been asked to implement the concepts of shared decision making in an environment that does not always support these changes, with economic and time pressures that do not reward good communication practices. Far too often, the end result is a clinical encounter that has the outward appearance of a shared decision making event but that, at its core, is not very different from old fashioned paternalism. This illusion of patient empowerment is particularly concerning because it transforms overt, clearly identifiable persuasive messages into more subtle (yet powerful) messages which may be harder for patients to identify as being persuasive.

In this article, we reflect on the current state of medical decision making, and whether it effectively promotes patients’ health care goals. Specifically, we describe findings from our analysis of prostate cancer decision making conversations, and related findings from the literature on shared decision making, to lay out several steps clinicians and medical educators—especially those teaching about the ethics of the physician-patient relationship—can take to improve decision making partnerships, and thereby more effectively promote patient autonomy.

SOURCES OF EMPOWERMENT FAILURE

We evaluated prostate cancer clinical encounters through the lens of shared decision making. Proponents of shared decision making contend that in situations where the best medical choice depends on patient preferences, an ideal clinical encounter will involve partnership between the patient (with our without loved ones) and clinicians (Charles, Gafni, and Whelan 1997). Shared decision making respects patient autonomy, by involving patients in their health care choices without forcing them to feel like they are making the decision by themselves. Standards of shared decision making indicate that patients should be informed about the pros and cons of their treatment alternatives, and that providers should explicitly factor in patient-specific values or preferences regarding those pros and cons when helping patients choose among those alternatives (Keirns and Goold 2009). In the case of localized prostate cancer, for example, the treatment choice should factor in how strongly patients desire to maintain sexual function. All else equal, a man who wants to maintain sexual function should be more likely to receive surveillance than one who does not.

In our prostate cancer research, we observed many examples of how the paradigm of shared decision making and patient empowerment has influenced clinical care. Urologists in our study almost uniformly provided patients with detailed information about the extent of their cancers, as well as thorough descriptions of the pros and cons of their three main treatment alternatives. Indeed, according to an established measure of shared decision making, Braddock’s Informed Decision making coding system (Leader et al. 2012, Pricea et al. 2012), the urologists in our study performed better than many physicians have performed in other studies (Holmes-Rovner et al. 2015).

Evidence of shared decision making was not limited to information provision, but was also apparent in the way urologists discussed decision making with their patients. Some urologists in our study took pains to emphasize that the patient had the last word in the decision making process. As one physician stated: “You can weigh the options; you can talk to people and then, you know, you can make what you think is the best decision for you.” Somewhat surprisingly, physicians’ embrace of shared decision making even extended to a forthrightness among many urologists to acknowledge their own biases, in an effort to minimize the chance that they would unduly influence their patients’ decisions. In some of the encounters, surgeons disclosed these biases, saying things such as “I’m a surgeon, so I’m biased in favor of surgery” (Sah, Fagerlin, and Ubel 2016).

Despite physicians’ efforts to promote shared decision making, however, we observed serious failures both at the level of information provision—helping patients comprehend the pros and cons of their treatment alternatives—and in physicians’ attempts to share in decision making responsibilities with patients, helping them choose the alternative that best fits their preferences. In the following, we outline what struck us as the most prominent of such failures, by comparing these interactions with accepted standards for how to achieve shared decision making (Charles, Gafni, and Whelan 1997).

Difficulty Providing Information

Ideally, health care providers will communicate with patients in a clear, concise, and unbiased manner. Our observations reveal many physicians falling short on all three accounts.

A. Failure to Communicate Clearly

In most encounters, urologists informed patients they had cancer and then, before discussing treatment alternatives, explained the stage of the patients’ cancers in hopes, presumably, that patients would better understand their situations. A typical staging explanation ran as follows:

Urologist: We grade prostate cancer on how it looks under the microscope. We give it a score between 6 and 10.

Patient: Is that the Teason?

Urologist: That’s the Gleason Score

Patient: Oh, Gleason Score, ok

Urologist: Yep, so 6 is what we consider the most low-grade, least aggressive looking, but it’s the most … it’s just abnormal enough for us to call it cancer. If it were any less than that, if there were less atypical looking cells, we couldn’t call it cancer. So it’s just enough to get a grade of cancer and then that goes all the way up to a score of 10 which is very abnormal looking and is more aggressive.

Patient: But 6 is the beginning number?

Urologist: 6 is the least aggressive, 10 is the most aggressive.

Patient: I’m used to like, 1 (laughing)

Urologist: Yeah well, the way we typically split it up is into thirds; low risk, intermediate risk, and high risk.

Patient: Right

Urologist: Low risk is Gleason 6, intermediate is usually 7’s, either 3+4 or 4+3, depending on how it looks under the microscope, and then 8, 9 and 10 are all high risk. So yours was an intermediate risk. So it’s in the middle. It was 3 + 3 and 3 + 4, so just enough of the atypical cells of the grade 4 to make it 3 + 4, which means you’re intermediate risk.

We are confident that this elaborate description of the patient’s Gleason score would not have taken place if the paradigm of medical decision making had not shifted from a perspective of “Doctor knows best” to one of patient empowerment. This urologist appears to be making an earnest effort to educate this patient about his disease. Moreover, the general thrust of this conversation was meant to comfort the patient; after all, his tumor had “just enough” of the grade-4 cells to achieve as Gleason score of 7.Any fewer and his cancer would have been a Gleason-6, the lowest possible score!

But the comfort of these words is probably lost in translation. The urologist seems to have forgotten that words like “Gleason score” and “atypical cells” would be unfamiliar to the patient. The urologist also seems to overlook the possibility that the patient will be confused by all the numbers being bandied around—the 3 + 3 and 3 + 4 somehow amounting to a score of 7.

Use of medical jargon is common in clinical encounters and can lead to undue anxiety when patients misinterpret physician lingo. In a study of patients who had undergone aneurysm repair, for example, some physicians referred to a minor imperfection of the graft (an imperfection that requires no further treatment) as “an endoleak.” This term caused patients to worry that they were at risk for imminent doom (LeBlanc et al. 2014). And who can blame them? Who wouldn’t be scared to learn that the tissue holding their artery together is leaking? Respect for autonomy depends upon patients understanding their health conditions and their health care alternatives. The urologists in our study seemed to acknowledge the importance of their role as information providers, but did not necessarily have the ability to explain medical concepts in ways that most patients could understand.

B. Failure to Communicate Tersely, Due to an Ethic of Information

Urologists in our study typically described each treatment alternative in elaborate detail and, from what we could see among those who interacted with multiple study patients, delivered standard “spiels”—general ways of describing information to patients. Development of such patterns helps make clinical interactions more efficient. However, many of the urologists in our study were wedded to lengthy descriptions, complete with details that appeared irrelevant to the decision at hand. The routine nature of such details revealed itself when patients interrupted the flow of their physicians’ spiel:

Urologist: There are three different things that we can do, I’m going to go from the least invasive to the most aggressive. One is to do active surveillance, which means that we don’t actively treat right now, we watch it.

Patient: Do blood, with blood work?

Urologist: Yep, let me, let me finish

Patient: Okay

Urologist: And what we’ll do is we’ll have you come back in about four to six weeks, repeat the biopsy. When we repeat the prostate biopsy, we would take more than the twelve, we’ll take about 18 or so.

In extreme circumstances, patients tried to interrupt urologists’ spiels to reduce the amount of time spent discussing alternatives that the patients were not interested in. Such interruptions were also met with resistance, as urologists pushed back to provide patients with thorough information about all their alternatives: In one encounter, the urologist had begun to explain the risks of erectile dys-function from surgery when the patient interrupted:

Patient: Okay, half that equation doesn’t bother me. The sexual aspect of it doesn’t even bother us. My wife had breast cancer; we haven’t had sex in, God now, several years. The point being that that’s not a concern with us.

Urologist: I understand.

Patient: Okay.

Urologist: But it’s my job as a physician to explain the risks and side effects. So the worst-case scenario is that I do surgery on you, and then …

Across the encounters, we came away with a strong sense that urologists were committed to an ethic of information. They put great effort into delivering information about each of the treatment alternatives to each patient. This effort took time; more than half of most encounters was spent informing patients about their alternatives. For this reason, it is understandable that urologists did not want to be pulled off their educational script. Efficient clinical practice can demand that clinicians deliver this information to patients within the confines of a busy clinic schedule. When this informational ethic collides with time constraints, clinicians may feel less able to engage in back-and-forth dialogue with their patients.

C. Failure to Provide Unbiased Information

The information urologists provided to patients was shaped by a mixture of conscious and unconscious biases. Such biases are inevitable, but recognizing them is important in assessing how well these encounters promote patient abilities to make autonomous choices.

For example, physicians sometimes explicitly disclosed their biases to patients (Sah, Fagerlin, and Ubel 2016). This typically occurred in the context of their descriptions of the relative pros and cons of surgery versus radiation, when they would admit to being biased in favor of surgery. Such admissions of bias, however, did not cause patients to discount the physicians’ opinions. Instead, after statistically adjusting for a wide range of variables, we discovered that patients whose surgeons admitted to being biased in favor of surgery were more likely undergo surgery (Sah, Fagerlin, and Ubel 2016). The admission of bias appeared to make patients trust physicians more, thereby making them more susceptible to influence (Cain, Loewenstein, and Moore 2005). We do not believe that physicians were aware of this paradoxical effect of admitting their biases but, rather, believe that physicians were honestly trying to suggest that patients take their opinions with a grain of salt.

As another example of potentially unconscious biases, consider the way urologists described robotic prostatectomy, a newer approach to surgical removal of the cancer that was unavailable at one of our four research sites. We discovered that urologists at that site were less likely to describe robotic surgery as being superior to more traditional “open” surgery (Scherr et al. 2017a). For example, one urologist at that location explained the impact robotic surgery would have on incontinence as follows: “You may hear things about … getting control of your urine back sooner, and I would say there is a lot of talk about that. But the evidence is not there to really say whether we know that or not.” At the other three sites, urologists were more likely to explain that the robot reduced the chance of urinary incontinence. As one urologist put it: “I think I’ve really seen a difference in the urine control.”

To be clear, we do not believe any of these urologists were purposely misrepresenting, or even spinning, the truth about robotic versus open surgery. Instead, we expect they were honestly informing patients about the truth as they perceived it. Nevertheless, it means that patients’ perceptions of robotic surgery were, unbeknownst to them, shaped by whether the urologists had access to that technology.

Difficulty Promoting Effective Decision Processes

Promoting patient autonomy and shared decision making is about more than informing patients about their treatment alternatives. It also requires giving patients a proper role in the decision and, perhaps even more importantly, making sure that choices reflect patient-specific values and preferences. For all the failures we observed in the way physicians informed patients about their treatment alternatives, we are even more concerned with the way physicians handled the process of making a decision.

A. Failure to Account for Patient Emotions

For patients to participate in health care decisions, they need to be psychologically prepared to engage with their clinicians. But the physicians in our study often seemed to overlook strong patient emotions that could impede optimal decision making. Consider the following example, one typical for our study. Less than 2 weeks earlier, the patient had undergone a prostate biopsy to determine whether his PSA (prostate-specific antigen) blood test was elevated because of cancer. He arrived at the urology clinic and the urologist broke the bad news:

Urologist: So, we took twelve cores out of your prostate. Out of those, there were three cores that had cancer in them, and the percentage of the cores with cancer was fairly low, it was under thirty percent. So out of those three cores … a third of them had a little bit of cancer in them. So those three cores out of twelve says that there’s probably not an extensive amount of prostate cancer in your prostate. But we should talk about different treatment options.

There is a lot going on in this brief soliloquy, but for now we want to focus on something that is absent—the urologist did not give the patient time to recover from the emotional shock of learning he has cancer.

The decision psychology literature provides no simple take on when emotions improve versus impair people’s ability to make good decisions. Studies have shown that a lack of any emotion, or of normal emotional abilities, dramatically impairs people’s decision making (Damasio 1994). But strong emotions can also lead people astray (Loewenstein et al. 2001). Most experts would conclude that extremely strong emotions, in the setting of non-urgent decisions, are not generally conducive to good decision making (Andrade and Ariely 2009). In addition, strong emotions can impair people’s abilities to absorb decision-relevant information (Joëls et al. 2006). In the case of this particular patient, he had just learned he had cancer and yet the physician appeared ready to launch into decision making mode. Rushing to such a decision is not a recipe for patient empowerment.

More than 40 urology trainees participated in our study. Yet few of them, after informing men that they had cancer, acknowledged and dealt effectively with patient emotions.

Consider the following interaction, typical for what we saw across many encounters:

Urologist: what we found is that you do have a prostate cancer it looks like.

Patient: I do have?

Urologist: Uh-huh

Patient: (gasps)

Urologist: it’s what we call a low risk prostate cancer. We sort of stratify prostate cancer into three categories—low, moder-ate and high risk …

This urologist went on to describe the technical details of the diagnosis, perhaps because he was uncomfortable dealing with the patient’s emotions or because he felt that the technical details of the diagnosis would comfort the patient. In either case, the patient was left emotionally hanging, expected to absorb the subsequent information despite his distress.

The urologists may, in part, have avoided patient emotion because they hadn’t been trained on how to handle such feelings. Studies of senior oncologists, with many years of experience, have shown similar inattention to patient emotion (Pollak et al. 2007). In one study of audio-recorded encounters, researchers identified explicit instances in which patients verbally expressed negative emotions. The researchers discovered that senior oncologists responded appropriately to these emotions only one in five times, either by acknowledging the emotion (“I understand you are frightened”) or by addressing their emotional needs (“I know you are scared, but I think I can help you with that …”).

If clinicians are blind to their patients’ emotions, or are unprepared for how to respond to such emotions, it is difficult for them to effectively partner in making health care decisions.

B. Failure to Recognize When “Medical Recommendations” Incorporate Value Judgments

One could argue that every medical decision is, in theory, preference sensitive—that all patients have the right to refuse medical care or to decide whether the best alternative is one they are interested in receiving. A person could decide whether to be casted for a broken arm, whether to receive antibiotics for a nasty infection, or whether to take pills for dangerously elevated blood pressure. In these cases, physicians would seem obligated to discuss the pros and cons of these interventions in detail, and thereby empower patients to make these choices. In practice, however, physicians usually tell patients what is going on (“You have a hairline fracture of your ulna”) and what they are going to do about it (“We will put you in a cast for six weeks and then, if the follow up x-ray looks good, you’ll be set to go”.) Whether this practice is paternalistic is debatable. Under most of these circumstances, physicians can reasonably assume that patients are coming to them to be treated for their problem—they want their bone fixed, their infection treated, or their blood pressure brought down to less hazardous levels. In these cases, the appropriate treatment is mainly a matter of medical evidence. The physician will know what position to cast the patients hand in, and what antibiotic and blood pressure pills work best for that particular patient’s medical condition.

The challenge for clinicians is to differentiate these medical situations—where the medical facts and implied preferences of the patients point largely in one direction— from more value-laden situations—where reasonable patients could disagree about what is the best course of action. Our analysis of prostate cancer decision making encounters reveals that many physicians have a hard time meeting this challenge. Physicians frame treatment recommendations as pointing toward the medical standard of care, even though patients could reasonably make alternative choices based upon their preferences.

For example, prior to urology appointments, we asked patients to indicate what treatment or treatments they would be most interested in if they discovered they had localized prostate cancer. A number of younger men in our study (those under 65 years of age) responded that they were only interested in surveillance. Some of these men were sexually active and did not want to take on the risk of erectile dysfunction from surgery or radiation. But urologists often told these patients that surveillance was not an appropriate alternative for them, given their relative youth. Indeed, analysis of our data revealed that patients’ treatment preferences prior to seeing their urologists did not have any association with the treatment that they ultimately received (Scherr et al. 2017b). By contrast, treatment was strongly associated with physician recommendations. For example, when urologists recommended active treatment, more than 80% of patients received such treatments, but when they recommended against active treatment, less than 5% of patients chose to go against their recommendation.

The strong influence of physician recommendations is concerning because physicians typically framed such recommendations as incontrovertible medical facts. “You are young and healthy,” they would say, “so active surveillance is not an option for you.” This type of judgment was based on controversial evidence about whether active treatments like surgery and radiation reduce mortality in younger men compared to surveillance (Bill-Axelson et al. 2014). Different studies have come to different conclusions about this issue (Wilt et al. 2012, Bill-Axelson et al. 2014). But the urologists in our study clearly believed that active treatments would increase long-term survival in younger and healthier patients. Yet they did not back up their treatment recommendations with any explanation of the magnitude of this potential survival benefit, nor of the controversy about whether such benefits exist. These younger patients were given little idea of what the real trade-offs were—what risk of erectile dysfunction and incontinence they would face in order to increase survival by a certain amount.

C. Failure to Diagnose Patient Preferences

Not only did physicians typically frame recommendations as incontrovertible medical facts, they also failed to take patients’ preferences into account when making recommendations. Patients often look to physicians for advice. Even in the context of preference-sensitive decisions, patients often ask physicians “What should I do?” or “What would you do if it was your dad?” They seek such advice in part because they, too, might fail to recognize the important role their own preferences play in making choices. They may also seek advice because they are scared, or overwhelmed, or do not want to bear the responsibility for making the decision (Ubel 2002). Not all patients want to take a prominent role in their health care choices. Respect for autonomy obligates clinicians to allow patients to cede decision making authority when that fits patients’ values.

But even when patients want doctors to make decisions for them, or when they want to follow their doctors’ treatment recommendations, they should expect that physicians will make decisions that reflect their values. In other words, physicians ought to understand individual patient’s preferences well enough to give them good advice. In the setting of localized prostate cancer, for example, urologists should inquire about patients’ desires to be sexually active before recommending surgery. Mulley and colleagues refer to this as “diagnosing preferences” (Mulley, Trimble, and Elwyn 2012).

Unfortunately, across the more than 250 encounters in our study, explicit discussion of patient preferences was not the norm. In terms of sexual function, for example, urologists asked more than two-thirds of patients whether they had normal erectile function, but in less than 15% of the encounters did they ask patients how important sexual activity was to them or inform them that their preferences regarding sexual activity should potentially influence their treatment choices (Scherr et al. 2017b). Physicians cannot give patients good advice if they do not make the effort to understand what patients care about.

D. Failure to Activate Patients

When our research team first began analyzing these clinical encounters, we were struck by how much the urologists spoke compared to the patients. Partly this relative verbosity reflects the need for urologists to impart large amounts of information on their patients. But even later in encounters, when the patients had been educated and were now trying to decide about treatment choices, urologists still frequently dominated the conversations.

The passivity on the part of patients is understandable. After all, these men had just learned they had cancer. The shock of the diagnosis could have reduced their abilities to assert themselves. In addition, the overwhelming amount of jargon-laden information could have silenced them. These clinical encounters also occur in a hierarchical setting—the physician with her medical degree and white coat, the patient with, perhaps, a high school education and street clothes. It is not surprising that in such a social setting, one party is more talkative than the other.

Nevertheless, physicians frequently missed out on opportunities to activate their patients. Rather than draw them out with open-ended questions (“What’s important to you?”), they would ask closed-ended ones, and often the type of questions that encouraged submissive responses (“Do you understand what I have explained?”). It is difficult to achieve patient empowerment if patients are mute.

Across the encounters we studied, physicians frequently made references to patient decision making authority, acknowledging the new paradigm of patient empowerment. But often such recognition was framed in ways that minimized patient power. Consider the following encounter in which a patient, after learning his treatment alternatives, leaned toward active surveillance even though the urologist had tried to downplay the appropriateness of that alternative:

Patient: I think I would go with the surveillance myself, because I’m 67 now. I mean what the heck, anyway.

Urologist: As I was saying: for low risk cancers we would recommend surveillance. For your type of cancer, we recommend treatment. You’re 67, but you’re … a very healthy 67-year-old guy. You probably have 20 more years to live. It’s better to kind of treat it initially before it gets spread. Now I mean, you are your own boss. And it’s always up to your choice.

The concept of this patient being his “own boss” is undermined by the urologist’s strong treatment recommendation.

Or consider another urologist who had recommended surgery, which the patient was not inclined to receive. His response was to say: “It’s your decision. I’m not going to drag you back to the operating room kicking and screaming.” This type of language acknowledges patient decision making authority while simultaneously under-mining it.

HOW TO AVOID THESE FAILURES

When bioethicists began writing about patient autonomy, the concept was controversial. The idea that patients deserved a say about their health care choices was not universally accepted, much less the idea that they should have the final say about those choices. Our study of prostate cancer decision making suggests that, at least in some clinical contexts, physicians in the United States now largely accept the importance of patient autonomy. Across the encounters we studied, we discovered numerous instances of urologists acknowledging patient autonomy. Almost universally, for example, urologists framed the decision as one that patients could make. Even when the urologists made strong clinical recommendations to their patients, they frequently reminded the patients that they could choose alternative treatments if they wanted to. It was not the importance of autonomy that was being denied in these encounters but instead, in Carl Schneider’s memorable phrase, it was the “practice of autonomy” that was being mishandled (Schneider 1998).

We do not believe these practices will be improved primarily through better education about ethical principles. It is not lack of familiarity with the bioethical principles laid out by Beauchamp and Childress (2006) that stands in the way of patient empowerment. Instead, physicians need the kind of communication training, and the reinforcement of such training, that will develop their abilities to partner more effectively with their patients. Consider several well-established communication techniques, techniques largely lacking in the prostate cancer encounters we studied.

Assessment of Baseline Understanding

Urologists in our study rarely took the time to assess patient baseline understanding of their disease or of their treatment alternatives. As mentioned earlier, our study took place in the context of a trial of two decision aids, meaning that most of the patients had read about prostate cancer treatments before their appointment. Yet urologists did not try to assess what these patients knew about their treatment alternatives before launching into long discussions of these treatments. Not that such assessment necessarily takes much time. Indeed, such assessments may shorten other aspects of the clinical encounters, by relieving physicians of the need to explain things that patients already understand. More importantly, such assessments could help physicians tailor their communication to fit patients’ informational needs.

Teach Back

Once physicians explain new information to patients, communication experts recommend that they ask patients to explain this information back to them in their own words (Jager and Wynia 2012). This request can be handled in nonthreatening ways that discourage submissive responses: “I have just explained a lot to you and want to make sure I did a good job. Can you tell me, in your own words, what you heard me say about your treatment alternatives?” Such “teach back” further enables physicians to address patient confusion, and to tailor their communication to patients’ informational needs. Perhaps even more importantly, this teachback technique gets patients talking. True clinician/patient partnership is not achievable if patients remain silent about their thoughts and perceptions.

Diagnosing Preferences

Instead of assuming that patients will spontaneously verbalize their values, clinicians should strive to find out what is important to their patients. A simple acknowledgment of patient preferences can go a long way toward promoting better discussion: “I am the expert on the medical facts, but you are the expert on you. I need to understand what matters to you before I can work with you to figure out what treatment is best for you.” Sometimes experts contend that shared decision making is not for all patients, because some people prefer letting their doctors make decisions for them. But even in these circumstances, physicians ought to understand their patients well enough to make good clinical recommendations (Mulley, Trimble, and Elwyn 2012).

Education After Medical School

The teaching of these communication techniques should be continued past medical school. Consider the urology fellows participating in our study. We expect that none of them were allowed to make a surgical incision in a live patient until they had witnessed dozens of such incisions made by more experienced clinicians. We expect that when they began making such incisions themselves, they did so under close supervision for dozens of times, if not for the duration of their residency. Unfortunately, the same cannot be said for their communication training. Most of the urology fellows recorded in our study were not being observed by the urology faculty during these clinical encounters. Telling patients they have cancer and helping them make potentially life-changing decisions was not treated as a skill that requires regular observation and reinforcement.

We think ethicists can play an important role in improving how clinicians communicate with patients in the context of shared decision making. Many ethicists play important roles in the teaching of the doctor–patient relationship in medical schools. In those roles, we encourage them to emphasize the moral importance of developing strong communication skills, especially in the context of preference-sensitive decisions. Moreover, we expect such education will be more effective if it emphasizes concrete examples of communication failures, rather than abstract theories about autonomy and the like. Using transcriptions of actual physician–patient encounters could help students recognize the difference between the ethical ideals they have been taught and the way medical care is often practiced.

We also think that, whenever possible, ethicists should engage in post-medical-school training, especially residency and fellowship training, when many early career physicians develop moral habits that will persist for many years. Ethicists interested in promoting patient autonomy should collaborate with residency educators, to increase attention to important communication training.

CONCLUSION

Patient empowerment often succeeds or fails at the hands of clinicians. Especially in the setting of serious and/or emotionally charged health care contexts, many patients will not assert themselves without help from their physicians. Ethics experts have largely succeeded in changing the paradigm of physician–patient interactions, so that newly trained physicians no longer believe that medical decisions are purely medical in nature. In fact, we expect that the amount of clinically relevant information provided to patients today dwarfs that provided to them several decades ago. Thanks in large part to the work of ethicists, most physicians today recognize that patients have a right to decision-relevant information. Most even recognize that patients have the right to be involved in making important medical decisions. However, many physicians fail to interact with patients in ways that lead either to true empowerment or to decision making partnership with their clinicians.

Good ethics often requires good communication. If we hope to promote patient autonomy, we must go beyond teaching ethical principles and improve how physicians interact with their patients. To promote patient autonomy and empowerment, we need to teach physicians how to partner with patients in making preference-sensitive choices, and reinforce these teachings throughout their training.