Editor’s Note: A Commentary on this Research Report appears on page 415.
Efforts have been made to improve medical training in order to foster more compassion in doctors,1,2 yet many questions remain as to whether empathy can in fact be systematically taught; if so, how effectively3; and how medical educators should address these issues. Disagreement exists even as to whether empathy is a skill or an attitude.4 Medical students often learn values from attending physicians implicitly by modeling, rather than through explicit communication.5–7 Yet within medical schools, the need to teach growing amounts of scientific information increases pressures on faculty and students.5 Over the course of their training, medical students have been found to become more cynical and less compassionate.8 Recent efforts to increase students’ empathy through reading and discussing literature9 and interventions to improve communication skills may help in the short-term, though results have not been entirely consistent10,11 and the long-term impacts remain under-assessed.
Physicians who themselves have become patients occupy both roles, and thus potentially can provide critical perspectives on issues of communication and compassion. From their experiences as patients, they may gain insights for improving doctor–patient communication and relationships. These insights may then be applied to communication training in medical education.
The experiences of doctors who become patients have been explored anecdotally,12–16 and only rarely have been explored systematically, focusing on issues of self-doctoring.17 Still, these anecdotal reports suggest that the experience of switching between roles can foster insights that these doctors had not previously had concerning the experience of patienthood. In previous studies, I have examined how doctors who become patients alter their views of spiritual issues18 and of risks and benefits,19 but issues of physician–patient communication are clearly vital to investigate as well.
After the Columbia University Department of Psychiatry institutional review board approved the study in 1999, based in New York City, I conducted pilot interviews with 20 health care professionals—17 physicians and three other health care professionals—about issues concerning physicians who become patients. These interviews led to the development and refinement of an interview guide. In 1999–2002, I then recruited participants for the full study, which took place in two phases. In the first stage, my focus was on HIV-infected doctors. In a second stage, I expanded the study to include physicians with other diagnoses as well. I recruited participants via e-mailed announcements (e.g., “Are you, or do you know, a physician with a serious illness?”), Web sites, word of mouth, and advertisements in newsletters. Serious illness was self-defined, and then confirmed by the principal investigator. Potential participants responded to advertisements recruiting physicians with serious illness. I conducted two in-depth, semistructured, two-hour-long interviews concerning experiences of being a health care worker and becoming a patient with each of these 50 participants, who were interviewed in several cities (New York, Seattle, and Los Angeles). I conducted all the interviews at participants’ offices or homes or in my office—whichever was more convenient for participants.
As described below, account was taken of the fact that different themes might arise with different participants, due to different professional experiences. However, the themes raised concerning medical education among respondents did not differ systematically or substantially between them in this regard. Thus, these interviews were all included in the analyses here. I asked participants about their experiences both as patients and as providers, and about other aspects of their lives. Relevant sections of the interview guide, through which I sought to obtain detailed, “thick” descriptions20 of these issues, are attached (see List 1). To have sufficient time to cover all of the questions, I conducted two interviews with each participant. I transcribed and analyzed the interviews during the period in which the interviews were conducted. Interviews were conducted until major and minor themes became clear.
My analysis of the interviews was informed by grounded theory.21 Once the full set of interviews was conducted, subsequent analyses were conducted in two additional phases, primarily by me together with a research assistant who had social science training. At several points during the coding process, we also received input from an additional, senior expert in qualitative research.
The research assistant and I each independently examined a subset of interviews to assess factors that shaped participants’ experiences, identifying categories of recurrent themes and issues that were subsequently given codes. We assessed similarities and differences between participants, examining categories that emerged, ranges of variation within categories, and variables that may have been involved. We developed a coding manual and examined areas of disagreement about themes until we reached consensus. We discussed new themes that did not fit into this original coding framework, and made modifications accordingly to the manual, when appropriate. In the second phase of the analysis, we refined and merged subdivided thematic categories into secondary or subcodes, when suggested by associations or overlap in the data. We then used these codes and subcodes in analyzing all of the interviews. To ensure coding reliability, all interviews were analyzed by each of us. Again, we examined areas of disagreement until consensus was reached. Major codes included, for example, ways to improve physician empathy. Subcodes included, for instance, approaches that dealt with the process (as opposed to the content) of care, or with end-of life issues. We conducted thematic content analysis and triangulated methods, referring to the anecdotal reports and studies that have been published, as described above.
Though length of illness varied among participants, we collected data concerning their responses to illness over time, which enabled analyses of information from respondents regarding their experiences earlier as well as later in the course of their illnesses, and allowed for comparison of responses with regard to different stages in the progression of illness. The concerns and themes that arose regarding medical education did not appear to vary significantly between respondents based on the length of their illnesses.
Forty-eight doctors, one dentist, and one medical student (referred to below as “doctors” or “physicians”) responded to my inquiry. Their ages ranged from 25 to 87; 49 were white and one was Latino; 40 were men and 10 were women; and 27 were HIV positive and 23 had other medical problems, including cancer, heart disease, and hepatitis C.
As summarized in List 2, several themes emerged from the interviews. Overall, respondents raised questions first as to whether empathy could be taught or not, and if so, how well; and second, given disagreement in this area, about how nevertheless the process and content of physician–patient interactions might be improved. In general, two sets of domains emerged concerning ways of possibly improving medical education—related to process and to content— though at times these overlapped or were closely tied.
Can empathy be taught?
These physicians varied in their views of whether empathy can be taught, and if so, how. Several physicians suggested that ideally, medical students should be required to be admitted to inpatient wards to sleep in patient rooms in order to experience the disruptions, inconveniences, powerlessness, and humiliations that patients confront. Several participants suggested that students should be admitted to hospitals, since that experience—involving helplessness, loss of power over one’s very body and life, confusion, and confrontation with the unknown—could powerfully transform trainees. Indeed, in its ability to make physicians appreciate fully what it is like to be a patient, patienthood was seen as being unique. One respondent thought that only the process of becoming ill would make medical students more empathetic:
If you’re not personally affected, it’s impossible to imagine what it’s really like. I’m not sure what would make students more empathetic. Maybe you have to go through it yourself.
Yet other physicians thought that specific skills could be taught to heighten students’ sensitivity without having all medical students hospitalized. In general, several felt that much of doctor–patient relationships and communication was best inculcated not explicitly through lectures, but implicitly through role modeling. A pediatrician with cancer tried to instruct trainees
… almost by osmosis. They observe me. I try to get across … the way I relate to patients and parents. It’s just very informal. That’s part of pediatrics. But I am especially informal.
She suggested, though, that degrees of styles and informality vary between doctors.
Not surprisingly, questions arose concerning the effectiveness of instilling compassion in medical students. Several skeptics concluded that such teaching may have unclear efficacy, but nonetheless was worthwhile, given the low cost and possible benefit. A surgeon commented,
I go back and forth: students are either going to have that sensitivity and treat patients in a kind and decent way, or not … It’s still good to teach them … but I wonder if it does any good.
Hence, given these potential barriers, others remained cynical about the ability to teach empathy. An internist commented,
Med school applicants want to work in soup kitchens all their lives, and practice urban-missionary medicine. But many physicians won’t get out of bed at night to see a patient. So, I’m a cynic. My partners … are compassionate, caring. I don’t know how well you can evaluate that. I say to medical students, “In terms of your interaction with people, you’re going to be where you already are. Do you treat people nicely? Then, that’s going to be the way you practice medicine. If you’re not, there’s nothing we can do to change that.”
Thus, he doubts that students can change, questions how potential improvement could be evaluated, if at all, and observes wide a priori differences among trainees.
Yet despite these questions about whether empathy can be taught, many doctor–patients nonetheless offered specific, explicit suggestions for improving both the process and the content of patient-doctor communication. Indeed, becoming a patient led many to treat their own patients differently, and to reflect on how to improve these relationships.
Improving care: process issues
Since “extra attention” was not always easy to provide, these physicians sought even small ways to be available and accessible to patients. For instance, one internist learned from his own physician a tactic for spending more time in patients’ hospital rooms:
One doctor used to take the chart into the patient’s room, and sit there writing his notes, rather than sitting out by the nurse’s station. That gave him an extra 4 minutes. While he was writing, he might ask questions. I’ve tried to incorporate that.
Many tried to adopt such techniques and transmit these to trainees as well. A pediatrician felt that patienthood made him more likely, before leaving patients, to ensure that he and trainees asked if the former had any questions. Another physician found that simply saying to patients, “I’m sorry about keeping you waiting,” could diffuse potential patient frustration.
As a result of their own experiences as patients, many physicians reported offering more explanations to patients, divulging clinical reasoning and uncertainties more explicitly. For instance, one physician stated:
I’m much more likely to explain why I’m doing things. Other physicians … don’t say what they’re thinking. I say, “I’m probably right, but may be wrong, thinking we ought to change to this …” I give much more information: my rationale, doubts: “Basically, this is a crap shoot. We don’t know what we’re dealing with here. We’ll take our best shot.”
Yet as patients, even these doctors at times got overwhelmed by having “too much” information about their own illness. As such, they raised questions as to the point at which information becomes “too much” for an individual.
Others tried to communicate better by providing information about what to expect from treatment, giving patients a “road map” to guide expectations. One internist said, “I say [to a patient], ‘An hour before you get this, the nurse ought to be giving you a pill make sure you get it.’” But improving communication depends in part on the particular patient. He cautioned, “You can do that with some, not all patients.”
Many of these physicians tried, too, to “follow-up with details more.” Though sensitive and attentive in the past, some now became even more careful that details did not “fall between the cracks,” developing special approaches or mechanisms for following up. One doctor–patient’s comments are revealing:
A nurse told me, “You have a special way of talking to patients. You come back—you don’t just leave them, or not show up. You listen.” I didn’t want to say, “It’s because I am a patient, what do you expect?” I really feel compassion. I can’t explain it.
This inability to “explain” the difference she felt after becoming a patient may reflect in part the fact that this interaction involves emotional, not just intellectual, connections.
Doctor–patients came to realize, too, the importance of nonverbal as well as verbal interactions. Rubbing a patient’s back, for instance, can help alleviate an existential sense of aloneness and isolation. As an internist said,
The nurse … giving a back rub was so incredibly important to me. It was profoundly human—an act of caring. Even with painkillers, there’s suffering and pain. Those back rubs were … somebody affirm[ing] that I mattered.
The importance of such nonmedical interactions surprised him.
Awareness increased, too, of the potential for relying more on non-physician providers (e.g., patient advocates or physician assistants) who had more time than did physicians and who could provide services for which physicians had inadequate time. Tertiary care doctors in particular could hire more nurse practitioners (as one doctor–patient put it, “to do the things other than fighting the fires, given the pressures and reimbursement”).
Improving care: content issues
Some doctor–patients also made suggestions for improving communication about several specific content-of-care issues.
As both doctors and patients, these interviewees described particular challenges posed by discussions about taboo topics such as mental health or sexual problems. Patients often appeared to want to talk about these areas, and expected doctors to be comfortable doing so. Among these physicians awareness increased, for instance, of the need for sensitivity in discussing mental health problems. One physician made this remark: “The doctor asked, ‘How are you?’ I said, ‘so-so.’ It took several visits for me to say, ‘actually, I think I’m depressed and need treatment.’”
Methods arose for discussing these issues with patients better. Concerning sexuality, for example, medical students could be taught to use more direct communication skills (e.g., asking not “are you married,” but “do you have a partner?”). In speaking about safer sex, these physicians suggested using blunt, not technical, terms (e.g., “sucking” instead of “fellatio”).
These doctors advocated less judgmental approaches for discussing taboos. An adolescent specialist admitted that he had dissembled to his doctors, and hence had learned to normalize taboo or unhealthy behaviors (e.g., nonadherence to safe-sex practices or taking medications) when talking to patients. He said,
I won’t ask, “Do you use condoms?” but, “Do you use condoms some of the time, never, or all the time?” If they want to say 100% of the time, I say, “Absolutely, 100% of the time? Most people slip up once…”
Several doctor–patients thought that issues relating to death and dying were still not addressed and taught sufficiently in medical school, and that physicians’ own psychological defenses could impede communication with patients. One physician observed that physicians tended not “to recognize that they, too, will one day face” these issues. Such denial had psychological and adaptive advantages—avoiding painful truths about disease and death all around them—but disadvantages, as well, as this respondent’s comment shows:
A lot of … medical schools address end-of-life. But it’s not done enough or well enough. [Doctors] think, “This is never going to happen to me. It happens next door.”
Suggestions arose that every resident should receive hospice training and be able to talk to dying patients about spiritual or other needs.
In a related vein, some physician-patients reported that their sensitivity increased to the need to present poor prognoses better, grounding these in a realistic sense of patients’ experiences. Several saw the delivery of medical data itself as a “talent”—especially as many patients misunderstood and “twisted” information, yet this task took dedication. As one respondent commented, “Information is good, as long as it’s delivered in a way the patient can understand. But many doctors don’t have that talent, so it’s useless.”
Many became more sensitive and timely about discussing Do Not Resuscitate (DNR) orders. Even many physicians who felt they were sensitive to end-of-life issues before their own illnesses changed, as this physician’s remark illustrates: “I realized how much more cavalier I was than I ever had imagined.” In part, they learned, though not from their medical training, the importance of tone, “gentleness,” and other subjective aspects of such communication.
Confrontation with their own mortality often made it easier for these physicians to confront that of others, and hence to discuss DNR for patients. An internist said,
I’ve dealt with my mortality. A lot of doctors are not good at dealing with code status because they haven’t dealt with their own mortality. When somebody becomes demented, it’s really easy for me now to draw the line.
Similarly, several came to alter their attitudes and conceptual frameworks in no longer viewing a patient’s death as a failure.
The phrase “there is nothing more we can do for you,” ought to be banned. There is a lot more we can do … even if we can’t cure your current physiological problem—talking to patients and spending time, as opposed to just doing procedures.
Given their own experiences with unpleasant side effects of treatment, many of these physicians had become less “hard” on patients concerning treatment problems, especially low adherence. As one internist commented,
I use my experience with difficulty with medications, side effects, keeping a schedule, and rationalizations. When someone talks about missing doses, I try to pinpoint what’s happening. For me, it’s easy on workdays, not weekends. Other times, [patients] just get sick of the schedule, so we talk about that. Or, they’re feeling sick, so don’t want to take a pill that makes them sicker.
Thus, these doctors have learned to focus on specific obstacles, rather than patient resistance in general.
Many of these doctors were startled at the arduousness of sticking to a schedule, and hence developed more empathy for patients.
Usually, if I’m busy, I say [to patients], “Take these pills 3 times a day. We’ll see you in 6 weeks,” and out the door I go to the next patient. I took more time with [a patient this morning], talking with her about how she might be able to get a pill box, carry her pills with her, time them to some schedule—try to take the pills with breakfast lunch and dinner. … She doesn’t eat breakfast. I said, “Maybe you ought to start, so you can remember to take your pills then.” I would not have done that had this all not been fresh on my mind as a patient.
Physician–patients’ recognition increased, too, of the importance of appreciating what patients were able to accomplish in adherence to treatment.
Now, I acknowledge very small steps—if patients don’t do 100% of what they’re supposed to, but just a little bit. In the old days, instead of saying, “You’ve really made some improvements,” I’d say, “… You need to work a lot harder.”
Several doctors said that after having been patients, they were “less likely to lecture patients” about other lifestyle changes. This respondent’s comment is illustrative: “Folks with diabetes have to make major lifestyle changes. I’m seeing how hard that is for me!”
These physicians suggested, too, specific methods of promoting preventative care—in other words, instituting the equivalent of “well-child preventative care” for adults. For instance, a doctor–patient recommended “Getting on an exercise program, worrying about your bones if you’re on steroids. For me, doctors weren’t thinking about it.”
Several physicians drew on their own experiences to teach patients means of recognizing and stymieing early symptoms. For instance:
Since I’ve had a lot of herpes myself, I tell people how to recognize it: that the first thing you get is tingling. I said, “Keep acyclovir in your medicine cabinet. As soon as you feel that tingling, pop 8 200-mg capsules, and see if you can abort it!”
This internist suggested that doctors who have not had certain conditions themselves may not be as attuned to early diagnostic signs.
Doctor–patients became more aware, too, of nonmedical aspects of care, such as whether patients could afford medications, had transportation home after a procedure or appointment, and had adequate home care. A radiologist said, “I’ve become like a social worker,” suggesting the degree to which physicians ordinarily view such duties as outside their role. Many of these doctors had previously overlooked nonmedical aspects of care (e.g., applying for government supplemental security income). As well, some physician-patients had begun to address, as they had not before they were patients, psychotherapeutic issues—for example, patients coping, and working toward “life goals.” One physician–patient commented, “Part of my drive to survive is wanting other people to survive—which I would not really think about before. Before I would have said, ‘Okay, do whatever you want.’”
Obstacles to improving communication
Yet the maintenance of such heightened attention toward patients often remained difficult. Despite their best intentions, many physicians slipped back into old patterns. Given competing time pressures, fostering empathy emerged as a daily challenge. As one physician said, “Sometimes I catch myself about to leave the room in a hurry, and ask, ‘Do you have any questions?’” After a long day, some doctors felt tired and “dismissed patients’ complaints.”
Several physicians tried to think more before speaking, to avoid potential insensitivity, but were not always successful. For instance, one respondent remarked, “I try and stop a millisecond in my thoughts before blurting out something, but I’m sure I’ve blurted out stupid things. It requires a lot of self-control. You always have to be careful of your comments.” In fact, colleagues denigrated efforts to take extra time, as this physician–patient’s comment illustrates: “Everything is driven by efficiency now. If a fellow is slow, it’s seen as bad, even if it’s because he’s taking more time with patients.”
Troubling trade-offs arose as a consequence of physician-patients’ heightened sensitivity, since spending more time with each patient resulted in seeing fewer patients overall. Hence, some had left, or spent less time doing, clinical work to devote themselves more to other activities such as patient advocacy.
A few doctors did not feel instinctively communicative all of the time, and thus had to force themselves to enter the role of “the caring doctor.” One respondent revealed,
Part of what we do is play-act sincerity. We are there to be sincere, but some is playing our roles as confidante and dispenser of wisdom, security and hope. I never thought of that as part of my persona.
Patients may also have unrealistic expectations or demands for time. As this physician–patient commented, “I would spend tons of time with people; and with some people, it just never was enough.”
Among the doctor–patients I interviewed, questions arose as to whether and to what degree empathy could be taught, as well as how to improve communication with patients. Simple processes were suggested such as charting at the bedside instead of after leaving a patient, asking patients if they had any questions at the end of interactions, acknowledging having kept patients waiting, and increasing awareness of nonverbal interactions. Sensitivity increased for directly addressing taboo content areas, including mental health, sexuality, end-of-life care, poor adherence, and nonmedical issues, but doing so was not always easy.
These data can help physicians be more aware of the impact of time constraints on patient behavior, attitudes, and care. By changing their behaviors according to these suggestions, doctors can optimize as much as possible the time they have with their patients. The doctors in this study learned only as patients much that they had not realized before—truly learning what it was like to be in the other’s role. Illness taught them what books failed to. The question naturally arises, then, as to whether doctors have to become sick to become more sensitive. Efforts to reform and humanize medical education cannot depend upon that, but must instead overcome critical obstacles—from competing pressures on doctors’ time, to physicians’ insensitive responses, to interpersonal stresses and demands.
Moreover, these doctors described a wide gulf between their intellectual and experiential knowledge. Experiential learning involves deep emotional layers of the self that book learning generally does not. Medical educators need to be more aware of and address these differences as much as possible. Indeed, the philosopher John Dewey articulated the needs to approach education as an active rather than a passive process and to take into account the characteristics of individual students.22 Clerkships on medical wards teach approaches to clinical reasoning, but empathy clearly remains more difficult to foster— especially passively. The data here also suggest that individual students’ openness to empathic stances can vary widely. Hence, teaching medical students a range of approaches and techniques, as outlined above, may offer several advantages.
This study has several potential limitations. The sample size may be small in comparison to some quantitative studies, though for a qualitative study of this nature, it is large enough to provide insights into patterns of issues that emerge. The data are qualitative rather than quantitative, but as such can shed light on a range of crucial issues that future quantitative research can explore in further detail with a larger sample. Many but not all of these doctors were HIV positive, but overall, differences in themes did not emerge between respondents with different diseases. Concerns about stigma were particularly prominent among those with HIV, but by no means unique, given that cancer and life-threatening illness in general are also often stigmatized and seen as taboo—even if at times to somewhat lesser extents. The issues that physician-patients identified in perspectives and approaches concerning the education of trainees did not appear to differ substantially based on their type of disease.
These data thus raise important issues for future training, clinical care, and research. Medical education can benefit from addressing more fully these cognitive and emotional aspects—among others—of differences between intellectual and experiential learning, and how awareness of these variances can best be used to enhance teaching about doctor–patient relationships and communication. Physician-trainees can be explicitly taught, for instance, to see more clearly that they approach patient interactions from a specific point of view; and that this view is just one of many and can differ from that of patients in a sometimes restrictive fashion. Patients may view risks, benefits, and clinical experiences in vastly different ways than do their physicians.
Heightened awareness of these gaps and immersion in the narratives and views of others can help. Requiring medical students to stay in hospital beds will probably not be incorporated into medical education. But trainees can become more aware of these issues through narratives such as those here that reveal patients’ points of view and that share these doctors’ unique experiences of seeing the world through others’ eyes. Narrative medicine has been proposed to help widen trainees’ points of view.9 These physician–patients’ experiences may in fact be more poignant and compelling to fellow doctors than those of lay patients in encouraging health care providers to be more open and sensitive to problems in communication with patients. Other physicians may recognize more fully that the boundary between physicians and patients is, in the end, nonexistent.
Several hypotheses emerge for future research. Teaching trainees specific skills concerning the content and process of communication (e.g., asking if patients have any questions at the end of interactions; trying to write chart notes at the bedside) may be more beneficial than attempting to instill empathy per se for many trainees. Randomized trials of educational interventions can be designed that assess and compare, before and after, patients’ attitudes, satisfaction, and clinical outcomes regarding these different kinds of physician interactions. Interventions can also target increasing direct communication about taboos such as depression or sexually transmitted diseases. The impact of such interventions on rates of mental health symptoms and sexually transmitted disease acquisition, respectively, could be investigated.
Future research can explore in further detail, too, how specific structural or other factors (e.g., gender or history of serious illness in oneself or close family member) may be barriers to or facilitators of physicians’ educability and understanding of patients’ points of view; and what aspects of patient communication different doctors or trainees most overlook. In several critical ways, the results of this study can thus help foster improvements in clinical care.
The author would like to thank Shaira Daya, Jesse Green, and Mary DuVernay for assistance with coding and other aspects of this project, and Renée C. Fox for her input on the coding process.