In recent years, many medical schools have implemented curricula in end-of-life (EOL) care, motivated in part by a survey in which graduating medical students and internal medicine residents described themselves as unprepared to care for dying patients.1 Since 1996, organizations including the Liaison Committee on Medical Education, the Accreditation Council for Graduate Medical Education, the National Board of Medical Examiners, and the American Board of Internal Medicine have mandated EOL teaching in medical school and residency and have incorporated EOL care competencies into standardized examinations. The majority of formal EOL teaching is lecture-based, elective, and occurs in the preclinical years of medical school.2
Studies about ethics and moral development have shown that students may suffer an “ethical erosion” during their clinical clerkships.3 Such findings correlate with students’ exposure to the “informal curriculum,” the education provided by residents and attendings as they model skills, attitudes, and behavior that may or may not be desirable for patient care.4 In a large survey of students, residents, and faculty at U.S. medical centers, Sullivan et al.5 described the informal curriculum and its influences on EOL care. While all groups endorsed the importance of EOL care, students lacked exposure to dying patients, residents felt unprepared to teach about EOL care, and attendings did not give feedback to trainees about EOL care. Similarly, a qualitative study about the clerkship experiences of medical students described students as uncertain and worried about EOL care, lacking guidance, and seeking increased experiential learning about EOL care.6
We designed this study to investigate how third-year medical students, already trained with preclinical EOL curricula, experience death and dying in their first clinical clerkship in medicine. Our research focused on three questions: What are students’ experiences with death and dying during the core medicine rotation?; What are students’ responses to these experiences?; and How does prior EOL education affect these clinical experiences?
Study design and sample
During the first rotation of 2002–03, one of us (NR), a trained moderator, conducted semistructured interviews about experiences with death and dying with University of California, San Francisco (UCSF) medical students rotating through the core medicine clerkship at three university-affiliated hospitals.
We chose a qualitative methodology to facilitate open descriptions by respondents and uncover themes that the researchers may not have anticipated. Rather than using standardized questions with specified responses as in quantitative research, one-on-one interviews allow respondents to reveal their perspectives in their own way and in greater depth.
We selected respondents through purposeful sampling, targeting 32 third-year medical students at UCSF as they rotated through their first core internal medicine clerkship. The first clinical rotation of the third year presents students with their first clinical experiences with dying patients after completing the formal EOL curriculum.
At the UCSF School of Medicine, students participate in a second-year course called Foundations in Patient Care, which includes a seven-week module on EOL care. These four-hour weekly sessions combine lectures on pain control and cultural differences in EOL care, small-group discussions about personal experiences with death, one-on-one mentoring by clinician preceptors, visits to hospice, and storytelling by patients.7 Over half the students also participate in a preclinical elective “Healer’s Art,” a 15-hour, three-month course about grief, mystery, awe, and meaning in medicine. Finally, the third-year core medicine clerkship includes a two-hour standardized patient exercise about durable power of attorney, advanced directives, and wishes for emergency cardiopulmonary resuscitation or “code status.”
The medicine clerkship is an eight-week inpatient rotation in which students join teams consisting of an attending, a resident, one or two interns, and a subintern. Every four days, a student completes a history and physical examination on one newly admitted patient, and the student follows this patient daily throughout the patient’s hospital stay. The sites for the clerkship are a university tertiary-care hospital, a county hospital, and a Veterans Affairs medical center. In this period, students transition from classroom-based learning to experiential learning through patient care and have their first opportunities to apply any acquired skills and knowledge in EOL care.
In August 2002, one of us (NR) conducted semistructured interviews with consenting respondents. NR did not teach or evaluate students in the clerkship, and recruiting efforts via e-mail, poster, and oral announcements emphasized that participation was confidential and would not affect course grading. NR visited each hospital on advertised days, recruiting students between conferences and patient care activities. We offered two movie gift certificates as incentives. The UCSF Institutional Review Board approved the study.
The interview instrument included open-ended questions to elicit students’ experiences with dying patients, reactions to these experiences, coping strategies, and opinions of the value of prior preparation in EOL care (see Appendix). NR, who had received training in qualitative interview techniques for undergraduate thesis research, used additional reflective statements and probing questions to clarify and expand on concepts raised by the respondents. Conducted in private conference rooms in the hospital, interviews were audiotaped and transcribed verbatim by a professional transcription service. NR compared all transcripts for accuracy with the audiotapes prior to analysis.
We used a grounded theory approach, in which a theory to explain a phenomenon is derived from the data itself, rather than testing a predetermined hypothesis. This analysis is iterative, requiring frequent revisions and testing of the theory as it emerges.8 Using ATLAS.ti software (ATLAS.ti GmbH, Berlin, Germany), each of us independently generated from six randomly selected interviews a list of “codes,” the descriptive words or phrases that summarize respondents’ statements. After comparing codes and negotiating discrepancies, we refined and grouped the codes into larger themes—a process termed axial coding. We each independently applied these themes to 12 more interviews and compared transcript analysis. If we disagreed about the placement of a theme, we all discussed the theme until reaching concordance. One of us (NR) applied the final themes to the remaining transcripts, from which a randomly selected subset was independently analyzed by two of us (KH, AT) to ensure consistency. We then generated a theory from these themes (see Figure 1). Finally, we tested the theory against original transcript data, actively conducting discrepant case analysis to ensure that the final model fit the data.
Twenty-eight students (87.5%) participated in interviews ranging from 15 to 60 minutes in length. Four students declined to participate in the study for unspecified reasons. Participation from each site was consistently high, ranging from 83% to 92% of potential respondents. Respondents’ characteristics are shown in Table 1.
All respondents stated that they observed members of their medical teams caring for dying patients. Twenty-four of 28 (86%) respondents participated directly in the care of at least one dying patient. Many students were surprised by the amount of death and dying in the hospital and by the acuity and rapid progression of some patients’ diseases, as one student revealed:
I have seen a lot of death and dying around the hospital—more than I expected to, which is kind of a funny thing to say in a hospital, but I guess you have images of people coming in and getting better and going home.
Themes emerged around four related areas: relationships with patients, educational and personal background, interactions with their teams, and development of professional identity. The frequency of codes in the above themes is shown in Table 2.
Relationships with patients: intense connections
Compared with other team members, students had fewer responsibilities and more free time, affording the opportunity to develop close relationships with their patients and patients’ families. This connection was seen as a new privilege, as patients allowed students into their lives. A student commented:
I got to spend the most time with her, for sure, of anybody on the team, just because I was following fewer patients and it was just neat to get to know somebody.
Students’ empathy with patients affected the degree of this attachment and their reaction to patients’ deaths. Students identified with patients in their age group and often compared their older patients to their family members. One student said:
It made it even harder to deal with … the fact that one patient I had was the same age as me.
They also empathized with the grief felt by family members, as one student explained:
The hardest part was watching the family and kind of understanding their loss, and what it meant to have their father gone.
This empathy also extended to patients with whom students had no direct relationship. Students described sadness evoked by reminders of death that are part of daily life in an acute care hospital: an intern delivering a poor prognosis to a patient during morning work rounds, a code blue run by a resident, an intubated stranger in the intensive care unit, or a corpse being wheeled down the hallway.
In their relationships with patients, many students described their role as an advocate, motivated by an intense desire to help their patients in any way possible. In the words of one student:
There was only so much we could do. But when I would be up at night, I was like, “Could I bring this man chocolates in the morning? Could I do little things that would make this better for him?” Beyond my limited ability to act as a physician for him, medically-wise. Just to be a friend to him.
Students sought to provide care consistent with their patients’ goals and values. They tried to facilitate communication between their teams and their patients, at times mediating differences in their perspectives. As advocates, students also learned to view certain deaths as “good deaths,” conducted in the manner desired by patients and their families. One student remarked:
[She] chose to take off the mask at the end, and she did it. She died on her own terms and her closest family members were there.
Students’ educational and personal background: limitations of the foundation
Students expressed ambivalence about preclinical preparation for EOL care. Most stated that preclinical teaching provided a conceptual framework that helped them approach patients. However, all students stated that didactic curricula in EOL care exert less impact than do mentored experiences during the third year. Patients motivated students to learn and apply concepts in a practical and realistic way, and according to students, this direct relevance made them more receptive to EOL curricula than they were prior to the clerkship. One student declared:
I don’t think you can be taught these things … it’s good that we are presented with those things [in preclinical curricula] because if we weren’t, then we’d be totally unprepared. But, I really think 90% of the learning happens personally.
Four (14%) students had extensive experience with death, either in prior careers or within their families. These students were older and more familiar with family discussions, grief counseling, and palliative care. They tended to be more comfortable working with dying patients during the clerkship. One student emphasized:
It’s important not to emotionally overprotect people … most of us, by the time we reach medical school age have had some sort of experience with somebody’s death or somebody’s being very ill.
However, even those without personal experiences with death and dying described how sources outside their medical education—family, friends, ethnic background, and religion— provided the humanistic values and tools that informed their approach to patients.
Students described personal coping mechanisms that they brought with them into the rotation. Exercise was mentioned most frequently among the nonverbal strategies. Some students also cited writing, music, therapy, and prayer. Time pressures and fatigue kept some students from pursuing these activities as often as they had prior to the clerkship.
Students often coped by discussing their experiences with family or nonmedical friends. These conversations afforded opportunities to voice opinions and feelings without fear of academic evaluation. In addition, nonmedical friends provided an alternative perspective untainted by the medical system. However, most students realized that nonmedical confidants could not easily understand their stories, as one student relayed:
I feel as though my days are very full of these kinds of intense moments…. I think the intensity doesn’t always translate very well into an anecdote.
To satisfy this need for understanding outside of the team but within the medical sphere, students sometimes turned to one another for support. A student commented:
[Other students] are going through the same acculturation process and frequently having similar experiences—or dealing with it totally differently.
Team interactions: role modeling responses to death
Students described their experiences with dying patients most vividly within the context of their medical teams. They observed their team members’ attitudes and behaviors and their reactions to dying patients. The response described as most fundamental to students was the team’s acknowledgement of a patient’s death through explicit demonstrations, such as discussion, moments of silence, or poetry readings. One student revealed:
Rounding on somebody who died … helps bring closure to taking care of that person, instead of just dropping them off the list like somebody who had been discharged, we just fold up the card and let them go.
Students wanted to know that their residents shared their view of a patient as a person, not just a disease. They compared their responses to those of their team members to determine whether their own feelings of sadness, shock, or grief were attributable to inexperience or shared by more experienced clinicians. One student observed:
[The team members] talked about [the death] to me individually … it showed me that they were genuinely concerned … they expressed their own feelings about it too … it’s not like them saying, “Well, you’ve got to deal with this because you’re new”. … “It’s still kind of hard for all of us.”
During attending rounds or Morbidity and Mortality conference, students also learned to process the medical reasons for a patient’s death. When a death was unexpected or a patient decompensated despite aggressive efforts, these settings helped students resolve feelings of guilt and provided closure, as one student recounted:
They came and talked to us about oncology emergencies and what to look for. And I thought that was helpful, kind of like a little bit of a resolution.
Students also noted times when explicit acknowledgement did not occur. Implicit acknowledgement, through nonverbal cues signaling sadness or other emotions, helped some students process their own reactions. However, medical discussions devoid of emotion made students feel isolated and frustrated. One student declared:
It would be great if once in a while we just dropped all the medical side of it and just sort of talked about what this person was going through.
Students also noted systemic factors contributing to a resident’s inability to process these experiences with them. Time pressures and large workloads required residents and attendings to prioritize their care of other patients. A student noted:
We have a lot of patients and they are all very sick, and this is just this one experience. … Each of us have been dealing with it in our own ways, but we haven’t really mourned her, her passing. It was brutal. I mean, so sad.
Although some attendings and residents taught students didactically about EOL care, most students cited role-modeling as the primary mode of teaching. In their discussions with the team and with patients and families, attendings and residents modeled communication skills, behavior, and attitudes in EOL care. A student stated:
The most important thing was the leadership at the top, was our attending and how he dealt with these issues. I think that if he had been very clinical and detached, it would have signaled to me … that’s my role model, that’s how I’m supposed to act.
Role-modeling was particularly helpful for students learning about the transition to comfort care. Residents and attendings validated the importance of care rather than cure and redefined successful interventions in the context of dying. In the words of one student:
One of my interns pointed out [the patient and family] had come to the hospital with nothing and really they were leaving with something. We were going to get them plugged into palliative treatments and a sense of support they didn’t have. That was really perceptive … that we had actually offered them a great deal, even though we weren’t able to offer them cures.
In the few instances when students mentioned negative role-modeling, they described residents who appeared insensitive to the patient’s personal experience of illness. As one student claimed:
The only thing I ever heard from anyone was, “Such an interesting case, if you don’t love this, you don’t love medicine.” I was like, this poor woman’s going to die, why is everyone so excited? I felt that was just creepy.
Whether speaking generally about the rotation or more specifically about their experiences with dying patients, students wanted to contribute in valuable ways to patient care. Teams that discouraged active student participation generated frustrated students who felt disempowered in their low position in the medical hierarchy. Students emphasized that their level in school should not label them as incompetent in the humanistic aspects of clinical care. In the words of one student:
There’s a certain overprotection of med students … an assumption that, because we don’t know how to analyze an ABG, we lack a certain emotional maturity.
Those with less confidence in EOL skills stated that they would have appreciated invitations to family meetings about goals of care. Students who did participate actively in these challenging conversations felt more confident about their skills and more pride in their impact on their patients. As a student noted:
I think that one of the strengths of my team is that they really made sure that I was there for every discussion. I really did feel involved and I felt like my presence was valued … by both my team and by the family.
The emergence of professional identity
Students’ relationships with patients, their background, and their interactions with teams combined to influence students’ developing sense of professional identity. Through dealing with death and dying patients, students learned to control and express their emotional responses, understand the transition from student to resident, and feel self-efficacy in their care of patients at the end of life.
Students interpreted these experiences as lessons in how a professional expresses emotional responses to patient care. Students struggled to balance the perceived need for objectivity with their tendency to identify with patients and react emotionally to their deaths, as one student expressed:
I would like to be friends with my patients … but I also need to be objective and somewhat distant … to have my decision making be guided by sound, clinical judgment.
Students looked to their residents and attendings to model appropriate behavior. While some team members demonstrated that emotions should not be felt or expressed in the hospital, others expressed and openly acknowledged their emotional reactions. One student explained:
There’s that subconscious idea that maybe you should be professional, or tough. Or, you look and see, “How are they acting? I should kind of act in the same way.”
Another student agreed:
When my first patient died, my resident—who I just think was totally great—he had tears in his eyes [which said to me], “Everyone goes through the emotions I’m going through right now, whether it’s your first patient or your hundredth.” He was very professional, very informative, and very helpful to the family. But he was emotionally attached to what was going on, and I think that that was really key. I guess that kind of changed my idea of being professional.
Students recognized that the clinical training years represent a transition period. Some students noted that their initial lack of medical experience helped them step out of the role of medical provider and know the patient intimately, as one student recalled:
That was my very first rotation, so I felt like I had one foot in the lay person’s world and one foot in the medical world. It was also kind of tension-filled for me because it felt like putting some distance between myself and the team.
However, students worried that their capacity for caring about patients would diminish when they became residents. When residents and attendings did not acknowledge deaths, students blamed time constraints, sleep deprivation, and heavy patient loads. One student remarked:
If I spend a half hour in the room with my patient, it’s scary to think what my intern or resident or attending is spending with that patient, because they have less time than I do, and they have more patients than I do.
They also viewed detachment as a method for self-preservation. One student quoted her resident’s advice:
I think it’s really a balance … of being emotionally open, and allowing yourself to connect with people and not letting your emotional attachments or your grief for the really sad stories … keep you from doing the rest of your work. So, I think it’s not inappropriate that I got teary, but I’m glad that I’m sad, ’cause if I weren’t, then I’d have to start to worry about myself.
Students themselves were not immune to this detachment. One student emphasized that sadness is not always the primary emotion after a death. For some students, recognizing their own numb reactions provoked feelings of guilt and concern, as one acknowledged:
I’ve been surprised at how well I’ve been able to leave things here … sometimes it makes me concerned because I can really separate myself already.
The manner in which teams responded to death influenced students’ confidence about their ability to care for dying patients. When patients decompensated acutely in the face of aggressive curative efforts, some students shared their teams’ sense that death is a negative outcome resulting from errors in judgment or skill. One student reflected:
My resident just felt, “Should I have pressed the oncology team to really make sure what this was?” … The intern was kind of feeling the same way: “What should I have done?”… It sort of trickles down to me, and I’m feeling maybe I should have read more about what was going on.
However, some team mentoring helped students reframe the dying process as a potentially empowering experience for patients, families, and physicians. By working with patients to meet their goals at the end of life, students were able to experience the underlying principles of EOL care: the differences between care and cure and the importance of accompanying patients through the process. A student observed:
The sadness really was only there for like two or three hours, and then it was just kind of replaced. … She died in the way she wanted to go. She died with people who loved her and she really died in very little pain … so it’s kind of a blessing … in my head it switched so I was able to look at her death as a good thing.
This maturation in perspective decreased students’ death anxiety and increased their sense of professional competence. Students who actively participated in their patients’ care gained confidence in their own EOL skills, as one student noted:
It wasn’t a depressing experience at all. … It was incredibly sad. But I just remember feeling like it was incredibly profound. We had really done something that was going to help this man to face the end of his life—hopefully feeling a little safer and a little more cared for than he would have otherwise.
Based on the themes above, we derived a model for the emergence of professional identity of medical students as shaped by their relationships with patients, educational and personal background, and team interactions. (See Figure 1.) We also developed from the students’ perspectives a list of practical recommendations for residents and attendings when responding to a patient’s death. (See List 1.)
Third-year medical students have notable experiences with dying patients that affect their professional development. Our findings emphasize the importance of experiential learning, with mentored clinical encounters and active participation in patient care. This study also demonstrates how the informal curriculum—role-modeling by residents and attendings—affects not only the acquisition of skills but also the molding of behaviors and attitudes that influence greater questions about professional identity.
These experiences reveal a model for medical students’ professional development during the clinical clerkships. (See Figure 1.) Students had the opportunity to form close relationships with their patients, and students’ personal and educational backgrounds permeated their perceptions of clinical experiences. Students reacted to these experiences within the context of a team, seeking mentoring from a supportive group that acknowledged their experiences and reactions, role-modeled good EOL care, and respected the students’ participation in patient care. These factors—patient relationships, background experiences, and team interactions—thus shaped the students’ sense of professional identity during the crucial transition from classroom to wards. The critical incident of working with dying patients raises dilemmas: How do I empathize with patients but maintain control over my emotions? Will I lose my ability to care for patients when I become a resident? Can I feel competent by offering a good death to my patients even if I can’t provide a cure? Reflection on and resolution of these dilemmas are central to the development of professional identity9 and may influence how students view their professional role toward other patients, not just those at the end of life.
In our study, the experiences of positive role-modeling differed from the experiences in previous literature which described predominantly negative behavior by residents and attendings.3,5,6 Our more optimistic findings may stem from the students’ recent introduction to the wards or their lack of exposure to the other clerkships where EOL care is less established. They may also represent the influence of a successful palliative care service on the residents and attendings in the Department of Medicine.10 Finally, early in the academic year, residents may have higher enthusiasm for teaching, but this may be balanced by less experience in EOL care or less efficiency in managing high workloads. Thus, these positive experiences may reflect only a time-limited and selected part of the informal curriculum, which could be captured in a more longitudinal study following trainees’ responses to dying patients and professional development over time.
Our study had potential limitations. We did not collect data from residents or attendings to corroborate students’ perceptions of their experiences. Although we included three hospitals with varied patient populations, this single institution study may have limited generalizability to other medical schools. We did not collect demographic information about the nonrespondents to determine differences from respondents. Finally, because the interviewer was a resident in the internal medicine program, students may have modified their responses, limiting stories the students perceived to be negative or critical.
Our study’s strengths include a large sample size for a qualitative study and a high response rate from all three sites. We had high intercoder reliability and reached thematic saturation. Comparison with other literature on ethical dilemmas and professionalism indicates a similar influence of critical incidents in the clinical clerkships on students’ moral and professional development.11–13
Our findings illustrate how experiences with death and dying affect medical students’ emerging professional identities. Our study supports efforts to improve residents’ education about the core competencies in EOL care and to promote students’ reflection about the informal curriculum in their ward experiences.14,15 In addition, it lends support to calls for faculty development in EOL care given its potential to affect undergraduate clinical education.16 Future research and interventions should focus on improving mentoring relationships on the wards, helping residents and attendings to recognize the impact of their role-modeling and to guide students through the important clinical challenges that shape professional identity.
The authors would like to thank Kathleen Kerr for manuscript editing and review. Funding for this project was provided by the Office of Medical Education, University of California, San Francisco, and U.S. Health Resources and Services Administration Grant #2DEPE19179 Residency Training in General Internal Medicine.