Discussing Physician-Assisted Dying: Physicians’ Experiences in the United States and the Netherlands

Abstract

Purpose of the Study: This study aims to further our understanding of the experiences of physicians when discussing physician-assisted dying (PAD) within the context of doctor–patient relationships in various sociolegal settings. Although patients bring up PAD in diverse settings, little is known about how physicians experience these discussions, and such experiences have not been directly compared across countries. Design and Methods: A total of 36 physicians in the Netherlands and the United States (including Oregon) were interviewed using individual semistructured interview guides. Interviews were conducted by a single interviewer, audiotaped, and independently transcribed. Inductive qualitative analysis, aided by NVivo7 software, directed purposive sampling of physicians until saturation was met. Multiple coders in a multidisciplinary team analyzed emerging themes and developed theory. Results: PAD discussions, which ranged from theoretical discussions to actual requests, could be gateways to discussions of other end-of-life issues important to patients and could strengthen doctor–patient relationships. Physicians found discussions to be emotionally intense, but often rewarding. Where PAD was legal, physicians utilized existing criteria to guide communication, and discussions were open and honest with patients and colleagues. Where PAD was illegal, conversations were less explicit, and physicians dealt with requests in relative isolation. In addition, physicians’ views of their professional role, patients’ rights, and religion influenced both their willingness to have and the content of PAD discussions. Implications: Discussion of PAD is an energy-consuming, yet potentially enriching part of the doctor–patient relationship. Legal guidelines may help to provide structure and support for physicians when patients broach the topic of PAD.

Physician-assisted dying (PAD), which encom- passes both voluntary active euthanasia and physician-assisted suicide, has been debated for millennia, and as a result of distinct cultural and social discourses (Emanuel, 1994; Foley & Hendin, 2002; Quill & Battin, 2004), countries have developed different legal frameworks around PAD (Battin, van der Heide, Ganzini, van der Wal, & Onwuteaka-Philipsen, 2007). At the time of this research, Oregon was the only state in the United States to have legalized physician-assisted suicide through the “Death with Dignity Act” in 1997, whereas the Netherlands had legalized voluntary active euthanasia and physician-assisted suicide under specific criteria in the “Termination of Life on Request and Assisted Suicide Act” in 2002. Although the legal framework around PAD may differ, patients nearing the end of life in both countries—and throughout the world—bring up the topic and request help from their physicians (Emanuel, Daniels, Fairclough, & Clarridge, 1998; Meier et al., 1998). Where it is illegal and even where it is legal, PAD is discussed between physicians and patients more frequently than it is actually performed (Ganzini et al., 2000; van der Heide et al., 2007), and the discussion itself can serve a function in patient care (Norwood, 2005).

Physician–patient communication and discussions around end-of-life issues have been recognized as key components of quality end-of-life care (Doyle, 2004; von Gunten, Ferris, & Emanuel, 2000). Building on the conceptual framework that PAD discussion is a distinct entity and that physicians’ experiences of PAD discussion in different geographic and clinical settings needs to be further understood, the aim of this study is to elucidate how PAD discussions transpire within the context of doctor–patient relationships and what influences physicians’ experiences of PAD discussions in the distinct cultural and legal settings of the Netherlands and the United States. Previous studies within the Netherlands and the United States have explored aspects of the experience of physicians in individual settings when dealing with discussions around PAD (Borgsteede et al., 2007; Dobscha, Heintz, Press, & Ganzini, 2004; Georges, The, Onwuteaka-Philipsen, & van der Wal, 2008; Kohlwes, Koepsell, Rhodes, & Pearlman, 2001). Other studies have utilized the United States and the Netherlands for comparison regarding end-of-life decisions (Battin et al., 2007; Helton, van der Steen, Daaleman, Gamble, & Ribbe, 2006; Willems, Daniels, van der Wal, van der Maas, & Emanuel, 2000). This is the first study to directly compare physicians’ experiences of PAD discussions across these different environments in order to elucidate how themes are either consistent or nuanced and to build theory about the phenomenon that triangulates input from three distinct geographical areas.

Design and Methods

Qualitative methods, which allowed the respondents to express their experiences in their own words and convey their specific context, were utilized for this sensitive topic. Ongoing inductive analysis of completed interviews informed data collection and sampling.

Data Collection

In 2007–2008, one-on-one, semistructured interviews (McCracken, 1988) were conducted by J. R. Voorhees with physicians in the Northeast United States, Oregon, and the Netherlands (see Supplementary Appendix A for interview guide, which was piloted and adjusted with feedback from initial participants). Purposive sampling was utilized to seek out diverse physicians with potentially different experiences with end-of-life discussions. The snowball technique (Kohlwes et al., 2001) allowed respondents to suggest other physicians who may have had different experiences or viewpoints during the wrap-up portion of the interview. Physicians with a mix of specialties, gender, age, experience, and clinical setting were purposively chosen to create a diverse group of participants to seek a wide range of perspectives. The recruitment was directed by ongoing analysis of completed interviews, and interviewing ceased in each geographic area when the research team determined, through analysis of the completed interviews, that data saturation was achieved.

Physicians were initially contacted via e-mail or telephone. Interviews were done in-person at physicians’ offices or alternative places of their choosing and lasted 40–70min. Interviews were digitally recorded, transcribed by an independent medical transcription service, and checked for accuracy by J. R. Voorhees. Notes were also made after interviews, and J. R. Voorhees observed physicians in relevant clinical settings to understand the context of the experiences described. The study design was approved by the Human Investigations Committee of Yale University School of Medicine.

In all, 36 physicians were interviewed from a variety of clinical settings and specialties (Table 1).

Data Analysis

Multiple research team members, including a geriatrician (M. A. Drickamer), a psychologist and health services researcher (J. A. C. Rietjens), and a medical student (J. R. Voorhees), contributed to the inductive analysis in which themes were elucidated (Bradley, Curry, & Devers, 2007) and further theory was built (Glaser & Strauss, 1967). The five-step framework approach—familiarization, identifying a thematic framework, indexing, charting, mapping, and interpretation (Pope, Ziebland, & Mays, 2000)—was modified for this analysis. Cleaned interview transcripts were shared with the research team as they were completed, and inductive analysis began with familiarization of the interview content. This process allowed for inductive identification of themes to be built into a framework, which—after many iterations with subsequent interviews—became the final coding tree. The final coding tree (see Supplementary Appendix B) was then independently applied to all interviews by M. A. Drickamer and J. R. Voorhees. Through review, it was determined that there was a high level of consistency between coders, and discrepancies were discussed and resolved through a tendency toward inclusivity. The consensus code and respondent demographics were applied to transcripts in QSR NVivo7 qualitative software to assist with further analysis (Bazeley, 2007). Themes were then able to be analyzed by “pulling” all text associated with a code for an in-depth examination for patterns and exceptions. Theory was built by pulling areas of overlapping codes, which allowed for examination of the interaction of different themes.

This analytic approach encompassed several criteria important for ensuring rigor and validity in qualitative analysis: reflexivity, triangulation, fair dealing, attention to negative cases, and clear exposition of data collection and analysis (Mays & Pope, 2000). Throughout the research process, the research team openly discussed how the researchers and research process shaped the collected data, and J. R. Voorhees performed formal written exercises of reflexivity.

Results

PAD Discussions With Patients

Physician respondents described a range of experiences of PAD discussions with their patients across the various legal and clinical environments of the Northeast United States, Oregon, and the Netherlands. Discussions ranged on a spectrum from theoretical questions about PAD to actual requests for PAD (see Figure 1), often depending on the state of health or stage of illness of the patient. Physicians in all three geographic regions noted that the discussions were opportunities to address patient concerns, explore patients’ fears of dying, clarify information about symptoms control, and reassure patients that they would not be abandoned.

I’m happy that it’s open to have a discussion about it . . . It’s bringing up emotions but it also opens up the possibility to discuss with the patient, with the families, and then to talk about death and dying. Netherlands/4

When patients bring up questions about physician-assisted death, [they] are also opening the door to discussions about their care and about their symptoms and about their situation . . . “Why is this coming up right now? What’s been going on at home? How’s your pain?” US-Northeast/1

Ask why someone asks that question at that moment . . . what a person is afraid of . . . what makes him suffer . . . “Yes we can discuss that, but . . . why do you ask that? What are you afraid of? How do you think your last phase of life would look . . . How do you feel about that?” Netherlands/10

Though PAD discussions occurred in all three geographic areas of study, physicians in areas where it was legal had substantially more experience discussing PAD with patients.

We are in Oregon; Death with Dignity—it’s legal here, and so I get questions from patients or families about it all the time . . . It’s not a topic that I avoid . . . If I don’t give them the time that they need to address those issues and concerns, then I am doing them a disservice. US-Oregon/24

In the Northeast United States, where PAD was not legal, the topic was less expected by physicians, and they were not as prepared to discuss the topic in detail.

It took me by surprise . . . I suppose in retrospect . . . anybody with that disease [ALS] . . . would be thinking about it. US-Northeast/2

And in those instances I declined to give the prescription but I don’t recall any real extensive discussions about it. US-Northeast/1

Respondents described that when a PAD discussion occurred, the topic was generally initiated by patients. However, some physicians in Oregon and the Netherlands felt comfortable initiating discussions about PAD because they realized that patients had heard of these options and felt an obligation to have patients understand how PAD could or could not fit into possible treatment options.

I actively discuss options like palliative sedation and euthanasia, just for patients to know that there is a choice, because it’s no use not discussing it when the patient is thinking about it all the time . . . What I more and more learn is: actively ask if patients want to discuss it and see what kind of ideas they have for their own. Netherlands/12

Physicians in Oregon and the Netherlands described that the legal regulations helped to shape and guide their discussions.

The forms themselves have some really great and clear questions that are not a bad template for discussion. “Have you informed your family members? Do you know you can change your mind at any time? . . . Do you know you have other options in pain control? Is there any sense that anyone is pushing you in this direction or forcing this choice upon you?” US-Oregon/23

Doctor–Patient Relationships

Physicians described that previous relationships with patients affected and were affected by PAD discussions (see Figure 1). Strong doctor–patient relationships could facilitate the occurrence of PAD discussions.

Because you know the patient very well; you see him very often . . . You grow with the problems of the patients . . . I think you only can do this type of communication with the patient . . . when you know him very well. Netherlands/5

That was . . . why she asked me to do the euthanasia because . . . I knew her. I knew . . . she suffered that much already . . . “You know . . . I have suffered and just I don’t want this anymore,” and then she didn’t have to tell me more because I knew her for so long . . . five years . . . and I knew so much about her life story that I knew that this was too much for her to bear. Netherlands/7

The more . . . you make it clear that you are comfortable talking about end-of-life issues and the more . . . that you give them to understand that you are available, then they—in some . . . unusual circumstances—may ask whether you will help them die . . . I have had that very clearly twice. In both instances these were elderly women, who I had gotten to know very well over years, and they were utterly competent and straightforward in wanting help in having the option to end their lives. US-Northeast/27

Physicians described that the experience of the PAD discussion often strengthened the relationship as the conversations were intense and revealing of patients’ values and goals.

One develops an intimacy . . . because you’re asking them intimate questions . . . and so . . . you get to know . . . about their families, about their family belief system, about where they grew up, about what sort of influences affected their lives . . . you learn about them. US-Oregon/22

It’s a strange topic to talk about but it’s a way to come close to the patient . . . You really get to know what this person thinks about his or her complaints. Why, why is this person considering this symptom as too painful or too distressing to live with? And then you get to learn about the real motivations of the patient. Netherlands/15

I probably had the most intense relationship with this patient than any other patient because you are really taking time to discuss with him . . . you go much deeper into his feelings and his ideas. So, this was actually quite intense. I think, from both sides, it felt good. Netherlands/19

However, exceptions occurred in actual PAD requests if the physician and the patient did not agree about the appropriateness of PAD.

The moment I saw her I knew that she wasn’t competent . . . she was just fiddling with the pills. So I had to tell the family that I couldn’t allow her to take the medicines. And they were absolutely thunderstruck, “What nonsense is this?” US-Oregon/23

[The] patient was very disappointed and also the husband and the son was very disappointed, but I didn’t see any possibility to do it. So I said, “No.” . . . You have to disappoint people even in the end stage when they’re very ill . . . unfortunately. Netherlands/3

If there was a strong prior relationship, physicians felt that refusal of an actual PAD request did not end the relationship but sometimes left a strain.

I’ve never had a relationship end over this issue . . . When you’ve had a long-term relationship with someone, there’s enough trust there and there’s enough respect there that people are willing to respect your point of view. US-Northeast/30

Where there was not such a prior relationship, physicians described that a PAD discussion with differing physician and patient viewpoints could be more destructive.

The conversation with him was brief and kind of brutal—brutal coming from his side, because I continued to try and make contact . . . That was right at the beginning . . . while I was doing the history and physical . . . As soon as I told him I couldn’t do it, he broke communication with me, wanted nothing to do with me, was barely civil for the rest of the time he was under my care. US-Northeast/33

Physician Emotions

Physicians experienced intense emotions associated with these discussions. Emotions described were both positive and negative, and were at times conflicting as they could occur simultaneously. Positive emotions included the rewarding feeling of being there for someone and the relief of being able to talk openly with them.

Those two patients were very, very, very determined in their wish and really were grateful that I was willing to talk to them about it and was willing to consider it . . . I thought it was very rewarding. Netherlands/15

It was just right out there, “Is there a way to get this over with quicker?” As opposed to . . . the fact that probably everyone thinks about it, just doesn’t bring it up. So on one hand it was actually liberating to be able to put it right on the table instead of wonder what he was driving at . . . It was the directness of it that made it more remarkable. US-Northeast/32

Negative emotions included fear of or discomfort with the topic of PAD as well as the emotional exhaustion such an intense experience can bring.

But it was really hard [when the patient brought up PAD] . . . it was like a knife in the heart . . . I think I was shocked. US-Northeast/34

In our hospital, once you are actively involved in a euthanasia question, for the rest of the year you think, “Well please let someone else do this question.” It is not once you have crossed the line, it gets easy, and I think that is a good thing. Netherlands/16

Where PAD was not legal, there were feelings of guilt for not considering participating in PAD or of fear because of the risk involved if they did consider it.

This feeling that maybe I was letting him down . . . given the circumstances of ALS . . . It’s not an unreasonable thing to think about, but the way things are now because of what’s legal and what’s not . . . I wasn’t comfortable promising him that I could, that I would do that for him. US-Northeast/2

And I think . . . particularly in the United States, we’re always fearful of how we’re perceived. And there are doctors who probably wouldn’t want to be known for their euthanizing people—we’re cancer doctors, and we want to cure cancer. It’s not like we want to admit defeat in curing cancer. US-Northeast/36

He and his wife asked me to help them. Literally I felt as though the blood had frozen in my veins. I just felt just totally cold all over. I had no idea of what to do. I realized there was no help I could get from anywhere . . . I . . . felt as though I was . . . impotent to help them. US-Oregon/22 (before Death with Dignity Act)

Physicians described that when they considered actually participating in PAD, their emotions became more intense.

I was busy all the time with the question, “Is it possible, and is it morally defendable, and is it a good decision?” . . . making pluses and minuses about it . . . but . . . “What’s it doing to me? I’m going to kill someone tonight.”[respondent began to cry]. Netherlands/11

Discussions With Others

The degree to which physicians discussed patient interactions about PAD with others differed between the Northeast United States, Oregon, and the Netherlands (see Figure 2). When physicians did discuss these interactions with others, they found it to be helpful.

In the Netherlands, extensive discussions with other practitioners including physician colleagues, nurses, rehabilitation therapists, and chaplains were inherent to the process.

It starts with the patients, the family, the nurses . . . in the nursing home, the other disciplines, therapists and that’s part of the process . . . some are close to the patient. In one case it might be the physiotherapist. Other times . . . someone who does the activities or . . . the [chaplain]. Netherlands/6

In addition, the Netherlands requires an official second opinion, usually done by a Support and Consult Euthanasia Network (SCENarts) physician with special expertise in the legal requirements for euthanasia in the Netherlands. In Oregon, physicians discussed their PAD discussions with others including their physician colleagues, their own personal contacts, and occasionally nurses.

In Oregon and the Netherlands, physicians described their discussions with others as helpful by empowering them to decline participation in PAD with a particular patient.

I discussed it with the peer group and . . . it helps because they say, “Yeah we are with you when you cannot do this, when this goes over your own limits of what you are prepared to do . . . No matter what you shouldn’t do it.” Netherlands/3

In contrast, in the Northeast United States, physicians who seriously dealt with PAD requests were in relative isolation, speaking about it rarely with colleagues and only with personal contacts, such as a spouse.

Probably the only other person was my wife. That’s all. She is not a physician, but that’s who I talked about it with . . . Just to say that this is where I was. US-Northeast/27

Factors Influencing PAD Discussion

There were several factors that consistently played a role in physicians’ responses to patients asking about PAD including legality, the perception of the role of a physician, patient rights and autonomy, and religion. Legality was often fundamental to the physician’s response to patients’ requests, both in establishing the criteria under which the physician would agree to consider PAD where it was legal or as a first line answer as to why they would not consider participating where it was not legal.

I know that some of this goes on in other states . . . I am sure patients and doctors find ways to help, and I am just really thankful that I don’t have to do that in any kind of climate of fear or weird communication . . . I don’t know what I would do, but I am sure I would be challenged, and I am sure it wouldn’t be as good for the patients...Having the structure that the law provides . . . a couple of consultations and going through some checklist of questions . . . provides a really useful framework. US-Oregon/23

I’d made it very clear to them that while I was certainly understanding of their point of view and was supportive . . . I wasn’t going to actively assist in that. For both my personal ethical views, I am not . . . comfortable taking an active role . . . but also the legal issues. I wasn’t going to put my license and my freedom on the line for anybody. US-Northeast/30

How physicians defined the role of a physician and whether or not this definition included the possibility of PAD also affected their discussions with patients. Dutch physicians saw PAD as a possible last stage of their duty to care for patients, where United States physicians—especially in the Northeast—generally did not.

Dutch doctors don’t want to do euthanasia . . . If possible, I would run away. But I see it as the last part of my care. I have taken care of that patient for years and now at the moment . . . when she needs me most . . . I would be a coward to run away then. Netherlands/8

The profession doesn’t view physician-assisted suicide as really being within the mainstream or the purview of . . . roles of a physician . . . I respect boundaries and respect the notion that the kind of consensus that develops within the profession of colleagues is important. US-Northeast/1

I have to do no harm, and I just feel that if you’re assisting someone in dying . . . it’s against what I’ve been trained . . . It’s not up to me to decide when the patient dies . . . US-Northeast /29

The limits of patients’ rights and autonomy were repeatedly mentioned in the interviews with physicians in all venues. Most physicians wanted to honor patients’ autonomy, but not at the expense of violating their own personal morals. The general consensus was that patients have a right to ask for PAD, but not a right to demand PAD.

Issues of autonomy are very difficult . . . wanting to support someone’s autonomy as much as possible, but is there some limit . . . If someone’s autonomy says, “I want to end my life and you need to do it,” or vice-versa, “I absolutely am not going to die and you have got to keep me going, no matter what.” All of those extremes are issues of autonomy. The doctor’s autonomy, the patients autonomy; they are [a] blend of the both. US-Northeast/32

I can imagine that I will perform it once but only when I am convinced . . . as a person . . . that a patient really suffers that much that it cannot be helped in another way . . . I have . . . great difficulties with the view point . . . that people have a right to get euthanasia. And I think that that completely goes beyond the fact that you need another person to get euthanasia, and that’s me. Netherlands/10

Religion played a role not only in institutional policies but also in physicians’ personal comfort with the idea of PAD; the latter was particularly true in the Netherlands.

I don’t determine whether or not they are a candidate or not. [Religiously-Affiliated Healthcare Institution] has a policy that says that they will not participate; they won’t support it or not allow you to do it . . . and so that is the role I take is that I don’t participate. I can give them the information of who they can participate with, and I will sit down with them and answer their questions and help talk them through it and tell them the way. US-Oregon/24

It has to do with . . . my Christian belief. I think God has made us—as a human being, has made me—and it’s not up to a human being to take the life of another human being . . . I’m against death penalty. It’s in the ten commandments. And so I wouldn’t be able to . . . I know . . . how people suffer, but I still think . . . killing another person is not the solution. Netherlands/9

Discussion

Whether or not PAD was legal, physicians described instances when they were asked about PAD. These discussions were significant and often very intense experiences for the physicians studied. Such discussions can facilitate a reexamination of the situation by the physician and optimization of palliative care, and can lead to a stronger relationship between doctor and patient. Our findings are consistent with Norwood’s conclusion that “euthanasia talk” in the Netherlands is a separate entity from the act of euthanasia, with its own function and purpose (Norwood, 2007). Our study is unique in that it presents the experiences of physicians in three distinct environments with different influencing factors, and yet found consistency in the existence and importance of these discussions.

Understanding the experiences of physicians who have discussed PAD with their patients can inform other physicians and help them to be more comfortable with allowing a conversation by recognizing that the topic can be a gateway to discussion of many important end-of-life issues that deserve attention. In addition, the understanding of how physicians deal with requests for PAD can inform those who create the regulations, which govern such practices. One concept that resonated in Oregon and the Netherlands was the importance of open discussions, both with patients and with others. This openness provides a transparency to the process, potentially making it safer and more manageable for physicians to navigate, because the legal framework and criteria to discuss and consider PAD have been established. Physicians in these areas felt as though they were able to have a rewarding interaction with patients at a crucial juncture in the doctor–patient relationship and found that the regulations gave a helpful framework for these discussions.

One of the fundamental objections to the legalization of PAD is the fear of the slippery slope (Beauchamp & Childress, 2009; Foley & Hendin, 2002; Norwood, 2009). However, physicians articulated, if anything, the opposite of a slippery slope; that the emotional burden of performing PAD made them think very hard before agreeing to participate. These findings are consistent with previous findings of Dutch primary care physicians avoiding future conversations about PAD after performing the procedure (Georges et al., 2008). Where PAD was legal, physicians expressed that on several occasions the regulations and guidelines had given them firm grounds for refusing to participate in PAD. Previous quantitative work has suggested the lack of a slippery slope in areas where PAD has been made legal (Battin et al., 2007). The qualitative data from our study demonstrate the thought processes and experiences that affirm these observations.

Limitations

This study was limited in that it investigated PAD discussions only from the point of view of the physician. It would be useful to interview patients and relatives about such discussions as other studies have done (Back et al., 2002; Norwood, 2005), but it is also logistically difficult and ethically challenging. In addition, since physicians were recalling conversations and experiences as they remembered them, the information is potentially subject to the physicians’ recall biases. However, since our focus was the physicians’ experiences, their recollections of the experiences are what will likely influence their future practice. The physicians could have been biased to give a socially desirable response. However, steps were taken to ensure confidentiality and comfort during the interview to minimize this effect.

Although the findings are representative of the groups studied, it cannot be assumed that the findings can be applied to other geographic areas. However, the similarities and differences across the three areas suggest potential for further study in other countries. Given the qualitative nature of the research, there are no statistical inferences about the frequency of certain experiences among different groups. However, given the complexity of the subject matter, it would be difficult to obtain useful data of that nature, and the value of that data would be diminished by the lack of context. Finally, these results represent a set of experiences and opinions of physicians at one point in time, and the changing legal landscape and other influencing factors mean that these results may be different if the study were conducted at a different point in time.

Implications

Since this research was carried out, several other states in the United States have explored legalization of PAD through ballot initiatives and court rulings, with Washington and Montana succeeding in forms of legalization. The experiences of physicians in diverse settings who attempt to address patient requests for PAD provide useful information for legal policy makers and physician groups who may be trying to operationalize some form of legal PAD or to establish a framework for dealing with PAD discussion in areas where there is not a move toward legalization, but there is an acknowledgement that these conversations occur. The experiences described in this study suggest that an appropriate legal framework should allow physicians and patients to talk openly about PAD in order to fully explore the reasons for the request, and yet contain specific criteria that safeguard patients and support physicians through a difficult decision-making process by encouraging discussion with medical peers. In areas where PAD remains illegal, these experiences can inform guidelines for physicians dealing with requests that emphasize asking why the patient is bringing it up and recommend strategies to reassure the patient that he or she will not be abandoned even if the physician will not be able to participate in PAD.

PAD is a complex topic to address in the context of the doctor–patient relationship with many influencing factors. Whether or not a physician considers participation, the discussion with the patient can be an opportunity to improve care and address important needs as the patient nears the end of life. The decision should be made in the open, with agreement between patients, family members, and medical peers if possible, and physician experiences should be fundamental in formulating policy changes that provide frameworks for broaching this topic.

Funding

Dr. Voorhees was supported by the Yale School of Medicine Richard K. Gershon One-Year Medical Student Research Fellowship and an NIH-NCRR CTSA-T32 Medical Student Research Fellowship through the Yale School of Medicine Office of Student Research to complete this work as her medical school thesis. Project funding was also provided by the Section of Geriatrics, Department of Internal Medicine, Yale School of Medicine.

Acknowledgments

The authors wish to thank the 36 physician participants, whose willingness to share their experiences made this work possible.

References

Author notes