What do physicians want at the end? An international qualitative study on physicians’ personal end-of-life preferences and what influences them

Study design

Qualitative study using semi-structured in-depth interviews.

Context and setting

We selected physicians practicing in jurisdictions with diverse cultural environments and varied levels of experience with assisted dying legislation. We included physicians in Belgium (Flanders), Italy, and the United States (Wisconsin).

Participant recruitment

We used a purposive sample in each country. Due to their varied levels of experience treating adult patients at the end of life, we sought to include three types of physicians: general practitioners, palliative care physicians, and other medical specialists with a high likelihood of seeing patients facing end-of-life issues (i.e., cardiologists, emergency medicine, gastroenterologists, geriatricians, gynecologists, internal medicine, intensivists, nephrologists, neurologists, oncologists, pulmonologists). Our goal was a distribution of physician types that included a minimum of five general practitioners, five palliative care physicians, and five medical specialists in each jurisdiction, amounting to 15 in each jurisdiction and 45 total physicians. Physician participants were identified through the networks of our research partners in each country, as well as through medical associations and professional societies, and self-identification following participation in the PROPEL online survey. This qualitative exploration is part of a multi-method study called PROPEL: Physician Reported Preferences for End-of-Life Decisions. Following initial contact by email or telephone, physicians received an email invitation detailing the study, information on the research team, and eligibility for participation.

Inclusion and exclusion criteria

Currently practicing physicians in the following groups were eligible for participation: general practitioners, clinical specialists (cardiologists, emergency medicine, gastroenterologists, geriatricians, gynecologists, internal medicine, intensivists, nephrologists, neurologists, oncologists, pulmonologists), or palliative care physicians. Exclusion criteria were having less than 2 years of clinical experience and being unable to do the interview in English. We originally planned to include only physicians with 5 or more years of experience. However, since we encountered recruitment challenges, eligibility was expanded. Most of the recruited physicians had 10–40 years of experience.

Data collection

Data were collected between January and November 2022 using a semi-structured interview guide. The interview guide (Appendix A) was pilot-tested with three physicians and adjusted for clarity and appropriateness before use. Most interviews were conducted online using a video conferencing platform (Zoom or Teams) due to COVID-19 restrictions. However, five interviews were conducted in person, at the office or home of the participant, and one at a public library. The interviews varied in length from 30 to 80 min and no repeat interviews were conducted. Notes were made following each interview. All interviews were conducted by the lead author (S.M.), a female doctoral researcher with previous experience conducting qualitative research with physicians (credentials – MSc, MPH). S.M. had no prior relationship with the participants, apart from two Wisconsin physicians who had previously provided care for her mother. Due to the complex nature of the interview topic, participants were offered language support through the attendance of a second researcher who spoke their native language and could explain complex questions or responses. Some European physicians accepted this option (Italy 4, Belgium 1). A demographics form was completed at the start of each interview.

Data analysis

Interviews were audiotaped and transcribed verbatim. Data collection and analysis were informed using the reflexive thematic analysis approach. A reflexivity statement was developed by the lead researcher (Appendix B). Reflexive thematic analysis is an ideal approach for this study as its flexibility allows for a deep examination of the multifaceted nature of end-of-life preferences without the constraints of rigid frameworks, making it well-suited to capture the complexity of this topic. ¹⁴ It emphasizes personal meaning and perception and aligns with our focus on physicians’ subjective experiences and reflections. The iterative, reflexive nature of this method ensures that the researcher’s role in interpreting sensitive, personal data is acknowledged, providing room for richer, more nuanced insights. ¹⁴ It excels in exploring the intersection of broader social and cultural contexts with individual decision-making, making it ideal for understanding the varied influences on physicians’ preferences. Identifying patterns across the data helps uncover shared and divergent experiences, facilitating a comprehensive understanding of the study population. Additionally, it is an inductive, data-driven approach, which allows for open exploration, making it particularly useful for research without predefined categories, enabling the discovery of unexpected themes. ¹⁴

Interview transcripts were imported to Lumivero’s NVivo (version 12) for analysis. Our recruitment goal was met, and thematic saturation was achieved in each jurisdiction. Reaching saturation suggests that the themes generated from the interviews are comprehensive and robust enough to address the central research aims, allowing for a nuanced understanding of the phenomenon being studied without the need for further data collection. The lead author (S.M.) coded all transcripts, and another team member (F.D.) independently coded four transcripts for comparison. Transcripts were not shared with participants. Initial codes were discussed and refined (by S.M. and F.D.) and themes and sub-themes were developed. The themes were then refined and discussed further by the full research team and similar or overlapping themes were collapsed or removed. A thematic map was then developed and evaluated, and associated interview quotes were identified, discussed, and selected by the full research team. Participants did not provide feedback on the findings. Consolidated criteria for reporting qualitative research (COREQ) were followed (Supplemental Material).

What do physicians consider for their own end of life

The sample reported a wide range of personal preferences related to various end-of-life decisions, shared information about the influences on their preferences, detailed elements they consider important for having a good death, and undesirable scenarios they hope to avoid.

We developed three main themes related to physicians’ consideration for their own end-of-life preferences: (1) Notions of a good death – physicians have developed ideas about what a good death would look like for themselves; (2) Preferences can evolve – physicians recognize their preferences could change over time; and (3) Reflection and discussion bring comfort – death and dying become normalized through reflection and discussion (Figure 1).

OPEN IN VIEWER

Figure 1.

Thematic map of physicians’ consideration for their own EOL preferences.

We found that most physicians have considered their personal end-of-life preferences, and for many, their ideas have clarified over time. Many physicians shared that they think regularly or often about their own mortality and have given some consideration to end-of-life issues. Palliative care physicians reported having reflected a lot on their own end-of-life preferences, which was not the case for general practitioners or other medical specialists.

Many physicians stated they have formed strong ideas about what constitutes a good death after witnessing experiences with family members or patients they considered peaceful and positive, and, conversely, situations they would prefer to avoid.

Physicians often cited the elements of a good death as being at home or in hospice, anticipating death, loved ones nearby, enough time for goodbyes, not suffering, pain and symptoms controlled, spiritual and practical affairs in order, a clear mind, and autonomy and dignity preserved.

Physicians largely preferred a death at home or in a peaceful environment. However, a small number said they have become accustomed to medical environments, and they would be fine with dying in a hospital. Many physicians specified that they would prefer to know that death was coming so they could prepare by putting their practical affairs in order and say goodbye to loved ones.

“Ideally, I am a much older man. I have a terminal illness that I see coming, so that way I have time to prepare and do the things that need to be done, say the things that need to be said. Family and people I care about around me, not in the hospital, in my own bed. Fairly free of symptoms, though I think a certain amount of suffering is just inevitable in end of life.” (Palliative care physician/internal medicine, Wisconsin)

Physicians also shared that they would prefer to avoid suffering and uncontrolled pain, and some specified they would like palliative care support and a competent medical team. If suffering could not be controlled, many physicians said they would want palliative sedation, and some would request assisted dying.

“I think (palliative sedation) it’s in line with the good death . . . I have seen what it can do and how people can say goodbye with more time and I think it’s a good alternative. It’s a good way of dying if you choose for it. If I would choose for it, I would. I think everything will be okay.” (Medical specialist, Belgium)

Other physicians shared that dignity and autonomy and having the ability to maintain cognitive clarity at the end of life are primary issues of concern to them.

“It’s all about what we call dignity. And it’s a very broad term, but I think (to) be aware of the situation is very important. Being able to say goodbye to your relatives, to your loved ones, to be as comfortable as possible and to be without fear and without worries. Because it’s not because I work in palliative care that I’m not afraid of dying. And I hope that I will be surrounded by people who can give me comfort and who can, because I think I will be very frightened.” (Palliative care physician, Belgium)

For some physicians, preserving a level of independence and not being a burden on their families at the end of life is a significant concern. Several palliative care physicians stated that part of a good death is that it is not medicalized.

Physicians shared that they have noticed patients and/or family members change their minds and want very different things at the end of life than they might have considered earlier. Personal preferences and what one considers acceptable can evolve as an illness progresses and physicians acknowledge this could also be the case for themselves.

“My intention . . . is to revisit that document every ten years to make sure it’s still applicable because, gosh darn it, so many patients I’ve run into change their mind. Over time you can get acclimated to new baselines you never thought that you could.” (Palliative care physician, Wisconsin)

“How you change in the course of dementia, how you as a person change. And even if you think this will be burdened to me, once you are in that state, that burden often is not that clear anymore just the same as you would think. I don’t want to have chemo in my life because see how sick people are. But once you suffer from a cancer, you are willing to get chemo, so things change.” (General practitioner, Belgium)

Physicians expressed a level of acceptance about the fact that they may change their minds in time. In fact, some physicians cite this as the reason they have not yet done an advance directive, which was the case for many participants.

“I think it is much more important at regular times to speak with my family (about) how I’m looking and maybe like I look now to health and to diseases and end of life circumstances. This is a dynamic process. Probably I will change my mind in the next years. That is possible and acceptable.” (Palliative care physician, Belgium)

Other physicians say their preferences have changed due to exposure to specific clinical situations or they have become more nuanced. For example, some physicians described having an opposition to assisted dying as an option for themselves early in their careers but becoming more open to assisted dying practices after seeing various difficult end-of-life scenarios. Others say their preferences developed early, such as avoiding aggressive life-prolonging measures, and have only deepened over time. Other physicians believe their core preferences have stayed constant over time and their wishes have altered little.

“At least six months of my residency was spent pretty heavily pondering what a good death can look like and what good deaths look like. And then, you know, probably at least monthly something crops up where it makes me reflect on my own mortality. Either at work or at home or in the universe. I don’t think I spend more than 30 seconds thinking about it, because I feel like, my opinions aren’t changing, my views aren’t changing.” (General practitioner, Wisconsin)

Some physicians reported feeling uncomfortable and uncertain during end-of-life conversations early in their careers but after more personal reflection on their own feelings and preferences and more clinical experience they came to feel more at ease initiating and navigating difficult end-of-life conversations with patients and families.

“Thinking about my own end of life preferences has made me, continues to make me more open with discussing different scenarios with my patients. So I think that has helped me, just, I think the more I’m introspective about things the better I’m able to talk about those scenarios with my patients. Not necessarily influence the way I talk about it, but just, you know, makes me more open to those conversations with people.” (Medical specialist, Wisconsin)

Physicians’ perceptions of how their own end-of-life preferences have been impacted by personal, professional, cultural, legislative, social, and religious influences

When we explored how physicians have reflected on their own end of life and perceive their preferences have been impacted by socio-cultural influences, physicians shared that their personal reflection on end of life is often triggered by a specific catalyst. These motivators include personal interest/curiosity, treating patients of a similar age, patient situations, history of illness in family, getting older, personal/family end-of-life experiences.

From the interview data, we identified four main themes and two sub-themes, including: (1) Death of loved ones – physicians have been significantly impacted by the death of loved ones; (2) Clinical practice experiences (sub-theme: Undesirable situations) – physicians have been influenced by their clinical practice, particularly negative experiences; (3) Assisted dying laws – physicians have been influenced by the existence/absence of assisted dying laws; and (4) Socio-cultural context (sub-theme: Religious beliefs/faith) – physicians have been influenced by a variety of factors including culture, ethnicity, education, social norms, personal/family values, and particularly, religion.

Many physicians shared experiences of having acted as a caregiver for a loved one or seeing the impactful death of parents, grandparents, family members, or other loved ones. Often the illness trajectory and/or cause of death was mentioned as shaping the experience as positive or negative and influencing their attitudes toward preferring or rejecting certain end-of-life decisions.

Many physicians discussed being strongly influenced by seeing loved ones die after a prolonged illness with dementia or cancer, which sometimes involved choices or treatments they would not prefer for themselves.

“That (mother’s illness) was probably the primary one. I saw her suffer through cancer for almost 18, 19 months and chemo, radiation, everything, and I, I think if I had a terminal diagnosis like that, I probably wouldn’t do anything.” (General practitioner, Wisconsin)

“I think going through it personally was also helpful with my father’s experience, and he actually made choices I would not make for myself. And we respected and supported them . . . personal experience really made me much more comfortable and realizing how important it was to have these conversations.” (Palliative care physician, Wisconsin)

Other physicians discussed the sudden death of family members, especially young family members, as a particularly challenging experience and a significant influence on their lives and personal preferences.

“I would say my child’s death in some ways made it more personally difficult for me, but yet also a little bit more of a calling to try to make it be a better experience for, for patients and families.” (General practitioner, Wisconsin)

Seeing difficult end-of-life patient situations and helping patients make end-of-life decisions profoundly impacts what physicians prefer for themselves. Many physicians described challenging end-of-life scenarios that involved prolonged illnesses with high levels of physical or psychological suffering or situations with cognitive impairment and dependency as being influential. Many also described seeing patients receive aggressive or futile treatments that they would strongly prefer to avoid at the end of life. Some physicians described being impacted by treating patients of similar age to themselves. Other physicians described peaceful deaths, or sudden deaths, or assisted deaths they thought might be preferred, or not preferred, for themselves.

“I just vividly remember this, an intensive care unit where every single person was a young man with Aids and pneumocystis pneumonia on a ventilator. None of them would survive, and it just would be one after another die, die, die, die. And, you know, finally . . . a lot of hospitals either said to people, you know, we’re not going to put you on a ventilator with this condition because you’re not going to survive it. So you’re better off just having palliative care or, you know, I think there was a certain amount of assisted, you know, assisted dying going on. You know, maybe sort of under the table, so to speak. But, you know, that was just horrible to see, you know, people just, you know, sort of be kind of bright and perky, get intubated, they’re awake at the time, and then that’s it, you know, then that’s the last moment that they’re awake or interacting with anybody. And I just thought, this is just really awful. So that, that I think really influenced me.” (Palliative care physician, Wisconsin)

Some physicians explained that their clinical experience working in an intensive care unit influenced them to prefer another type of environment at the end of life and others shared that exposure to palliative care teams or receiving palliative care training was influential and exemplified the kind of care they would prefer to receive at the end of life.

“My time in critical care is probably the biggest (influence). Well, and, if I could choose a second, I would say my interaction with the palliative and hospice care services, to know that that existed and to kind of get what they do and to think, wow, this feels so comfortable. And, and actually, it’s a strangely joyful place to spend time. And so my time in critical care showed me a lot of what I don’t want. And my time with in working with the palliative and hospice care folks was shorter, but probably equally meaningful.” (General practitioner, Wisconsin)

Sub-theme: Undesirable situations – Physicians described being deeply impacted by seeing situations and deaths they would not want for themselves in both their personal and professional lives. Undesirable situations were interpreted as a cross-cutting sub-theme across Theme 1 – The death of loved ones and Theme 2 – Clinical practice experiences.

The undesirable situations detailed by physicians in the interviews included: sudden deaths, being in the hospital or intensive care unit at the end of life, having uncontrolled symptoms or severe pain, being on a ventilator long term without the ability to communicate, receiving futile or aggressive treatment, experiencing physical and/or psychological suffering, being alone, being dependent on others for care, being confused or cognitively impaired, and being a burden on their families.

“I hope I will not die alone. That’s what we see in some care units or that there we see patients that have not a lot of family, and where only the doctors and palliative care personnel is there for the patient, so I think that’s very important to me. Another thing is, is that the things you hear a lot like suffocation or a lot of pain that are the two main reasons that I hear a lot and that are also praying for myself that I hope to avoid. And I think we can avoid at the end of life.” (General practitioner, Belgium)

Most physicians expressed the desire to avoid life-prolonging treatment at the end of life including cardiopulmonary resuscitation, artificial nutrition and hydration, and mechanical ventilation. The concepts of being dependent and physical and mental suffering also came across as a main concern for physicians considering possible end-of-life scenarios.

“I guess (I hope to avoid) a lot of pain at the end where there’s trouble with pain relief. Also, like the physical indignities of the end stage where you are incontinent and you need to depend on other people. I think also like more the emotional burden, if you would be, if I would be very young to die and I would know it in advance. And you see the suffering of your family. I have four children myself so I think that would be horrible to know that I would have to leave my children. So I think it’s both the physical suffering as the, as the mental suffering.” (General practitioner, Belgium)

Some physicians feel assisted dying laws are a positive and important end-of-life option for themselves and patients and are grateful to live in a jurisdiction that allows these options or express a wish to have an assisted dying option available to them. Other physicians are opposed to assisted dying for themselves, or for everyone, and expressed concern and conflict about the notion of having to practice medicine in an environment that allows assisted dying should it become legal where they practice. These beliefs seem to be often aligned with the legal situation in the jurisdiction of the physician.

“I think this (assisted dying law) will influence us as physicians and just in our own personal view, if you can talk about something freely, I think it will alter your way of thinking also. Yeah, but in this way, I’m happy that we can. We don’t have to have any fear about this topic.” (Medical specialist, Belgium)

“This can influence you because if nobody among you, nobody in the hospital, nobody among patients, nobody among your friends even have to speak about assisted suicide or something like that, you never start to think about it and you just think implicitly that this is not a solution to consider.” (Medical specialist, Italy)

“I think not only Wisconsin, but, you know, the United States in general. So absolutely, it affects, it limited my scope of thought process because when I was filling out the survey, I thought, oh, wow, I really haven’t even thought about this. And it’s because, it’s not, it’s not an option for me or my patients or my family.” (Medical specialist, Wisconsin) (The “survey” refers to the related PROPEL quantitative survey study)

Some physicians acknowledged feeling undecided or conflicted about assisted dying practices or believe their preferences have not been impacted one way or another by the presence or absence of an assisted dying law. We also found that palliative care physicians strongly preferred palliative sedation to assisted dying at the end of life. Many general practitioners and other medical specialists said they would consider assisted dying if it were a legal option and all physicians said they would consider palliative sedation an option to relieve suffering at the end of life.

Physicians shared personal stories and traditions related to death and dying within their families, often of their parents or grandparents, as well as impactful exposures to films, books, media stories, etc. For example, the books Being Mortal by Atul Gawande and The Death of Ivan Ilyich by Leo Tolstoy were mentioned as being influential for some physicians, as well as exposure to media coverage of patients who remained in a persistent vegetative state for extended periods. These exposures motivated some physicians to reflect on end-of-life issues and influenced their own preferences.

“I traveled a lot for purpose of studying and working, and that really helped me a lot in seeing different perspectives in different relationships with doctors and their patients. The cultural situation and also practical aspects of care and, and I’m sure that they helped me a lot in feeling more comfortable, because I think if I just lived in Italy, probably I would be more narrow minded because I would have I would not have seen so many things. That really impacted the way that I opened the discussion about end of life.” (Palliative care physician, Italy)

“Reading about people dying, even in a fictional setting, or probably more so in a fictional setting, has influenced me. I mean, there’s some of the most beautiful and accurate passages about death in literature.” (General practitioner, Wisconsin)

Sub-theme: Religious beliefs/faith – Some physicians feel they have been influenced by their religion/faith, or non-religious views and believe these views have impacted their personal preferences for end of life.

Some physicians described having a strong connection with God and/or their faith and a comforting belief that their lives and end of life are in the hands of a higher power. Some cite their religious beliefs as the basis for their opposition to assisted dying practices and belief that the end of life should not be hastened. Other physicians describe being influenced by treating patients with strong religious beliefs, who received what the physician considered overtreatment due to religious objection to ending treatment. This seems to be a key issue as several physicians expressed concern and objection to receiving aggressive or futile treatment at the end of life and this was influential on their personal preferences.

Our study found that physicians with more strongly held religious beliefs were more opposed to assisted dying practices. This was prominent with physicians from Italy but also with the more religious physicians in Belgium and Wisconsin.

“I think that religion influences me only to artificial death . . . I’m against to artificial death. I think because of religion, cultural aspects, really in myself and I think that this aspect could influence.” (General practitioner, Italy)

“And the way you are raised as a child, for example, I’ve been raised in a rather, a very Catholic family, actually. So my, mother and father, they went to mass every Sunday. We prayed. I had to go there, too. I don’t do it anymore now. So I’ve lost that kind of necessity but I still consider myself a Catholic, so I didn’t. I still consider myself as being a Christian, a Christian man. But it all shapes you to what you want also. You . . . have to think about it because it influences, of course, your way of working with people. Which might, which might be a danger because you might want to fill in for the patient like the way you want it to be.” (General practitioner, Belgium)

Other non-religious physicians shared that they held spiritual or other personal beliefs that influenced their preferences. Some non-religious physicians noted that it is hypocritical to argue that medicine should not interfere at the end of life since medicine has developed in significant ways and often intervenes and impacts the beginning of life.

“I’m not religious. So for me deciding or taking end-of-life decisions, for me, there is no religious component in, or like, life is sacred. I think that medicine had already developed. I always think it’s a bit hypocritical to say that we can’t decide for God, because we do it already. If you see what we do with little children when they’re born, when they’re 23, 24 weeks, we already play for God. So when we can play for God in the beginning of the life, I don’t see why we shouldn’t, can’t do that at the end of life. Yeah. That’s really, so I totally, I respect the religious people but I totally disagree with that because we do it already.” (General practitioner, Belgium)

Strengths and limitations

A strength of our study is the cross-national qualitative design, including physicians across three countries representing varied legal and cultural environments and the participation of three groups of physicians – general practitioners, palliative care physicians, and other medical specialists. Our preferred balance of physician groups was successfully recruited and the total sample of 45 physicians was ideal to gather comprehensive insights given our research questions, the diversity within the sample, and our analytical approach. By focusing on how physicians have considered their end-of-life preferences and the factors that influence them, we provide a much-needed addition to the current literature on physicians’ end-of-life preferences. The study had certain limitations. Although the final sample included a variety of physician groups in three countries, it was ethnically homogenous, and there may be selection bias as the interview study could have attracted those with a particular interest in end-of-life issues, resulting in a sample that may have reflected more intensely on the topic.

Policy and practice implications

These study findings suggest several important implications for policy and practice. First, clear policies and ethical guidelines are needed to help physicians manage moral conflicts and maintain professional boundaries to ensure that clinical recommendations are based on patient preferences rather than personal beliefs or biases, particularly in end-of-life care. Second, ethical guidelines would help physicians navigate complex decisions, such as those involving physician-assisted suicide or euthanasia. Third, promoting palliative care training, which fosters a person-centered approach, could help reduce the influence of personal preferences on patient recommendations, making it crucial to integrate palliative care principles into general medical education. Fourth, addressing cultural and religious influences on physician decision-making through cultural competency training is also essential, especially in jurisdictions where strong religious traditions impact patient care.