Physicians’ moral distinctions between medical assistance in dying (MAiD) and withdrawing life-sustaining treatment in Canada: a qualitative descriptive study

Abstract

Background

Medical assistance in dying (MAiD) was legalized in Canada following the Carter v. Canada ruling of 2015. In spite of legalization, the ethics of MAiD remain contentious. The bioethical literature has attempted to differentiate MAiD from withdrawing life-sustaining treatment (WLT) in an effort to examine the nature of the moral difference between the two. However, this research has often neglected the firsthand experiences of the clinicians involved in these procedures. By asking physicians if they perceive the major bioethical accounts as clinically useful, we seek to distinguish between aspects of the contemporary bioethical landscape which are useful at the bedside and those which are divorced from the realities faced by clinicians.

Methods

We applied a qualitative descriptive approach to explore physicians’ experiences and bioethical distinctions in providing MAiD and WLT.

Results

Semi-structured interviews were conducted with 21 physicians, and the transcripts were thematically analyzed to identify common patterns and divergences in their perspectives. Three core themes were found: (1) consensus on MAiD’s moral equivalence with WLT despite differences between the practice, (2) discord regarding the use of the term ‘killing’, and (3) disjuncture between bioethical debates and practice. Theme 1 comprised of three sub-themes: (1.1) no moral difference between MAiD and WLT, (1.2) physician versus underlying medical condition as cause of death, and (1.3) relief of suffering.

Conclusions

In order to have practical utility for clinical practice, it is essential for bioethicists to engage in dialogue with patients and their medical providers pursuing MAiD or WLT. Theoretical debates that are divorced from the realities of terminal illness do not assist physicians with navigating the ethical terrain of ending a patient’s life. This research captures meaningful accounts regarding MAiD and WLT that is rooted in the lived experience of the providers of these services in order for bioethical debates to have substantive impact in clinical practice and in legislation surrounding future health policies.

Introduction

In February 2015, the Supreme Court of Canada unanimously struck down the ban on medical assistance in dying (hereafter referred to as MAiD) in the case of Carter v. Canada [1]. The following year, MAiD was legalized by the passage of Bill C-14, the federal law permitting physicians and nurse practitioners to perform the procedure on patients who meet the specified eligibility criteria [2, 3]. This legislative shift introduced profound ethical and practical challenges for Canadian healthcare providers, particularly regarding how MAiD compares to the withdrawal of life-sustaining treatment (WLT). WLT involves ceasing interventions such as ventilators, dialysis machines, or tracheal intubation, whereas MAiD entails the active provision of medication to end a patient’s life. Although both practices can result in death, the distinctions between them—rooted in causality, physician intent, and the active versus passive nature of actions—have sparked significant bioethical debate [4,5,6,7].

The relationship between MAiD and WLT is central to ongoing discussions about the moral permissibility of these practices. It is legally required to withdraw life-sustaining treatment from a patient who requests that it be withdrawn, and WLT is almost universally accepted within the medical community as morally permissible. In the literature, the permissibility of WLT has been used to support [8, 9] or to contest [10, 11] the permissibility of MAiD. If MAiD and WLT are effectively equivalent practices, then MAiD must be considered permissible as well, on pain of inconsistency. The judgment that MAiD is morally impermissible requires a meaningful difference between it and WLT. For these reasons it is imperative to understand how clinicians experience the similarities and differences between MAiD and WLT. Bioethical discussions of these subjects often fail to capture the lived experiences of healthcare providers who must navigate these ethically complex practices in their day-to-day work. While some studies have examined healthcare providers’ opinions on MAiD logistics, policies, and willingness to participate [12,13,14], few have delved into their moral and ethical experiences interacting with patients seeking to end their lives [15]. Though many sociological studies have been conducted on physicians’ experience with end-of-life decisions in general [16,17,18,19,20], clinical experience with MAiD [21,22,23], and with withdrawing life-sustaining treatment [24,25,26], few empirical studies have explicitly attempted to compare the experiences of physicians with MAiD and WLT using a qualitative approach. Even fewer have sought to assess clinician perspectives on the viability of bioethical distinctions for end-of-life practices in the Canadian context.

Our study seeks to address this gap by exploring how physicians across Canada morally navigate their roles in providing or abstaining from MAiD compared to WLT. It aims to capture their lived experiences and assess the applicability of bioethical debates to clinical practice. As MAiD legislation evolves through policies like Bill C-7—extending eligibility to individuals whose deaths are not reasonably foreseeable—understanding these experiences is crucial for ensuring the procedure is delivered ethically and effectively. An interconnected aim of this study is to critically evaluate the clinical utility of specific bioethical debates differentiating MAiD from WLT. These debates often focus on causality [27], physician intent [28,29,30], and active versus passive involvement in bringing about death [31,32,33,34]. An interconnected aim of this study is to critically evaluate the clinical utility of specific bioethical debates differentiating MAiD from WLT. This study underscores the importance of grounding bioethical discussions in the experiences of practitioners and patients to inform public policy and healthcare delivery meaningfully.

Methods

Study design

This study applied a qualitative descriptive approach to elicit the lived experiences and moral beliefs of physicians involved in providing either MAiD, WLT, or both practices. Qualitative description is a pragmatic and adaptable method that allows for the exploration of human experiences within a naturalistic setting [35]. It facilitates a rich account of beliefs and experiences while remaining close to the language and syntax of participants [36]. This is consistent with this study’s goal of elucidating physician’s experiences and examining how they rationalize them through a bioethical lens. Remaining connected to participants’ language also ensures the findings can directly inform clinical practice [35, 37]. This design was informed by the framework described by Stumpf and Rogalski [38] to examine the practical utility of popular bioethical distinctions used to differentiate MAiD and WLT. The study design was applied across all phases, including interview facilitation, data collection, and thematic analysis. Collaboration between investigators from philosophy, sociology, and health services research further shaped the methodological approach and ensured interdisciplinary rigor.

Sampling and recruitment

The study recruited 21 Canadian clinicians with experience in either MAiD, WLT, or both (see Table 1 for a full demographic profile). Purposive and snowball sampling methods were used to identify participants involved in MAiD (provision, assessment, or referral) and/or WLT. Recruitment targeted physicians from diverse specialties and varying levels of professional experience to capture a broad range of perspectives. Of physicians who had experience with MAiD, 14 were MAiD providers, 3 had assessed patients for MAiD eligibility but had not provided MAiD, and 4 had referred patients for MAiD but had neither assessed for nor provided MAiD themselves. An advertisement was posted on the Canadian Association of MAiD Assessors and Providers (CAMAP) website for recruitment purposes. Physicians practicing in any Canadian province or territory were eligible to participate. Initially, the study aimed to conduct in-person interviews with clinicians in Ontario’s Waterloo-Hamilton region. However, COVID-19 protocols necessitated a shift to virtual interviews, which enabled greater geographic variation in the sample. All participants received a consent form, a document outlining the study’s goals, and a demographic survey before the interview. The consent process was repeated verbally on the interview day. Ethics clearance was obtained from the University of Waterloo Office of Research Ethics (ORE #40801).

Table 1 Demographic characteristics of interview participants (N = 21)

Data collection

Data collection involved in-depth, semi-structured interviews designed to explore participants’ perspectives on key bioethical distinctions between MAiD and WLT. The interview guide was developed collaboratively by a philosophically trained investigator (AS), a sociologist investigator (KB), and a health services research investigator (MM), with additional input from end-of-life care practitioners. This iterative process ensured the questions were reflective of bioethical literature, free from undue bias, and appropriate for clinical contexts.

The interview guide was finalized following feedback from stakeholders and a mock interview conducted within the research team to refine the format and estimate duration (see Supplementary File 1). Interviews were conducted virtually via Microsoft Teams due to COVID-19 social distancing measures. KB and MM co-facilitated the interviews, which ranged from 30 min to 2 h. Written and verbal consent was obtained for audio recording. All transcripts were de-identified to ensure participant confidentiality.

Data analysis

Data analysis was conducted using NVivo software (version 1.0) for coding and thematic analysis [36, 37]. MM manually transcribed all interviews to ensure that participant data was accurately captured. Transcripts were edited and reviewed for accuracy and were then reviewed by the members of the research team to become familiar with the data and assist with the early identification of meta-concepts. Themes were independently reviewed by two primary analysts (MM and KB) and areas of convergence and divergence were discussed until consensus was reached.

Reflexive journals were maintained by KB and MM throughout the data collection and analysis phases. These journals documented personal biases, perceptions, and reflections, ensuring transparency and credibility in the research process. By systematically addressing preconceptions, the research team promoted reflexivity and trustworthiness in interpreting the data.

Results

Interviews for this study were conducted between October to December 2020. 22 physicians were interviewed, and 1 withdrew from the study. The results comprise the findings obtained from 21 participants. The virtual format allowed for participation from physicians across diverse practice settings, medical specializations, and provinces. Table 1 captures the participant characteristics. The majority of participants were physicians who had practiced in Ontario (n = 17), with other participants from British Columbia (n = 2), Manitoba (n = 1), and Saskatchewan (n = 1). Most participants had extensive experience with WLT (n = 19), and more than half had experience both assessing eligibility for and providing medical assistance in dying (n = 14) and 3 physicians who had experience as assessors only. Just under half of participants were general practitioners (n = 8), with the remaining representing palliative care (n = 6), internal medicine (n = 5), psychiatry (n = 1), and obstetric (n = 1) specialties.

The findings conceptualize physicians’ personal distinctions between withdrawing life-sustaining treatment and active provision of medical assistance in dying. Our study found three overarching themes: (1) consensus on MAiD’s moral equivalence with WLT despite differences between the practice, (2) discord regarding the use of the term ‘killing’, and (3) disjuncture between bioethical debates and practice. Theme 1 sub-divided into three sub-themes: 1.1 no moral difference between MAiD and WLT, 1.2 physician versus underlying medical condition as cause of death, and 1.3 relief of suffering (see Table 2 for a full summary of themes).

Table 2 Summary of themes

Theme 1: Consensus on MAiD’s moral equivalence with WLT despite differences between the practices

This theme explores the nuanced bioethical discourse on whether there is a moral distinction between Medical Assistance in Dying (MAiD) and Withdrawal of Life-Sustaining Treatment (WLT). While MAiD remains contentious among bioethicists and practitioners, WLT is widely deemed morally acceptable. Despite significant variation in perspectives, most participants ultimately did not perceive a moral difference between the two practices, even as they differed on the underlying reasoning regarding causality and intent. This theme aims to integrate and elucidate these areas of discord while emphasizing the shared focus on patient-centred care and relief of suffering.

Divided perspectives on causality and intent

Participants exhibited considerable variation in views regarding whether it is the physician’s actions or the underlying illness that causes death in MAiD and WLT. Seven participants argued that MAiD directly causes death, while WLT does not. Among these, four participants opted not to engage in MAiD, citing the act of administering a lethal drug as a defining distinction.

“There’s a difference. I wasn’t causing my patient’s death when I did not prescribe antibiotics for an elderly man with pneumonia when he asked me not to. The pneumonia was doing it, and he was not wanting treatment. I was not causing his death, whereas I am causing the death when I inject a lethal dose of medication” (Participant R: family physician, 30 + years of practice, MAiD provider).

Five participants saw both practices as causing death, asserting that the distinction is less meaningful due to the shared outcome:

“In both, it’s a deliberate act. The end result is clear. The getting there is somewhat different, but by and large, we’re provoking death… in my book, it’s the same thing.” (Participant K: palliative care specialist, 11–20 years of practice, MAiD provider).

Nine participants contended that neither practice causes death, attributing it instead to the underlying illness. For example:

“I really strongly feel the underlying disease is the cause of their death [in both] … even after death in a MAiD case, I will remind them that on the death certificate, I will put metastatic cancer of this or that… that’s really what killed them” (Participant P: family physician, 30 + years of practice, MAiD provider).

Relief of suffering as a Shared goal

Participants consistently emphasized relief of suffering as the primary intent in both MAiD and WLT, underscoring the centrality of patient-centred care. However, perspectives on the prevalence of this focus varied. Many participants described relief of suffering as their foremost goal when providing MAiD.

“I’m intending to relieve suffering. If I was ever asked to withdraw treatment or perform MAiD and I didn’t feel that it was a way to relieve suffering, I could not do it. In fact, my intention is not to cause death, my intention is to provide relief of suffering, and that is true for both WLT and MAiD” (Participant E: family physician, 30 + years of practice, MAiD provider).

Others highlighted the importance of patient autonomy in determining the path to relief:

“In both situations, you’re intending the end result of death, but in both situations, you’re trying to achieve patient-centred care, meaning we’re trying to do it in the way that the patient wants it done. It’s how they want to achieve that same end result” (Participant G: palliative care specialist, 21–30 years of practice, MAiD provider).

Moral Equivalence despite Discord

Despite these divergent views on causality and intent, the majority of participants (n = 16) did not perceive a moral difference between MAiD and WLT. Participants explained that both practices involve actions leading to an inevitable death, with distinctions largely concerning the mechanisms and timing rather than moral principles. For instance, on timing of death.

“When you administer a lethal substance, the patient dies in 10 minutes… and then the withdrawal of treatment can take several hours [or] days… there is quite a distinct difference. Mind you, I do think they’re both ethically acceptable” (Participant P: family physician, 30 + years of practice, MAiD provider).

Factoring in considerations of autonomy and personal choice, another participant found no meaningful distinction between practices:

“…voluntariness and personal choice. Morally or ethically, taking someone off a ventilator with the certainty that they are going to die is absolutely no different [than MAiD] to me. Both are actions where there is a certainty that a person is going to be deceased… I’m not sure it matters what the mechanism is” (Participant V: internal medicine specialist, 21–30 years of practice, MAiD provider).

Theme 2: discord regarding the use of the term ‘killing’

A question posed to interviewees was whether it is more appropriate to describe WLT and MAiD as killing or allowing to die. Historically, these terms have been used by bioethicists seeking to refine the debate from a perspective based in moral theology [34, 39, 40]. Responses to this question revealed differing perspectives among medical practitioners about the appropriateness of these distinctions and their applicability in real-world practice.

Some participants expressed offence at what they perceived as provocative or emotionally charged language, particularly regarding the term “killing.” Of the 10 interviewees who emphasized that neither MAiD nor WLT should be described as “killing,” 6 found the term offensive, arguing that it carries connotations of criminality and moral wrongdoing.

“The people who [use the word] kill – why? “Put an end to him?” Why pick the term kill? We kill enemies, I don’t kill patients. In reference to health care, it’s just objectionable” (Participant U: internal medicine specialist, 30 + years of practice, MAiD assessor).

Another participant compared its emotionally charged nature to rhetoric used by anti-abortion interest groups:

“I wouldn’t use the word kill… you’re using a word that has emotional baggage. That’s like in the abortion world, where people say pro-choice as baby killers. You’re using highly emotional words” (Participant R: family physician, 30 + years of practice, MAiD provider).

In contrast, fewer participants expressed strong opinions about the term “allowing to die.” Almost half of respondents (n = 10) preferred this terminology over “killing” if language had to be applied to both MAiD and WLT. One participant emphasized the nuances of the term “killing” and the role of patient autonomy in reframing its application:

“The word killing implies that it’s involuntary, that you are doing this to someone who doesn’t want you to do it to them, and this is different. This is someone who is saying, ‘I want to die, I want you to do this.’ It’s voluntary. Some individuals may be likely to take their own lives if you don’t help them with MAiD, so are you therefore allowing them to die because they would have died by suicide if you didn’t help them? You might actually just be allowing them to die and not killing them” (Participant M: psychiatrist, 11–20 years of practice, MAiD assessor).

Two participants, neither involved in MAiD, found the bioethical distinction between “killing” and “allowing to die” useful, viewing it as a valuable tool for abstract discussion rather than a practical reflection of clinical realities:

“I find the distinctions are helpful and valuable that the bioethicists are making. It provides some terminology, just looking at the issue at a different level” (Participant D: palliative care specialist, 6–10 years of practice, no MAiD involvement).

Theme 3: disjuncture between bioethical debates and practice

The final theme captures the clinical utility of bioethical debates. The majority of participants stated that they did not find practical utility from current distinctions. Most respondents (n = 18) found that the abstract bioethical debates comparing MAiD and WLT were not helpful because they are not grounded in the reality of patients’ lived experiences with pain, illness, and desire for control over their end-of-life circumstances. Differentiating between the experiences of an ethicist and a medical practitioner, one participant shared the perspective that:

“The difference between the ethicist and me is that tomorrow I’ll be on the ward with a real patient and have to make a decision” (Participant H: obstetrician, 30 + years of practice, MAiD provider).

Similarly, respondents stressed that any debate pertaining to the ethics and moral differentiation between MAiD and WLT must be grounded in real-world practice, or else it holds no meaningful significance:

“I think if you are [engaging] in the debate, you do actually have to see the suffering that people go through in the situation that’s being discussed” (Participant O: family physician, 30 + years of practice, MAiD provider).

One participant shared the opinion that the debate is outdated and not grounded in present realities, thereby limiting its usefulness to medical practice:

“I think it’s old, right? I think it comes back to when we talked about active versus passive euthanasia if you want to use those terrible old terms, but I think to me, it’s old thinking” (Participant G, palliative care specialist, 21–30 years of practice, MAiD provider).

Three of the four respondents who chose not to be involved in MAiD said that the debates were useful. For instance, they cited the bioethical debate as useful as abstract distinctions in considering questions as to intent of the medical practitioner:

“I think they are helpful because it helps you to categorize different conflicts that are at play. It helps you crystallize your thoughts around certain concepts, such as intention, what is actually the cause of death, and what is intention as a practitioner. It helps us… understand where our overall judgments are actually coming from” (Participant J: internal medicine specialist, < 1 year of practice, no MAiD involvement).

Discussion

The findings from this study highlight the moral nuances and complexities physicians navigate in comparing MAiD and WLT. These accounts indicate gaps in the bioethical debate concerning these two procedures by pointing out ways in which philosophical accounts can fail to relate to actual clinical practice. Two particularly significant conclusions emerged from our data. First, practitioners generally view MAiD and WLT as equivalent in terms of moral permissibility, despite recognizing differences between the two practices in terms of intent, underlying cause of death, and physician active vs. passive participation. In general, participants did not consider MAiD and WLT as morally distinctive since both procedures occur in circumstances of terminal illness. Division was found as to whether the physician or the underlying medical condition causes death; however, almost all physicians involved in either practice cited relief of patient suffering as their core intention. This challenges the bioethical literature which treats death as a negative consequence of an intention to relieve suffering [30, 41], as our participants believed death to be a necessary end for intolerable suffering.

The results of the study are relevant to other issues currently being debated in Canada. The attention to clinical experience demonstrated in our study also relates to broader questions about the scope and goals of medicine. For instance, can MAID be considered a therapeutic procedure? The majority of our participants indicated that they viewed MAiD under Bill C-14, for patients whose deaths are reasonably foreseeable, as a tool used to relieve patient suffering, which is consistent with the literature that refers to the procedure as therapeutic [42, 43]. Participants used the terminal nature of patients seeking MAiD as a key ethical factor to argue for equivalence between the two practices. But Bill C-7’s extension of MAiD eligibility to individuals whose deaths are not reasonably foreseeable, and to those suffering solely from a mental health condition [44] challenges the justification of MAiD based on a comparison with WLT. Several physicians expressed concern regarding the ethical implications associated with Bill C-7, and how this expansion may change the way they perceive their involvement and moral understanding of MAiD. Providing MAID for non-fatal conditions, including certain mental illnesses, is necessarily distinct from WLT at a conceptual level since WLT cannot take place in regard to non-fatal conditions. Attention to the lived experience of physicians would help to confirm or disconfirm this conceptual argument.

The second major conclusion is that, from the perspective of medical practitioners, abstract philosophical discussions regarding end-of-life care often fail to capture the kinds of real-world nuances that should be prominent in such debates. This study responds to the abstract nature of much debate about the distinction between killing and allowing to die in the bioethical literature, as outlined in the framework by Stumpf and Rogalski [38]. While the discussions of classical bioethics focus on intent and causality as key aspects of the debate concerning MAiD, these factors were not central for most participants. Rather, participants found concepts such as respect for autonomy and attending to patient suffering and needs to be more salient. Paying attention to the lived experience of clinicians involved in practices that lead to patient death provides a more concrete approach to appreciating the difference, of lack thereof, between MAID and WLT. Whether there is a difference between the two practices is not only a conceptual matter, but a practical and clinical matter. The question should be addressed not solely through philosophical thought experiments, but rather an understanding of the embodied moral and clinical experiences of practitioners who are required to be involved.

Concerning the semantic debate of ‘killing’ and ‘allowing to die’ predominant in bioethical literature, participants expressed distaste for such politicized and emotionally laden terms, drawing parallels to the inflammatory rhetoric used in pro-life circles to oppose abortion. This finding is contrasted against the existing corpus of MAiD research, where many providers show a secular understanding of the word [45,46,47]. Our findings aligned with the literature which has asserted that contemporary bioethical debate has erred toward abstractions that are divorced from real-world considerations; namely, the perspectives of physicians who are actively engaged in both practices [38, 48, 49].

Bioethics is a diverse field, and there is no general consensus among theorists about what method is most appropriate for dealing with the varied moral problems that arise within the field. Approaches that favor top-down applications of “high moral theories” such as consequentialism or deontology have been roundly critiqued for, among other things, being too abstract to provide helpful answers to fine-grained and complex ethical problems. Numerous concrete approaches have been recommended, including casuistry or case-based moral reasoning, narrative ethics, feminist theories and approaches that focus on moral experience. The results of this study suggest that health practitioners do not find abstract moral theorizing to be helpful in informing clinical moral judgments or in framing and interpreting clinical experience. More concrete approaches that explicitly attend to relational, cultural and contextual features, such as Sherwin’s relational theory [50] and Carnevale’s interpretive, “thick” framework for reconciling bioethical dilemmas [51] hold more promise as resources to help guide clinical moral reflection.

Limitations

An important limitation of this study is that recruitment for participation occurred through the CAMAP, a pan-Canadian pro-MAiD organization which supports clinicians involved in the practice, thus inclining the interview sample toward physicians who are involved in MAiD. Most participants who were not involved in MAiD were recruited through convenience sampling through connections with the research team. The potential overrepresentation of physicians involved in MAiD provisions through our recruitment measures may have resulted in a type of sampling bias. We recommend that future research in this area emphasize how involvement in MAiD provisions (or lack thereof) may sway clinician’s emotional responses and opinions toward the current bioethical debates.

As well, most of our participants were located in urban regions of Southern Ontario. Future research should balance the perspectives of both MAiD- and non-MAiD medical providers in connection with the dominant bioethical schools of thought, as well as expanding participation to include a larger variety of physicians across the nation, particularly those in rural regions. Lastly, the relevance of bioethical debates as they relate to MAiD and WLT may continue to evolve in the wake of Bill C-7, the removal of the foreseeable death eligibility criterion for MAiD which will expand to those suffering solely from a mental health condition in March 2024. Since our interviews were conducted in late 2020, our interviews do not capture this recent development, which should be explored by future studies.

Conclusion

This study has sought to contribute to the contemporary bioethical debate regarding the distinctions between MAiD and WLT in a novel way through exploring the personal perspectives of clinicians engaged with one or both of the procedures. The purpose was to advance the development of bioethical thought which is both informed by and of practical utility to physicians rather than an abstraction of academic philosophical thought. Our findings showed that the current core bioethical debates seeking to differentiate the practices are divorced from the real-world understandings of physicians involved in both procedures; namely, using outdated or offensive language, overemphasizing the distinction of a physician’s causal role in bringing about death, and intention underlying both practices. Overall, participants believed that these bioethical distinctions do not have practical utility at the bedside and must be re-conceptualized in collaboration between clinicians and bioethicists. The impact of such findings serves to develop a meaningful bioethical account of both MAiD and WLT that is based on the lived experience of providers of these services so that these debates are able to exert substantive impact on clinical practice and future health policy development.