Use of palliative sedation following Medical Assistance in Dying (MAiD) legislation: A mixed-methods study of palliative care providers

Objectives. Palliative sedation (PS) and Medical Assistance in Dying (MAiD) are options for end-of-life (EOL) care in Canada, since the latter was legalized in 2016. Little research to date has explored the potential impact of MAiD on PS practices. This study investigated physicians’ perceptions of their practices surrounding PS and how they may have changed since 2016. Methods. A survey ( n = 37) and semi-structured interviews ( n = 23) were conducted with palliative care providers throughout Ontario. Questions focused on PS practices and explored potential changes following the implementation of MAiD. Codes were determined collab-oratively and applied line-by-line by 2 independent investigators. Survey responses were analyzed alongside interview transcripts and noted to be concordant. Themes were generated via reflexive thematic analysis. Results. Thematic analysis yielded the following themes: (1) Increased patient/family knowledge of EOL care; (2) More frequent/fulsome discussions; (3) Normalization/repositioning of PS; and (4) Conflation and differentiation of PS/MAiD. Across these themes, participants espousedincreasedpatient,family,andprovidercomfortwithPS,whichmaystemequallyfrom theadventofMAiDandthegrowthofpalliativecareingeneral.Participantsalsoemphasized that,followingMAiD,PSisviewedasalessradicalintervention. Significance of


Introduction
Palliative sedation (PS) and Medical Assistance in Dying (MAiD) are options for end-of-life (EOL) care in Canada, since the latter was legalized in 2016 (Carter v. Canada 2015;Criminal Code 1985). PS is a tool in managing refractory symptoms at EOL when prognosis is felt to be very short (generally under 2 weeks), wherein medications are used to deliberately reduce consciousness (CSPCP (Canadian Society of Palliative Care Physicians) 2017; Henry 2016). It has been defined as follows: "(1) the use of (a) pharmacological agent(s) to reduce consciousness; (2) reserved for the treatment of intolerable and refractory symptoms and (3) only considered in a patient who has been diagnosed with an advanced progressive illness" (Dean et al. 2012). In contrast to PS, which is not intended to hasten death, MAiD entails the deliberate provision of medications specifically to hasten death. MAiD has clearly outlined legal requirements including a serious illness, disease, or disability with an advanced state of irreversible decline and which is causing enduring physical or psychological suffering. MAiD can only be requested and consented to by the individual and requires the need for capacity at the time of the formal written request and at the time of eligibility assessments (GC (Government of Canada) 2022).
There is very limited literature addressing the impact of legalization of medically assisted death on PS. Two studies from the Netherlands found an increase in PS after the introduction of voluntary euthanasia (Rietjens et al. 2008; Van der Heide et al. 2007). Similarly, a prior 2 A. Mills et al. study at a tertiary care centre in Toronto, Canada, documented a rise in the use of PS after the implementation of MAiD in the Palliative Care Unit (PCU), though rates of PS remained stable in acute care (Nolen et al. 2022). Research more generally on the impact of MAiD on palliative care is similarly scarce and has not focused on PS (Mathews et al. 2021).
The goal of this study was to explore palliative care providers' perceptions of their practices surrounding PS, including the potential impact of the legalization of MAiD on the use of PS.

Methods
This mixed-methods study included both a survey and interview component, open to palliative care providers in Ontario, Canada, who had been in practice for at least 5 years. The survey was administered through Qualtrics. Recruitment proceeded via an invitation included in the newsletter for the Ontario Medical Association Section on Palliative Medicine (an email newsletter for physicians self-identifying as palliative care physicians). In addition to demographics, questions were designed to ascertain participants' views on their use and frequency of PS and how this might have changed over time (see Supplementary Appendix A for full list of questions). At the conclusion of the survey, respondents were invited to provide their contact information if interested in participating in the interview component of the study. Of note, having completed the survey was not a prerequisite for scheduling an interview as participants were also able to self-refer after learning of the study via word-of-mouth.
Semi-structured interviews were conducted with palliative care providers throughout Ontario (n = 23). The interviews were conducted over Zoom from March to July 2022 and were approximately 30 minutes in length. The interviewer was a Family Medicine resident with prior experience in interview-based studies. The interview guide was amended after the first 5 interviews following transcript review by the investigators. The final interview guide is attached as Supplementary Appendix B.
Participants were asked for permission to record the interviews which were subsequently transcribed verbatim, then individually reviewed for accuracy by the research team. Data collection and analysis proceeded simultaneously and iteratively (Charmaz and Belgrave 2012;Lingard et al. 2008;Merriam and Tisdell 2015). A series of codes and subcodes was determined collaboratively and applied to the transcribed interviews in a line-by-line fashion. Each interview was coded by 2 independent investigators and the codes were adjusted periodically; code saturation (Hennink et al. 2017) was reached after the first 4 interviews. Themes were generated via reflexive thematic analysis (Braun and Clarke 2019) and theme/meaning saturation was reached after the first 12 interviews, at which point recruitment efforts were ceased, though already-scheduled interviews were completed. As there was significant overlap between the data generated by the survey and interviews, and the research team felt there was general agreement, the comments from the survey were analyzed alongside the coded interviews, which changed neither the codes nor themes. For the purposes of knowledge translation, the available themes were separated into different categories; the paper at hand specifically presents themes addressing the relationship between MAiD and PS, whereas themes pertaining solely to the diversity of PS practices will be addressed separately.
This study was approved by the Sunnybrook Health Sciences Centre REB. All participants provided informed consent prior to participation.

Results -survey
The survey was available from March to May 2022 and was completed by 37 respondents (n = 4 incomplete responses). This represents a very small response rate in terms of the number of people who theoretically had access to the link (n = 1145); however, it is impossible to know the percentage of recipients who opened or engaged with the email newsletter. For the demographics of the survey results, see Table 1. With respect to frequency, 48.5% of respondents reported providing PS at the same rate compared to prior to 2016 and 33.3% answered "slightly more frequently. " The openended questions of the survey overlapped considerably with the questions asked in the semi-structured interviews. Particularly in light of the low response rate, these comments were analyzed alongside the interview transcripts and were noted to be concordant. As such, the themes presented in this paper represent both the survey and interviews. Survey comments are integrated throughout, particularly to ensure adequate representation of physicians who wanted to remain anonymous, and are attributed to Respondents (R) 1-33.

Results -interviews
Semi-structured interviews were conducted with 23 palliative care providers (see Table 2 for demographics). One provider was deemed to be ineligible on the basis of length of time practicing palliative care and was subsequently excluded from analysis. Broadly speaking, when asked about changes in frequency since 2016, 11 (48%) participants reported some degree of an increase in the use of PS; 9 of these participants endorsed an increase in their own practices, while the other 2 endorsed an increase generally, but felt their personal rates were unchanged. Of the remaining participants, 5 reported similar rates, 4 reported decreased rates, and 2 were uncertain. Quotations from interviews are attributed to Participants (P) 1-22. For all the themes presented below, please see additional supporting quotations under Table 3.

Results -themes
Theme 1: Patient/family knowledge of EOL care Several participants discussed an increase, not only in patient knowledge of EOL interventions but also in their willingness to initiate and participate in EOL conversations. Specifically, they noted that patients and families were increasingly familiar with MAiD. I think patients have become a lot more comfortable about it too. You know, there's a lot more in the media about it. Obviously, there's a lot more sort of public awareness of it.
[…] So I think everybody's evolved a bit over the past five or six years and that the legalisation of MAiD has probably facilitated that either directly or indirectly by sort of forcing us to have those conversations. (P10) While this increased awareness largely pertained to MAiD, some participants described an increase in requests for PS, or at least knowledge of it as an option, whereas others felt PS remained relatively unknown.
… when I said increased awareness, I'm thinking particularly about MAiD, I'm not entirely convinced that lay people understand palliative sedation as we would understand. (P22) In addition to increasing patient conversations with clinicians directly, participants also highlighted that the legalization of MAiD  Theme 2: Frequent and fulsome conversations The majority of participants reported having more frequent and in-depth conversations surrounding EOL interventions with their patients following the implementation of MAiD. Importantly, these conversations were not just patient-driven (secondary to the awareness discussed above) but clinician-driven as well. As with the above, inquiries about MAiD, originating either from patient interest or physicians discussing MAiD routinely, were often the inciting event for extensive conversations that included PS. Interestingly, in addition to discussing EOL interventions, some participants noted that these conversations often unearthed previously overlooked symptoms and suffering. Of note, although participants did discuss a causal relationship with respect to the implementation of MAiD, palliative care in general has continued to develop and crystalize as a discipline during this time. As such, more frequent Goals of Care discussions may be a natural consequence of the evolution of palliative care.
I think all of that is just growth in the whole world of palliative care, both with physicians and with patients and their families. (P16) I think there's a general awareness and comfort with it that's growing. I think that's related to training, like there's been a big push for increasing palliative care education in training programs and not just specialist palliative care (P17).   PS more palatable to some care providers P4: "There's some people who might prefer palliative sedation if, for religious, personal reasons MAiD is not acceptable." P16: "I certainly know when you have patients where for religious reasons they would never agree to MAiD, but they would see palliative sedation as being different and they'd be more comfortable with it."

Theme 4: Conflation and differentiation
Confusion understanding the difference between PS and MAiD: P5: "I think also there's a little bit of a feeling that [PS] is sort of akin to MAiD. Like it's in the grey zone, you know, not everybody's comfortable with that." P6: "I think nursing staff unfortunately don't get a lot of training in, you know, medical ethics, when it comes to things like palliative sedation. They confuse it with medical assistance in dying. There's a general lack of understanding amongst physicians, nursing staff, allied health, everybody about, the principle of double effect and what is and is not appropriate medical therapy." P9: "I think some patients more recently, I would say sometimes view palliative sedation as a form of MAiD. Like they haven't qualified for MAiD because of capacity. So, then they want, the family or POA want palliative sedation. I find that very challenging cause I don't view them as the same thing." PS and MAiD being discussed as equivalents P23: "…we felt at times there was this misconception by referring services that if somebody, there was a whiff of, they wanted MAiD, they can't consent so come do palliative sedation. And we would kind of re-center on this perception that palliative sedation was 'MAiD light"' and we wanted education on that. That there's a criteria for initiating palliative sedation." Does proceeding with MAiD constitute "failure" of palliation?
P7: "And I feel like I would be I never want someone to choose maid if it's for a symptom that I'm controlling poorly, so I can see that that impulse might be stronger for someone who doesn't provide maid because they have to sort of physically transfer care and sort of admit their 'failure' to somebody else as, this person is requesting MAiD." Theme 3: Normalization/repositioning After MAiD was legalized and implemented, many clinicians found that PS was viewed as a less extreme option -that it began to occupy the middle ground within the landscape of EOL interventions.
There's probably still some level of discomfort with MAiD among palliative care providers. And so some people may offer [palliative sedation] as an alternative that sits a little bit more comfortably with them. I think maybe it also, in our collective consciousness has kind of like opened up the idea that, there's a continuum to how people want to experience end-of-life. And so maybe palliative sedation, which was previously on one complete end of that continuum, now feels a little bit more in the middle. (P12) Several participants noted that this was true not only for physicians but also for the interprofessional health-care team: I notice that colleagues, in different professions, not just physicians, per se, but physicians and other health professionals who work on the various interdisciplinary teams and the levels of care that I work in are more readily bringing up palliative sedation since the institution of medical assistance in dying. This repositioning of PS, so that it is no longer seen as the most radical option, may also have implications in how practitioners view PS from a medicolegal standpoint. There may be a greater level of comfort on the part of the physicians that they aren't going to be questioned and possibly even disciplined or sued.
Finally, PS may be seen as a more palatable option for those with religious, moral, or personal objection to assisted death.
My experience suggests that patients/families see palliative sedation when a patient is close to the EoL as a preferred option -not only because it is "easier", but because some family members often have ethical conundrums about assisted dying, so PS is the easier ethical choice. (R27) Theme 4: Conflation and differentiation Several clinicians discussed instances wherein, despite the increased comfort and discussions highlighted above, there was confusion on the part of the patient, family, or interprofessional team regarding the differences between PS and MAiD.
There is still the misconception going back to the team and even family members. Oh, that means you are killing my loved one, or you're 6 A. Mills et al. killing the patient. Even though I try to be very, very clear that no, we are not. (P14) Many people, both laypersons and allied health workers, continue to confuse palliative sedation with MAiD, making it a challenge to propose and get buy-in from the team for palliative sedation. (R29) Participants who were also MAiD providers tended to find this confusion problematic when it came to discussing PS alongside MAiD for patients expressing interest in the latter. They particularly highlighted that other clinicians sometimes discussed the 2 interventions as though they were equivalent options for the patient: When patients and teams are discussing MAiD, I'm finding that teams are suggesting or offering or proposing palliative sedation as an alternative. And at the time when they're first just exploring, when it's just an exploration, when a person's not necessarily, making a MAiD decision […] I find that very problematic because I think that, that it is oversimplifying it as a treatment strategy, it's oversimplifying how it fits in, in terms of decisionmaking and the clinician decision-making around its use.
[…] Palliative sedation -it's not a means to death. It's a means to managing a symptom. They're distinct interventions that have very distinct paths to why they're being used. (P1) Of course, many participants discussed scenarios where patients originally had requested MAiD but, after becoming ineligible, went on to receive PS; they also highlighted that discussing PS could bring solace to patients worried about that exact scenario. These physicians had different practices regarding how to discuss the 2 interventions (e.g. routinely or only in the context of specific patient questions), but they stressed that PS and MAiD are distinct interventions and that conversation should be more complex and individualized than simply offering them in tandem.
I think there's increasingly a blended line between medical assistance in dying and the use of palliative sedation and, you know, if we can't get one, we'll just use the other kind of thing. That creates a lot of complicated ethical questions, which can ultimately lead to delays and barriers in the use of, say palliative sedation. [..]a lot of patients will say things like, oh, well it's okay if I can't get MAiD, then I'll just use palliative sedation as a backup kind of thing. And, well that's not the indication for palliative sedation. But they'll sometimes pressure us […], even if it's in the situation where we might otherwise not have even offered it. And that becomes a difficult conversation. (P17) I think that's a mistake that is happening right now because I think some people present palliative sedation as an alternative to MAiD, and it is not alternative to me. […]. MAiD is an intervention that leads to death. […] Palliative sedation, it's a medical intervention, used to manage and treat refractory suffering in patients.
[…] So unless the patient has met the criteria for refractory suffering, usually they're not eligible for palliative sedation. And for those patients, I tell them I can't really sedate someone just because they wanted to get MAiD. (P18) Finally, there were a few practitioners, particularly those who participate in MAiD, who raised a related and salient sentiment. These participants voiced that they might feel a sense of personal or professional disappointment if their patients make a formal MAiD request, as it can be taken as a reflection of the quality of the palliative care they have received up until that point. The same was not true of PS, which, as the above quotations showcase, was regarded as a "palliative care intervention. " As a MAiD provider, I make every effort to decrease the need for MAiD by practice, by trying to practice excellent high quality palliative care. I believe in the patient's right for an assessment […] for MAiD. And that's why I provide it. I want to address their suffering in every way possible. So it may be though that with me, I might work pretty hard at trying to address their palliative needs before it comes to MAiD. (P15)

Discussion
This exploratory study has revealed important information about how perceptions of PS among palliative care practitioners have evolved with the legalisation of MAiD in Canada and how this may influence their practice. Participants described an increase in knowledge and willingness of patients and caregivers to discuss the spectrum of EOL care options, driven in part by awareness of MAiD. There was a shared perception that patient inquiries about MAiD prompted more in-depth exploration of suffering and EOL preferences, including PS. Participants emphasized that the legalization of MAiD repositioned PS as a less radical option, thus increasing comfort with its use amongst the health-care team. They also described instances of confusion among the public and even health-care practitioners, whereby MAiD and PS were conflated and presented as equivalents.
Much has been published examining the use of PS worldwide, and previous studies have shown that the frequency and practice of PS varies significantly by geographic region and clinical setting. A systematic review of 10 studies examining the clinical practice of PS across all settings showed considerable variability in the proportion of patients receiving PS (mean frequency, 34%; range, 14.6-66.7%) (Maltoni et al. 2012). In Canada, 1 study from Calgary showed prevalence of PS prior to MAiD as 3.3% of deaths in acute care, 4% of deaths in hospice, and 22.2% of deaths in PCU (Abdul-Razzak et al. 2019). Another study from Toronto showed that PS was used in approximately 5% of deaths in acute care and the PCU prior to MAiD (Nolen et al. 2022).
Less is known about how the introduction of assisted dying legislation may in turn influence practitioners' use of PS. Clinical practice guidelines exist for PS (Cherny and Radbruch 2009;Dean et al. 2012) and there is a legal framework by which MAiD must be administered in Canada (Bill C-7 2020;Bill C-14 2016), thus positioning the 2 as ethically and legally distinct medical interventions. Despite this, the interplay between PS and MAiD in clinical practice is complex and not well understood (Koksvik et al. 2022). As discussed, limited data from Canada and the Netherlands suggest that rates of PS may have increased following the implementation of medically assisted death -at least in certain settings (e.g. PCU as opposed to acute care). There exist different possible explanations in the literature to account for this increase. Previous research from Belgium indicates that continuous sedation was sometimes offered to patients who had requested euthanasia but subsequently lost capacity, or was preferentially offered by the physician for moral reasons (Robijn et al. 2017;Seymour et al. 2015). Some practitioners also offered PS over euthanasia due to practical challenges in obtaining medications or completing necessary paperwork in timely fashion (Robijn et al. 2017). Other research has proposed a greater inclination to initiate PS due to an increase in patient awareness of all EOL choices in the post-MAiD era, and a clearer distinction the 2 entities (Nolen et al. 2022). Our study contributes the direct perspective of palliative care physicians who have experience administering PS both before and after the legalization of MAiD in Canada, in order to understand how MAiD has altered their practices.
Across the presented themes, palliative care providers clearly articulated increased comfort with PS on several levels, including both conceptual comfort with PS as an EOL intervention and comfort in their interactions with patients, families, and interprofessional colleagues. Part of this comfort is no doubt attributable to the passage of time in general as well as the development of palliative care as a discipline; however, most participants described a noticeable change following the implementation of MAiD in 2016. This increased familiarity, discussion, and potentially provision of PS appears to stem from both patient-and provider-driven factors, with 1 important mechanism being that patient inquiries surrounding MAiD can unveil unaddressed symptoms and lead to more fulsome Goals of Care discussions. Within provider factors, many interviews highlighted the potential amelioration of medicolegal concerns pertaining to PS, noting that understanding of PS has evolved in part due to its distinction from MAiD, with the attendant clarity that the goal of PS is not to hasten death.
One key finding of this study is the potential challenge in discussing PS and MAiD in tandem with one another for interested patients. There was significant variety in standard practices, including those would almost never discuss both and those who do so routinely (which may or may not be protocolized). Although the majority of participants confirmed certain scenarios wherein they would review PS in the context of a MAiD request -e.g. patients who might derive clinical benefit from knowing about theoretical alternatives -the physicians in this study were strongly opposed to treating them as direct equivalents, or offering both as concurrent choices, given the differences in their intent and eligibility criteria. As summarized by P1: "Palliative sedation -it's not a means to death. It's a means to managing a symptom. They're distinct interventions that have very distinct paths to why they're being used. " Nevertheless, many relayed a tendency on behalf of patients, families, and medical colleagues who were not MAiD providers, to conflate PS and MAiD. The imperative to separate PS and MAiD as distinct entities is reflected within the limited literature on this topic (Booker and Bruce 2020) and echoed in particular by MAiD providers in this study.

Conclusion
This study sought to elucidate provider perceptions on how practices of PS in Canada may have changed following the advent of MAiD -a topic which has not been researched to date. Salient themes included (1) Increased patient/family knowledge of EOL care; (2) More frequent/fulsome discussions; (3) Normalization/repositioning; and (4) Conflation and differentiation. Across these themes, participants espoused increased patient, family, and provider comfort with PS, which may stem equally from the advent of MAiD and passage of time in general. Overall, it is likely that both MAiD and the natural evolution of palliative care as a discipline have increased the extent to which providers discuss EOL options. A potential consequence of these more frequent discussions is the ideological conflation of PS and MAiD, which may pose a unique challenge for palliative care providers trying to navigate these distinct EOL interventions with patients. As participants in this study stressed, any request or inquiry into MAiD should prompt an individualized and nuanced conversation that explores all avenues of symptom management -the results of which may or may not include PS.

Limitations
As discussed, even though the results of the survey correlated well with the results of the interviews, the n = 37 of the survey component represents a very small response rate overall. Additionally, although there was good representation in terms of type and length of practice, recruitment strategies were solely focused on Ontario, even though MAiD is legislated federally. It is thus possible that results may not generalize to other Canadian provinces, let alone international jurisdictions.
Supplementary material. The supplementary material for this article can be found at https://doi.org/10.1017/S1478951523000706.

Competing interests.
The authors declare none.