How can we improve the experiences of patients and families who request medical assistance in dying? A multi-centre qualitative study

Background Medical assistance in dying has been available in Canada for 5 years, but it is unclear which practices contribute to high-quality care. We aimed to describe patient and family perspectives of quality of care for medical assistance in dying. Methods We conducted a multi-centre, qualitative descriptive study, including face to face or virtual one-hour interviews using a semi-structured guide. We interviewed 21 english-speaking patients found eligible for medical assistance in dying and 17 family members at four sites in Canada, between November 2017 and September 2019. Interviews were de-identified, and analyzed in an iterative process of thematic analysis. Results We identified 18 themes. Sixteen themes were related to a single step in the process of medical assistance in dying (MAID requests, MAID assessments, preparation for dying, death and aftercare). Two themes (coordination and patient-centred care) were theme consistently across multiple steps in the MAID process. From these themes, alongside participant recommendations, we developed clinical practice suggestions which can guide care. Conclusions Patients and families identified process-specific successes and challenges during the process of medical assistance in dying. Most importantly, they identified the need for care coordination and a patient-centred approach as central to high-quality care. More research is required to characterize which aspects of care most influence patient and family satisfaction. Supplementary Information The online version contains supplementary material available at 10.1186/s12904-021-00882-4.


Introduction
Since the passing of Bill C-14 in 2016, over 13,946 Canadians have received medical assistance in dying (MAID), representing approximately 2% of all deaths [1]. Bill C-14 legalized assisted dying for capable adult patients, with intolerable suffering whose death is "reasonably foreseeable, " and specified procedural safeguards, including the need for two witnesses of a written request for MAID; a 10 day waiting period; and confirmation of patient's right to withdraw request, including immediately prior to MAID. While Bill C-14 detailed criteria by which a patient could become eligible for MAID, it provided little detail on how MAID should be implemented in clinical practice, and as a result it allows both assisted suicide (in which a patient self-administers a fatal dose of medication) and voluntary euthanasia (in which a clinician administers a fatal dose) [2,3]. Though the MAID process has a straightforward technical outcome (a rapid and painless death), understanding whether or not the process leading to that outcome has been of high quality is subjective, and must be understood by exploring the experiences and perceptions of those who take part, particularly patients and families, as described in a recent systematic review [4]. While new Canadian Federal reporting requirements for MAID took effect in 2018, the collected data is broadly descriptive, and little to no indepth data on the patient and family member experiences of MAID to guide clinical practice and guide research [1]. To address this gap, we conducted an exploratory qualitative study to understand patient, family, and clinician perspectives on quality of MAID care.

Setting and participants
The research team use a pre-specified protocol to guide study procedures, including recruitment, eligibility, interview process, and analysis. We conducted interviews with MAID-eligible patients, their families, and health care providers at four Canadian centres, two academic hospital-based with interdisciplinary teams (Hamilton and Ottawa), one a community hospital program lead by nurse practitioners (Niagara) and one an outpatient clinic (Vancouver). This manuscript includes patient and family descriptions of the quality of care during the MAID process. We reported HCP interviews separately as these described clinician-specific operational and institutional aspects of MAID care, such as remuneration, scheduling, and institutional supports which would only indirectly impact patient and family experiences [5].
After approval by the research ethics board (HIREB 3991 November 2017), clinicians at participated study sites asked MAID patients if they would be willing to be approached for research. If so, we invited patients who had been found eligible for MAID to participate in the study. If they agreed to take part, we subsequently invited one family member (as chosen by the patient who could be biologically related or not) and health care provider (preferentially one of the MAID assessors or providers, though we also included nurses, social workers, and managers who were directly involved in the patient's MAID process) to participate as well. At least two of the three potential participants had to provide documented written or verbal consent to an interview to be included in the study. The only exclusion criteria was if a potential participant was unable to speak and read English. Our purposive sampling strategy was "typical" case sampling, aiming to recruit people who could provide experiential knowledge across the depth and breadth of the most common MAID experiences, as opposed to purposely recruiting 'extraordinary' or 'unusual' cases.

Data collection
We developed a semi-structured interview guide which aimed to elicit participant perspectives of quality of care throughout the MAID process, and sufficiently general to apply across settings and a variety of care models. The guide was developed by the research team (including both non-clinician researchers and MAID clinicians), and asked participants about their experiences, facilitators and barriers to accessing MAID, and perceptions of what made for a positive MAID experience. The interview guide was updated iteratively during the study design and REB approvals process, however only minor wording changes were made after the first interview, as it appeared to elicit a broad range of experiences sufficient for our analysis (Supplementary Material). Investigators at each site without a relationship to the participant (DEC, SO, DC, MK) conducted and recorded interviews in person or by phone as per participant's preference. Patients were interviewed shortly after study enrolment, while families were interviewed at a time of their choice after the patient's death, usually 4-8 weeks later. At the time of interviews, participants also completed a basic demographic form. Recordings were transcribed, de-identified, and uploaded to Dedoose, a cloud-based qualitative data management program, for organization and analysis [6]. To allow for adequate breadth and depth of data, the target sample size was 60 interviews, including 20 patients, 20 families, and 20 HCPs, which we estimated would allow us to reach thematic saturation [7,8].

Data analysis
Aiming to provide high descriptive and interpretive validity, so as to identify practical improvements to the MAID clinical process, we used a qualitative descriptive approach of interview data. Given the legislative requirements for MAID and their chronology (MAID request; MAID assessments; preparation for dying; death and aftercare) we organized codes and themes according to this sequence [9,10]. The primary analysis team (SO,DEC, PA) conducted open coding on the first three interviews to develop an initial codebook, including rules for when specific codes should be applied to the data. Analysis continued alongside participant recruitment and interviews, with the team meeting periodically to review new interview results and revise the codebook as new findings emerged. In an iterative process of thematic analysis, we developed primary and secondary themes related to perceptions of quality of care within our theoretical chronology of the MAID process [11]. To assist with knowledge translation of these findings, we developed practice suggestions from two sources, either direct study participant recommendations, or inferences based upon participant descriptions of successes and challenges in the care process [12]. As the study participants were either deceased or grieving, we did not use member-checking or allow for participant review of transcripts.
We incorporated several strategies to increase trustworthiness of our results [13]. Confirmability is enhanced by its use of a pre-specified protocol, interview guide, and adherence to reporting guidelines (COREQ); credibility by its use of multiple interviewers including clinicians and non-clinicians for triangulation during analysis, as well as its descriptive analytical approach; and transferability and dependability applicability by its multi-centre recruitment strategy; inclusion of hospital and the community settings.

Results
Between November 2017 and September 2019, we recruited 24 patients and 17 family members to participate in the study. Twenty-one patients and all 17 family members completed an interview, meeting our goals for thematic saturation (identifying no new codes or themes) with fewer families than anticipated. In all 17 family interviews, the corresponding patient was interviewed; in the four remaining interviews, the patient and a HCP (but not the family) provided interviews. In one case, the patient died before interviews were completed; in the others, only the HCP completed an interview. Participant demographics are reported in Table 1. Similar to general demographics of Canadian MAID patients, cancer was the most common qualifying condition for MAID (n = 15, 63%), followed by cardiac or respiratory disease (n = 4 17%) and neurologic disease (n = 3, 13%). Two patients (8%) had other conditions. The patients all had significant functional limitation, with Eastern Cooperative Oncology Group grade of 2 (n = 3, 13%), 3 (n = 8, 33%), or 4 (n = 8, 33%) [14]. In contrast to patients, who were roughly evenly divided between genders (n = 11, 46% female; n = 13, 54% male), family members who reported gender were most often female (n = 13, 76%).
We organized themes related to perceptions of quality of care as either "process" themes related to a specific step of the MAID process as outlined in the interview guide (MAID request; MAID assessment; preparation for death; death and aftercare), or as a "transcendent" theme which which themed across multiple steps in MAID process (Fig. 1). Below, we summarize these major themes, and provide exemplar quotes. A complete list of themes, sub-themes, and exemplar quotes can be found in Supplementary Material 2. We summarize the practice suggestions in Table 2.

Themes related to MAID requests
i) Promote MAID awarenessParticipants described the need for efforts to raise awareness of MAID both in the general community and amongst clinicians.
Patients required opportunities to be exposed to the possibility of MAID, as for many it was not clear if it was available, or how it could be accessed: iii) Facilitate written requests. Patients described struggles in identifying two independent witnesses for the written request, and those who relied upon friends and family were cognizant of the emotional burden this could put upon the witnesses.
iv) Move MAID requests forward quickly. Early delays were perceived to be due to the actions of the clinical team, due to either inefficiencies or even purposeful obstruction, highlighting the need for an efficient referral system for MAID requests:

Themes related to MAID assessments
i) Assess in a safe, comfortable space and ii) Assessors need to build trust.
The clinical purpose of MAID assessments is to determine whether or not a patient meets eligibility criteria for an assisted death. The exploration of a patient's suffering and their decision to pursue MAID, is a complex and sensitive activity. High quality assessments require trust of the clinician and comfort with the setting, whether it is an in person or virtual/telephone assessment: The assessments were cognitively and emotionally challenging; patients described feeling the pressure to perform for the assessors, to maximize their chance of being eligible.
"I just wanted to give as much information as I could to persuade them that this was the right thing for me" (Patient #411) iii) Provide transparency around procedural safeguards.
A family member or friend would sometimes be present during an assessment, described as a valuable experience for both patients and families as a support mechanism. Both parties appreciated the transparency around the purpose of the assessment, and the safeguards built into in the process: iii) Support patient and family reflection. While awaiting the MAID death, patients and families reflected frequently on death and the decision for MAID, and the tradeoff between avoiding suffering and reduced life:   These reflections were important for patients to decide if MAID was the right option for them, and for families to come to terms with the patient's decision for a hastened death.
One major concern was that participants would lose capacity, and therefore become ineligible for MAID. Intertwined with this worry were decisions around symptom control, with patients and families sometimes deciding to withhold pain medications to avoid sedating effects and the development of delirium. Coordinating the timing of events precisely was important, as even small changes in the scheduled time could be a source of anxiety.

iii)
Prepare the family for the patient's death and offer support afterwards. Family members valued education on what the patient's death could look like, as many had not been present at the time of death of a loved one, or were unaware of how a MAID death would look and could feel. Others describe the value of clinicians providing support and information after the patient had died:

Two transcendent themes: coordination and patient-centred care
Alongside these four "process" themes and sub-themes, two themes were developed across multiple steps of the MAID process. The first, coordination involved descriptions of how the complex MAID process was carried out in a smooth, professional, and predictable sequence. Central to this was continuity of care, as patients transitioned between sites, steps in the MAID process, or between health care teams. One of the central roles of the clinician or team coordinating MAID care was providing clear and consistent education and managing expectations. Patients and families were grateful when HCPs supported and facilitated access to MAID: I have to say, I never expected that level of kindness and cooperation, and information. It's like an information highway. You want to know something, we'll tell you, just ask. . (1-1-009) The second theme which transcended all steps in the MAID process was patient-centered care. Central to this was exploring a patient and family's previous experiences with death; individual needs and preferences, and restoring dignity by providing human connection and compassionate care: Crucially, this also involved taking care of the whole family: patients described the importance of their families having a positive experience and support after the MAID death. Caring for the patient entailed caring for the family as well.
Investigator: What was the most important thing that the healthcare team could do to help patients?

Participant: Look after my family. (Patient #131)
Families acknowledged that supporting their loved one through the MAID process as a rewarding but emotionally complicated experience, and the need for counselling and bereavement services:

Discussion
We identified sixteen process themes and two transcendent themes related to patient and family perceptions of quality of care for MAID, along with a list of suggested practices, summarized in Table 2. While not supported by quantitative evidence, these suggestions are based upon the experiences of patients and family members who have been through the MAID process. Health care clinicians and organizations in Canada may use these to guide practice, while researchers may use them to inform research. For international readers, these results can still be applied to practice (if legislative parallels exist). These results can also assist readers in jurisdictions where MAID is not legal to understand the perspectives and needs of patients who request a hastened death, and perhaps to alleviate some aspects of their suffering.

What this study adds
In addition to the practice suggestions we provided, the two transcendent themes-coordination and patientcentred care-are particularly noteworthy, as these speak broadly to the opportunities and challenges posed by MAID as a therapeutic intervention. The complex medico-legal nature of MAID requires considerable organization on the part of patients, families, and health care teams. This coordination straddles traditional health care silos such as community vs. hospital settings, or specific clinical programs. MAID is a novel clinical practice which requires dedicated organization to ensure smooth transitions in care, completion of paperwork and reporting, coordination of the numerous decisions leading to the patient's death. None of these activities clearly fall within a traditional clinical role. While some individual clinicians may function without dedicated MAID coordination services -either because their practice is dedicated to MAID or case volumes are very small-most are likely to benefit from the formalization of the role. Study participants' responses implied that a lack of coordination in the MAID process could be hugely disruptive, potentially resulting in prolonged suffering or loss of eligibility for MAID. Poorly coordinated care is unlikely to lead to a satisfying experience for any participant in MAID-the patient, the family, or the clinical team. The widespread formation of MAID care coordination services and the MAID coordinator role are thus likely to be a permanent, rather than a transitional aspect of MAID in many settings [15].
We identified a complex relationship between patients, families, and the health care team which suggests a need to recalibrate what "patient-centred" means in the context of assisted dying. Patients who request MAID are often concerned with the impact of their choice upon others, especially their family and friends. Decision-making around assisted dying is often framed in terms of either the physician-patient dyad or patient-family dyad [16]; by contrast we noted in our study that decision-making was more often a triad, involving all three of patient, family, and clinician. In striving for the best possible death experience, participants described a need to compromise-patients suffer longer to allow families to prepare for their death; families support their loved one's choice to die sooner; and clinicians bend policy to provide compassionate care to patients and families. The complex challenge of providing "patient-centred" care lies in balancing these conflicting values, needs, and preferences [17]. While patient-centred care is held as an ideal, reality may be a more complicated compromise between all three parties in MAID decisions. Fostering an honest discussion of values with all participants, rather than simply following on the patient's stated choices, is essential to creating a patient-centred MAID experience, even if it is ultimately the patients capable choices which clinicians strive to uphold [18].
A systematic review by Gamondi et al. summarized a number of studies describing the cognitive and emotional responses of family members to a loved one's assisted death, and be unique aspects of bereavement and aftercare in this context [4]. However, there are relatively few studies describing the perspectives of patients and families on procedural aspects of care, possibly due to the heterogenous medicolegal contexts of assisted dying around the world. Comparisons with other studies suggests there is some transferability of this study's findings, especially in Canada. Brown et al. conducted a qualitative study including 5 patients, 11 family members, and 14 health care providers in Saskatchewan, identifying sustainability, pathway ambiguity, lack of support for care choices, institutional conscientious objection, navigating care, post-death documentation, and legal stipulations as being primary procedural concerns of patients and families [19]. This study reaffirms that ambiguity around care processes and struggles coordinating care are major challenges faced by patients and families. A separate report of the same interviews explored considerations for providing patient and family centred care which are also align with those in the this study [20]. The present study builds upon this work by integrating both procedural and patient-centred aspects of MAID, and providing concrete practice suggestions.
This study's results also parallel those of Hales et al., who conducted a multi-methods study including 11 family members [21]. They identified six areas for improvement, 3 organizational (process clarity, scheduling challenges, 10-day reflection period) and 3 experiential (clinician judgement/objection, secrecy/privacy, and bereavement support). These correspond closely to several themes we identified. While family bereavement and counselling were mentioned, they were not as prominent in this study, possibly because the interviews with families occurred fairly quickly after the patient's death and interviews focused on providing support immediately around the patient's death, rather than long-term bereavement support (6-8 weeks). Participants perspectives of MAID may change with time after the death of a loved one, with bereavement playing a larger role in perceptions of quality after more time has passed [22]. Future research will be useful to determine which longterm supports will be most useful for families [23][24][25].
The coherence of results between this study, and others conducted in other Canadian centres suggests that the experiences and expectations of Canadian MAID patients and their family members may be similar in a variety of practice settings, with some caveats (described below). The results of this study will be used to develop patient and family satisfaction surveys, which will undergo testing at multiple Canadian centres and thus test this hypothesis. If the surveys demonstrate adequately robust psychometric properties, they will be useful in assisting clinicians and institutions in MAID research and quality improvement. Other important research topics can include testing communication strategies, screening tools for patients who may be interested in MAID, and bereavement supports for families.

Strengths and weaknesses of the study
This study has several strengths, including its multicentre recruitment strategy, which aimed to recruit a broad but representative sample of both MAID patients and their families. This is supported by the participant demographics, which are consistent with those reported by the Canadian Federal government, suggesting we interviewed a reasonably broad sample of Canadian MAID patients and families according to these criteria [1]. As well, by recruiting patients and families, we were able to identify practical areas to improve the MAID process from the perspective of both groups. Weaknesses of this study include the large proportion of respondents from Ontario, lack of representation from rural and remote communities, lack of data collection on patient and family race and ethnicity, and inability to account for the impact of centres' different practice models, care teams, and patient populations. These results may thus not be transferrable within specific socio-demographic contexts. Other weaknesses include the use of telephone/video for some interviews, which may have enhanced recruitment but also have had an impact upon participant contributions to the study. As well, we did not conduct member-checking with families to ensure our interpretations were consistent and resonant with their experiences, although bereaved relatives in MAID may have been willing to participate in this and long-term follow-up, as they are in other palliative care research [26]. Lastly, while we aimed for a low-inference, qualitative descriptive approach to our data, there is a significant interpretive aspect to our use of thematic analysis as we applied it to the study objective, and some would argue that content analysis would be a preferred analytic approach [27].

Conclusion
Patients and families described a variety of successes and challenges in the provision of high quality care during the MAID process. We identified numerous opportunities to enhance the experiences of those involved in a patient's assisted death.