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

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.

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].

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.

I don’t think I saw any MAID information over at the cancer centre. I don’t think so and it should be there, it needs to be there. (Family member #812)

The initial patient contact when that decision is made is critical. I think the conversation that happened … helped him move his decision making forward.…it’s not about handing somebody a pamphlet and saying, “Here, here’s what it’s about.” (Family #512)

“Right. If my doctor didn't want me to do it, then he should have referred me to somebody else. He should have just not said, you're not eligible. And to my way of thinking he was saying, just keep on suffering. Because I was…(Patient #821)

“Because I think most situations, or at least in mine, drawing the process out may cause a lot of undue stress and frustration.” (Patient #421)

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] both happened at the hospital. There was a really nice quiet room and they were pleasant. It was a conversation.. It was like having dinner with somebody without the dinner and very calm, no pressure … (Family member #812)

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)

They kept asking her if she wanted to change her mind. And that's a good thing, because it always gives her the opening. And they asked us if we wanted her to change her mind. And we all basically said, no it's her decision whether she wants to go.

(Family #822)

“I liked that when I was present at the assessment and I was there to support my sister, I liked that I got sent out so they could interview her alone to make sure that she was not being unduly influenced” (Family member #412)

The second interview, you know, [provided] a bit more detail about how it would actually be handled on the day, the injections and that type of thing. So, more general information that just gave me more information and made me relax more. (Patient #411)

If your life is scheduled and you’re given a timeframe to work with, you do the estate planning and settlements. I think that type of planning makes death a whole lot easier. (Family member #712)

Investigator: What makes the time between now and your death meaningful to you? Participant: Well, nothing. I wouldn’t say it was meaningful. The human instinct is to live as long as you can. It’s in there. You don’t know when it comes down to be your turn to die. As I said to my daughters even now, I’m dying to live but I’m not living to die.”(Patient #631)

You go up and down like an elevator, you know. Do I really understand this? Am I doing the right thing, and my family doing the right thing? Yeah, it's an elevator ride up and down, up and down. They just took all the fear away. I haven't been afraid of anything for the last couple months.....just a belief in what this program offers. I want it to get better but I like it the way it is so far. (Patient #911)

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.

She was happily surprised that she was further along to setting that date.....I think part of that was she was really worried she wouldn’t be competent to… You know, like it would be taken away from her. (Family #412)

Some patients and families described the waiting period as agonizing:

Oh I wish I was gone tomorrow…. But ten days. That’s a punishment. What for? I wasn’t that good a boy but that bad I wasn’t either. I’m lying here all day long waiting and waiting for the day to be over. And it’s a long, long day, you can believe me. When you wait for something. (Patient #521)

They offered me spiritual advisors, social workers, whatever I wanted. And they would assist me to get in whatever I needed if I wanted and I basically turned them down. I felt my belief system, my internal setup was secure and at peace with the decision I was making so not an issue. (Patient #411)

I think it’s a really good opportunity because most people, or a lot of people may feel that in choosing MAID it’s a very selfish thing to do. And this sort of helps balance that.(Patient #421)

I think it’s enviable being the author of your own last chapter, as it were. Not everybody has that opportunity to make that decision. (Family member #912)

He is the man for the job.....He knew exactly what he was doing and he knew exactly what to say and he knew exactly what to expect from people. We were totally happy and impressed with him. He just was totally on top of his game. He knew exactly, you know, what to say and do. (Family member #712)

Coordinating the timing of events precisely was important, as even small changes in the scheduled time could be a source of anxiety.

He made a noise after the first one that sort of all surprised us....Just a snoring noise sort of....That was a surprise to all of us. We sort of weren’t sure what was happening. Maybe mentioning something like that, you know, could happen. Then you don’t feel so thrown off at that point. (Family member #232)

In addition to the practice suggestions we provided, the two transcendent themes—coordination and patient-centred 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 long-term supports will be most useful for families [23‐25].

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.

This study has several strengths, including its multi-centre 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].