Hearing, sharing and applying: a qualitative study of the experiences of family caregivers’ end-of-life decision-making in Singapore

EOL decision-making in Asia

Two recent studies in Asia describe the EOL decision-making practices in Singapore and Korea. When care preferences were not in initial agreement, about half of the patients were able to convince family caregivers to go along with their choices.6 Here, EOL decisions were made bilaterally in consultation with both patients and family caregivers.

Among 990 patient-caregiver dyads in Korea,7 it was found that most patients (82.9%) and caregivers (89.6%) preferred bilateral decision-making. While there was low agreement about who was leading the decision-making, both parties nevertheless preferred bilateral decision-making within the family unit, particularly where it was characterised by collaboration and communication among family members. In Singapore, activities of daily living such as bathing or feeding are commonly outsourced to community services or a foreign domestic worker to reduce caregiver burden. Nevertheless, decision-making is a distinct care burden that remains vested within the family caregiver’s responsibility because it requires authorisation and personal knowledge, a responsibility that is not easily transferable.8 However, to our knowledge, the literature has not previously addressed how to support such families in reaching a consensus, or what factors affect their satisfaction or dissatisfaction with the decisions they have made.

In Singapore, the Mental Capacity Act requires that healthcare providers make the final care decision when patients lack capacity.9 However, providers are still encouraged to engage caregivers to elicit the patient’s preferences. Since 2011, the Agency for Integrated Care, set-up by Singapore’s Ministry of Health, has been promoting advance care planning to Singaporeans. The agency advocates early conversations and shared decision-making between individuals, caregivers and healthcare providers on one’s care preferences in the event one becomes seriously ill and unable to speak for oneself. However, the programme has not gained sufficient traction in the country owing to low public awareness and difficulties experienced by healthcare workers in documenting the care plans among competing clinical priorities.10 Even when presented with the space for these conversations in the care setting, challenges remain because doctors are unevenly skilled in conducting these conversations,11 patients may change their mind regarding their preference,12 simply defer to medical authority or their adult children and the previously mentioned Asian context that could enhance the pressure on family caregivers to choose life-extension by default.

The EOL decision-making process may have implications on caregivers’ well-being, long after their loved ones have passed. A prospective multicentre study in the USA showed that unresolved decisional burden can lead to ‘moderate or high levels of decisional conflict, which is associated with ‘decision delay’, ‘feeling uncertain’ and lower quality of life.13 Studies like this have resulted in more attention on reducing decisional burden among caregivers through the development of decisional aids and care planning tools in the USA.14 We understand that similar caregiver-focused interventions have yet to be developed in Singapore.

Developing these interventions for Singapore may be important, as they have been shown to result in less aggressive care at the EOL, better patient quality of life and caregiver bereavement adjustment. In addition, having an earlier discussion offered more timely referrals for EOL care, such as hospice placement.14

Study objectives

The preceding literature suggests that in Singapore, the predominant model of EOL decision-making may be bilateral between patient and proxy decision-makers. For those who take on EOL decision-making duties, there may be significant burden but little is known about how these burdens are managed in Singapore, or what may be appropriate interventions for these Asian caregivers. While there is growing recognition of the diversity in EOL caregiving roles, few studies have provided an Asian viewpoint on the decision-making function of family caregivers, how they come to undertake this decision-making role, their decision-making journey and the ways in which they facilitate discussion with the patient, family members and the healthcare team to arrive at a satisfactory conclusion.

With the research focus in Singapore shifting towards supporting caregivers to feel better prepared for care decisions,15 more needs to be understood about the unique relational dynamics of local family units and their norms and practices surrounding EOL decision-making. Hence, this study examines how family caregivers make EOL care decisions, by focusing on relevant moments in their caregiving journey such as information communication and preference elicitation from patients.

Participants and recruitment

Participants were recruited via convenience sampling. The inclusion criteria were as follows: (1) English-literate; (2) no cognitive impairment; (3) was significantly involved as a family caregiver of someone who was seriously ill; (4) has been involved in EOL decision-making (eg, whether or not to resuscitate in the event of heart-stopping) OR had a patient pass way in Singapore, (5) the patient was 40 years old and above at the point of caregiving, AND (6) if participants were recently bereaved (patient passed away less than 3 months ago at point of recruitment), they must bring a ‘trusted other’ for the interview (this was encouraged but optional for other participants who were not recently bereaved). ‘Trusted others’ refer to individuals who would join the target research participants in the research study interview as emotional support. As the ‘trusted other’ could not be excluded from the audio recording if they contributed any verbal content, their consent as research participants was also taken for the research. Individuals who did not agree to audio recording were excluded from the study. The participant criteria were selected for patients who were 40 years old and above as it is the age at which chronic illness screening begins in Singapore, coinciding with the early onset of adult chronic disease. Hence, the sample should include caregivers caring for loved ones who are middle-aged or older. This is also congruent with the literature review that largely discusses the care of older persons.

Participants were recruited via (a) a recruitment channel hosted on the Telegram application; (b) a recruitment platform hosted on our university’s online learning website; (c) personal contacts; and (d) invitations from participants recruited via (a), (b) or (c). In total, 23 individuals had registered their interest to participate (online supplemental appendix A). Of these, five did not meet the eligibility criteria, did not respond or later withdrew before consent-taking for the interview took place. 18 participants consented to be interviewed but only 16 participants were eventually interviewed. However, one participant’s audio recording was too noisy for transcription and could not be analysed. Of the 15 remaining participants, 14 were caregivers and 1 was a ‘trusted other’.

The demographic characteristics of the 14 caregivers who were interviewed and whose audio recordings were later transcribed for analysis are detailed in online supplemental appendix B. Of the 14 caregivers, 12 were adult children caring for their parents or grandparents, while 2 were spousal caregivers. Of the 14 caregivers, 11 were women and 3 men. All caregivers were Chinese, except for one who was Indian. The age range of caregivers was 20–29 years (29%), 40–49 years (29%), 50–59 years (21%) and 60–69 years (21%).

Data collection

Semi-structured interviews were conducted with 14 participants in English, over ‘Zoom’, a teleconferencing application. An interview guide (online supplemental appendix C) was used for the interviews, which lasted approximately 60–90 min each. Participants were asked to share the barriers and enablers they had experienced as caregivers preparing for EOL decision-making for their loved ones. While there was concern the presence of trusted others might restrict how open the interviewees might be, ultimately only one participant brought a trusted other, who did not speak much. The bulk of the evidence comes from one to one interviews. Only the audio recording from each interview was retained for transcription and analysis. In accordance with Saunder’s definition of data saturation,16 data collection ceased when no new themes or subthemes were generated from additional interviews.

Ethics

As part of consent-taking, participants were given a Qualtrics website link which contained the participant information sheet (online supplemental appendix D). All personally identifiable data was anonymised, and each participant was assigned a code (eg, P1). Mobile banking details (preferred mobile baking platform and mobile number) that were collected for the purpose of reimbursement (~US$7) were deleted after payments were effected.

Patient and public involvement

It was not appropriate to involve patients in the study design, conduct, report or disseminate plans of our study, as the study’s focus was on caregivers’ experiences. Further, most of the patients had passed on at the time of the interview.

Data analysis

De-identified audio recordings were machine transcribed by Otter.ai, then checked for accuracy. Participants were categorised based on1 the care roles they reported involvement in: (1) emotional support (n=11), (2) financial support (n=7), (3) physical care (n=8) and (4) decision-making (n=13), as defined by the caregiver,2 whether they described themselves in the interview as lead facilitators (n=8) or non-lead facilitators (n=6) in the decision-making process and3 the illness trajectory of their patient.17 A six-phase reflexive thematic analysis approach was applied to the transcripts, including (1) familiarisation; (2) generating initial codes; (3) searching for themes; (4) reviewing themes; (5) defining and naming themes and (6) producing the report.18 Given that the literature on caregiver EOL decision-making is not well-established, a largely inductive approach was taken in the coding process, with no specific conceptual framework or intent to theory-test in mind. The coding was done manually by MLO and relevant quotes were later transferred to a Microsoft Excel document. The first author coded five number of interviews first. The research team then met to discuss and refine the codes. The resulting codebook was applied to the remaining interviews in uncovering the central themes. Code suitability was assessed and amendments were made to the codebook when needed. The epistemological lens for this study runs closest to phenomenology, in its attempt to explore the unique experiences of caregivers.

Theme 1: understanding and accepting patient’s condition and options

Many caregivers expressed that the opportunity to embark on EOL discussions with family members and doctors was essential to arrive at a satisfactory decision. This observation was underscored by the experiences of several caregivers who faced limitations in visiting their family members due to COVID-19 restrictions in healthcare institutions just prior to the time of this study. This group of caregivers reported feeling that their loved ones were being de-prioritised by the healthcare team and had limited access to talk through care options with the doctor. Thus, an enabler of EOL decision-making would be communication access to the healthcare team, while limited or no access to the healthcare team constituted a barrier to EOL decision-making.

The family’s shared decision-making processes were aided by recommendations from the attending doctor or medical team, which helped to inform the eventual EOL decision. This was experienced by P6:

P6: …either to keep her on ventilator but the doctor said unlikely she will wake up ever again… The other option, which the doctor recommends. is more practical to let her go peacefully.

(P6_Male_20–29 years old_Chinese)

Having a clear prognosis from the medical team aided the consensus-building in EOL decision-making among the family members. For example, at each hospitalisation, the doctor for P3’s mother conveyed his mother’s diagnosis clearly and made recommendations, with the final decision being made by P3 after consultation with his mother and family members.

Conversely, not being told of the prognosis by the doctor was disempowering for the caregiver as this absence of communication meant an inefficient information conveyance. Further, the lack of a clear prognosis, delayed prognosis or unexpected prognosis impeded consensus-building in EOL decision-making among family members. Caregivers who were affected by this expressed their wish to have sufficient time and thinking space to make consensual decisions for their patients, and experienced regret and guilt when they could not. According to P4, she was not allowed to talk to her patient’s attending doctor over the phone but was allowed to speak with a nurse. This information asymmetry led P4 to bring her patient to a private doctor for a second opinion.

P4: Yeah, the first time the first time like when this (referring to COVID-19 pandemic) escalated right, I couldn’t go up (to the ward) at all. They (the healthcare team) didn’t think it was very critical. So I did not have the chance to go and talk to the doctor and everything …

…I felt that maybe if I if I was able to go and talk to the doctor, it might have been slightly better like we could have changed to another drug quicker.

(P4_Female_20–29 years old _Chinese)

Aside from clinical expertise, another source of prognostication was referencing past EOL experiences. Whether with people they had cared for directly, or observations of others who have passed, these experiences were used as reference points in the current EOL decision-making process. Caregivers who were healthcare workers had a particular advantage in this aspect of previous experience. They possessed prior knowledge and/or experiences relating to EOL treatment options and had higher acceptance of the patient’s terminal prognosis. Such is the case of this caregiver who worked as a healthcare worker in the operating theatre:

P1: So we all agree, but before that because I’m a healthcare worker, so the family members turn to me. “How, how, what do you think?”. So I said just listen to what the doctor says, it’s logical. Just let her feel comfortable. So then all agreed, so no dispute no dispute. So really thank God so then so towards the end then they give a morphine to relieve pain and let her drift off no CPR so that’s how she passed on. Because layman still think that there is hope. But because we have seen it in operating theatre. And the surgeon agree with those cases of operation it’s because the family members think that there is still hope so they let them have that hope whereas I’ve seen it, seen enough of it, I know what is it. In the end, its just extending. That life also will not be … quality life… the extension of that you know, … so no point.

(P1_Female_60–69_Chinese)

In contrast, the absence of past experiences meant caregivers may find it hard to accept a terminal prognosis. For instance, there were two caregivers who were especially pained by EOL decisions to not prolong life as they had hoped for a miracle, had no reference point from patients on their EOL decisions nor had they ever reflected on EOL options as caregivers. Further, these caregivers witnessed their loved ones undergo a sudden death, where the latter were still relatively well prior, but deteriorated suddenly and passed away within a week or so of being hospitalised. The interview excerpt below details the pain experienced by one caregiver:

P2: To be very frank, I’ve ever heard about death but I’ve never thought my own parents, my family, there’d be death. But I’ve never once ever thought that my own parents one day will leave us or anybody in my family would have left us. I've never thought of that. To me it’s always we are together. So when my mother left us, it was a big blow for me. It was a big blow for me.

Interviewer: So you’re the last person right that she saw. So when the decision-making was happening to not proceed for surgery, do you feel uncertain?

P2: Yeah I was.

Interviewer: Okay. What was your uncertainty at the time?

P2: To me, it’s like, I had this feeling she will survive, you see? But my family didn’t want her to go through the pain that’s all. So until now I still carry the guilt that if only I could have been firm about it, maybe she could be alive until today.

(P2_Female_50–59 years old_Chinese)

Theme 2: achieving family consensus alongside ongoing reflection of patient and/or family’s wishes

Caregivers seemed to experience particular satisfaction with the decisions they made when there is a sense of consensus or harmony among family members. Among our participants, there was an overwhelming inclination for multilateral, shared decision-making in EOL treatment and care options within the family unit. Such shared decision-making was instrumental towards consensus-building for the EOL treatment and care decisions made by the caregivers in consultation with or on behalf of their patients.

Among our study participants, there was no case of a unilateral decision made by a single family member. Having a tight-knit, larger immediate and/or extended family base encouraged shared decision-making that led to mutually agreeable EOL decisions for patients.

For example, during family discussions details such as financial considerations and the degree of caregiving burden associated with the care option were tabled. This open discussion was instrumental to helping P3 and P5 reach EOL decisions for their patient that they were personally comfortable with and also mutually agreeable within the family unit. This point can be aptly summed up with P1’s description: ‘…there was no dispute, no regret…we are all on the same page’.

Caregivers also took reference from their mutual experience of past family events in forming and norming the values and preferences of patients and family members. The quotes from P5 and P10 illustrate how witnessing the prolonged terminal condition of an elderly loved one had led to the evolution of values and preferences. In the case of P10, this came to bear when her own father turned ill, and expressed a desire not to prolong life. This wish was aligned with the family’s, and ultimately came to pass.

P5: I can tell like from my mom’s side, it’s more like, like the ethos is okay, let’s keep the person like alive for as long as possible. You know, just like her whole family also decided as a as a unit to keep my grandfather alive for like the past to keep resuscitating and like, you know, treating him for the past like four years of his life. Maybe a month or two ago, we were talking about my late grandfather, who was vegetative for maybe the last like, four years of his life. He had a hospital bed in his house, then they had 2 maids taking care of him. Oh, okay. So, then my dad kind of expressed, like, like, he just kind of said that. It’s quite a sad state to be like that for your last four years. Especially when you’re just like waiting to die. It just it feels quite… yeah, it’s quite like, difficult. You know, seeing someone you love like that.

(P5_Female_20–29 years old _Chinese)

P10: Decision like NG (nasogastric) tube and where he wants to pass on right? My mum didn’t bring this out. But really, because we have seen our grandfather bedridden and on NG tube. So it’s very clear. My dad was when he was diagnosed at stage two cancer, he already say very clear. He doesn’t want any tube inside his nose.

(P10_Female_40–49 years old _Chinese)

Overall, the analysis of the data indicates that having ongoing reflection and discussion of the values and preferences within a tight-knit family unit positively affected the eventual EOL decision-making process for the terminally ill patient.

Theme 3: families often have a lead facilitator whose role is to increase communication and collaborative decision-making

This theme arose when caregivers described a distinctive function in the decision-making process, where an individual in the family consistently facilitates discussion between the patient and the wider caregiving structure. This individual was responsible for coordinating information sharing and decision-making within the family unit to arrive at a satisfactory EOL decision for their loved one. Table 2 details the characteristics that suggest why a particular caregiver leads. We chose the term ‘lead facilitator’, as opposed to ‘main decision maker’, because the person operates with the context of multilateral decision-making.

Table 3 summarises the types of activities that were expected of the person taking on this role.

Typically, the spouse would be the default person who took on this role (such as P13, whose spouse eventually passed on due to breast cancer), especially if their children were still young. However, when the children were adults, or in the absence of a spouse (ie, the patient is single, widowed or divorced) the adult children or sibling tended to step in to facilitate the discussion.

The interviewees gave various reasons for a person to take on the role of the lead facilitator (table 2). For example, some came to assume the role due to traditional reasons such as seniority or gender.

On the other hand, lead facilitators perceived that they were thrust into this role because they were trusted by the patient who felt that the former understood them best or were meticulous in handling matters. Such qualities allowed these caregivers to navigate conversations between patient, family members and the healthcare team. This point is exemplified by the two quotes below:

P3: I wanted to avoid such thing (referring to family disputes). That’s why I get involved. I inform my brothers. Nowadays with WhatsApp, you know nephew, nieces, everybody in the group chat. So inform everybody just put the message there so whoever have anything to say will say, but normally my brothers are the first ones to say.

P3: …it is not that I am guided by her (referring to patient) vision…I roughly know she don’t want to suffer. I will listen to the professionals…So I will just make it more clear to her as simple as possible for her to understand. And then I’ll tell her this is what we're going to do. Are you okay with it? Like getting her to agree with what the professional is saying.

(P3_Male_50–59 years old _Indian)

Lead facilitators often put in effort to elicit, reconfirm and interpret patient’s EOL values and preferences. An excerpt from P11’s interview below demonstrates this point:

P11: My mum is one person that she always said, when it’s time to go, go. No need to prolong. And she also told us she would want to be sea burial…This was even prior to being sick. She had expressed this when she was sick with a red and green card… without fail, she goes to the green card because I say mum you told me you want to scatter your ash in the sea after cremation. Is that still there? She pointed to the green card.

(P11_Female_60–69 years old _Chinese)

Some data also suggested the lead facilitator sometimes needed to ‘take over’ and do what they thought was ‘best’ or to fill a gap in the decision-making in cases where they judged the patient to be lacking in capacity or disinterested (eg, P3 and P8). However, this was not the norm and most of the data implies that caregivers preferred to build consensus even as they led the decision-making.

P3: Not just to avoid that (referring to keeping all family members informed to avoid family disputes), but I always feel I don’t know. I’m still I think maybe I’m still Asian la, even though I study more, but I think for me as an Asian Asian value still is there because it must be done as a family because the parents are only one you see.

(P3_Male_50–59 years old _Indian)