2.1 The Setting and Participants

The ACP information sessions of 1.5 h were divided into two parts: the first, facilitated by a social worker, focused on exploring values and preferences. The second part, led by a geriatrician, covered end-of-life care, ACP, AMD, and LST. For ease of reference, a table highlighting the definition of terms can be found in Table 1 [8].

Initially, the talk sessions were held every 4–6 weeks in a hybrid format (both face-to-face and online) at a nearby university-affiliated facility during COVID-19 restrictions. Later, they transitioned to in-person sessions of 5–10 participants at the hospital. For those interested in further ACP discussions, follow-up meetings were arranged to complete the ACP process, including signing the ACP/AMD forms according to Hong Kong Hospital Authority guidelines [8]. The content of the ACP discussions included EOL care, LST, and AMD.

The inclusion criteria required participants to be family caregivers of PWD who attended the ACP talks. Exclusion criteria included individuals under 18, those with cognitive impairments or mental incapacity, and those unable to speak or comprehend Cantonese.

2.2 Interviews

Between March 2021 and April 2024, 28 talk sessions were held for 164 family caregivers of 117 patients. Follow-up calls were made reaching 124 family caregivers and successfully conducting 23 individual semi-structured interviews. At the time of the interviews, five families had completed the ACP documentation process and were all interviewed.

Interviews of family caregivers who did not proceed with the formal ACP documentation post-talk were conducted by three trained interviewers, resulting in 18 interviews. One researcher conducted interviews with the five families who completed the ACP process. These interviews took place between November 2023 and February 2024.

A total of 124 family caregivers provided feedback on the effectiveness and benefits of the ACP talks in aiding their understanding of ACP. A summary table of their feedback is available in the Supporting Information S1.

Family caregivers who had completed the ACP process further explored their experiences of the ACP process. Based on literature [9, 10] and clinical observations, a semi-structured interview guide was developed to meet the study's objectives.

2.3 Data Analysis

The telephone interviews were audio-recorded and subsequently transcribed into text. Assisted by NVivo data analysis software, these transcripts were then coded using a thematic content analysis framework [6, 7]. This method is beneficial given the novelty of ACP in Hong Kong, especially concerning PWAD. The research team met regularly to review the transcriptions and discuss the findings. Identified concepts were categorized into themes, which then reached a consensus. Consistent thematic schemes were applied across all transcripts. The emergence of new themes peaked around the 10th interview and gradually declined with further interviews. Eventually, in the last three interviews, it reached data saturation where no new codes, sub-themes, or themes emerged [11, 12].

2.4 Ethical Approval

The study received approval from the Hospital Authority New Territories East Cluster Clinical Research Ethics Review (CREC Ref. No. 2023.331) before the recruitment of participants. Participation was voluntary, with digital informed consent obtained immediately before telephone interviews.

3.1 Theme 1—Receptive to ACP

Family caregivers had not heard of ACP before. They were open to ACP when initiated by the healthcare team, particularly if patients had previously expressed a desire to forgo treatments, experienced prior crisis episodes, or had multiple hospital admissions. They believed healthcare professionals (HCPs) knew the best time to initiate such a conversation. In their recollection, there was more emphasis on forgoing CPR or DNACPR. Many accepted ACP due to patient's advanced age and ill health status. They believed that LST was meaningless, which prolonged unnecessary suffering. Some deemed it a waste of resources. Family caregivers who had a similar experience before, or had relatives who died in a palliative care setting, found it easier to discuss ACP. They had a good understanding of patients' health status, and the care demands due to their caregiving journey.

3.2 Theme 2—Mixed Emotions

Considering forgoing LST, most family caregivers were emotionally calm because they accepted death and acknowledged life's limitations. For some caregivers, it was a dilemma between not wanting patients to die or to suffer. After some mental processing, the decision to forgo LST was seen as alleviating patients' suffering, which provided family caregivers with a sense of relief.

Some found it difficult to accept letting go of treatments, which brought sadness. A family caregiver described that she could not accept forgoing LST for her relative initially. She was upset that HCPs were not trying to “save” the patient's life. In the subsequent readmission, the topic was brought up again by another nurse, and she was then able to accept it.

3.3 Theme 3— Barriers to ACP

Lacking awareness of the potential benefits of ACP is a major barrier in Hong Kong for PWD. Before the introduction of ACP by the healthcare team, nearly all family caregivers had not heard of ACP. Some welcomed a more open approach to let the families know about ACP in advance. They thought that the public should know about it. They believed that many families and the public had misunderstandings about LST. The ACP talk was useful for them to learn more about LST. They felt that there was inadequate public education and that only people with personal experience were aware of the possible benefits of ACP for PWD.

Another barrier was that ACP discussions came too late, and patients could not get involved. Many family caregivers believed that it was necessary to involve the patient, the earlier the better. However, they found it difficult to start the conversation due to a lack of skills to discuss with the patient and other family members. They also had many worries, such as upsetting the patient and causing emotional distress. Some caregivers believed the local Chinese culture was not permissive in free discussions of death and dying. Some expressed that they chose avoidance because they could not accept talking about death or believed it was taboo for older people. A family caregiver tried talking to the patient, but the patient refused to discuss his EOL care wishes.

At the beginning of this study, Hong Kong was still under a strict social isolation policy during the COVID era. Family caregivers found discussing ACP inconvenient. They could not tell what was wrong with the patient's health because they were not allowed to visit them at the hospitals or nursing homes. They could not directly communicate with the patient face-to-face. The patients were at the advanced age, 85–100, and their children were also older adults. Such a period created a lot of difficulties and suffering for patients and families.

Another top barrier was that family caregivers were occupied with other challenges, such as preparing patients for hospital discharge, finding nursing homes, running errands, such as obtaining self-paid personal items, and the like. Due to the shortage of hospital beds, patients were required to go home as soon as possible. This was often cited as a competing priority over ACP discussion.

Family caregivers believed that ACP discussion was not that helpful to existing patient care. They expected more follow-up care upon ACP discussion. A family caregiver expressed that providing more physical care to the patient or education on caring for PWAD was a modest anticipation post-ACP discussion.

3.4 Theme 4—Family Consensus in Decision-Making

Family consensus on DNACPR was readily achieved because family members were concerned about patient suffering. Patients were often excluded from discussions for various practical and health reasons. ACP discussion often came too late to involve patient participation. When there were no previous EOL care discussions, the patient's view may not be incorporated in ACP or EOL care. Not all family caregivers could understand the concept of substitute decision-making. They misunderstood the healthcare team seeking their perspective or the family's collective perspective.

Family caregivers assumed that patients were incompetent or unable to participate due to disabilities. They believed discussing with patients would not be fruitful. They also rationalized that patients were illiterate, and thus not knowing what they needed or that they were not willing to listen to others. Some feared that ACP discussions could cause unnecessary worries to patients. This lack of awareness of patient's values or preferences prevailed.

3.5 Theme 5—Benefits of ACP

Family caregivers found themselves more knowledgeable about ACP after the talk and discussion. They believed ACP helped meet patient's holistic needs. ACP discussion could enhance dementia care by considering aspects of care beyond the illness itself. They also found more communication among family members and more attempts to communicate with patients. In pondering ACP, family caregivers had more reflection. Although limited, they wanted more conversation with the patients so their preferences could guide them in future care decisions.

In ACP discussions, family caregivers had more psychological preparation for patients' death. Since families had more discussions, family cohesion helped them psychologically prepare and get support from each other. In the end, it was easier to accept patients' death as well as less regret for unfinished business. A few family members experiencing bereavement mentioned that ACP discussion had helped them to have less guilty feelings because they were there for them.

Many informants reported that they verbally discussed ACP in their family after the ACP talks, even though they did not formally complete ACP in written format. Caregivers also found the ACP process benefiting themselves. The process prompted them to ponder on their own ACP and their older family members'.

4.1 Low Take-Up Rate of ACP

Despite people being receptive to ACP in Hong Kong, it remains a novel concept, requiring time to gain momentum. Besides, individuals could set their priorities and choose whether to engage in ACP [14]. Many dementia patients in Hong Kong prefer family and HCPs to make decisions on their behalf, reflecting a personal choice.

Our findings align with previous studies indicating that ACP is not a priority for patients, caregivers, and HCPs due to ongoing crises and current issues, and they are not future-oriented [15]. Thus, measuring success by the number of completed ACP forms is not warranted. A better strategy may involve having information available and HCPs ACP-trained to facilitate ACP discussions whenever patients and families are ready.

Many local Chinese still treat death as taboo. Family caregivers frequently feel uneasy discussing ACP with patients or older family members, fearing it may cause unnecessary anxiety. As such, they often could not know patients' values or preferences, which undermines the core intent of ACP- to provide goal-concordant care. Therefore, timely and inclusive ACP discussions involving patients and family members are crucial for informed decision-making.

ACP is a gradual process and should be integrated into routine dementia care [14, 16, 17]. Understanding patients' and surrogates' perspectives on current health status and disease trajectory helps provide care that aligns with their preferences. When appropriate, HCPs can invite them to imagine certain circumstances where some medical treatments may bring more suffering than benefits that they would not want. HCPs can naturally introduce ACP by discussing potential medical treatments and their impacts, rather than focusing solely on death or CPR, which can deter patients and families from considering ACP [16], particularly in an abrupt manner.

Appointing a healthcare proxy, which is not part of Hong Kong's Advance Decision on the LST Bill, offers practical benefits by ensuring early discussions and reflecting patients' values and preferences. It can avoid leaving ACP to a late stage when the family must make surrogate decisions, which they may not be comfortable with. Proper documentation of these discussions is essential, rather than relying solely on the AMD form.

Legislation in Hong Kong tends to emphasize signing AMD forms without addressing the ACP process. The issue of form-filling versus documentation needs to be discerned. A comprehensive ACP process with clear documentation is more practical and effective than written documents like AMD forms alone [18], which causes hesitation to many. The AMD's focus on DNACPR overlooks other kinds of life-sustaining treatments. It assumes that basic supportive care includes hydration and nutrition. Since many PWAD cannot swallow without choking, the whole issue of artificial hydration and nutrition (with restraints, etc.) is ignored.

4.2 Cultural Practice in Decision-Making

Achieving family consensus on DNACPR stems from shared concerns about patient suffering. However, excluding patients from these discussions due to dementia is inappropriate. Many family caregivers lack an understanding of substitute decision-making, leading to misconceptions about expressing a collective family perspective.

Maintaining family harmony is a value of high importance in the Chinese context. Consequentially, ACP discussions are unlikely to be pursued proactively when they are perceived as a threat to family harmony. When ACP was prompted by HCPs, family caregivers are often unprepared to make decisions for their ill relatives.

Asians, such as Hong Kong people, are more family-centric, where family involvement in decision-making is a key feature in ACP discussions and EOL care [19]. The cultural expectation of family decisions often creates family conflict, and consensus building may not always be possible, leaving the surrogates feeling lonely without support and adding burdens [20]. Filial expectations, such as in the Chinese culture, could have a negative impact on surrogates experiencing stress and family conflict [21].

Family caregivers often face social pressure to conform with filial expectations and avoid being considered as lacking it. When the decision-makers view filial piety as pursuing all possible LST, forgoing LST will not be an option. Further, filial expectations and family dynamics could contribute to family conflict relating to birth order in EOL decision-making [21]. In Hong Kong, it is a common scenario where elder sons are assumed with higher weighting in family decision-making.

Older Chinese who had strong trust in physicians tended to place less emphasis on EOL discussions with family. And they valued counseling from care providers for EOL discussions with family. Patient education and family engagement in EOL discussions led by trusted healthcare providers may be effective approaches to ensure goal-concordant care [22]. Chinese culture was seen as collectivist, prioritizing family over individual rights, or the individuals must surrender their rights for the betterment or harmony of the whole family. Recent research, however, indicates a shift towards individualism [23-25]. Due to the unique history of Hong Kong, the concept of collectivism may only be one of the facets among contemporary local Chinese. As evident from this study, individual family caregivers hold varying views on ACP, suggesting patient preferences should be solicited early to provide person-centered dementia care. Evidence shows that ACP positively impacts the quality of end-of-life care. ACP is associated with improved outcomes for PWD and their caregivers [26]. Moreover, families could have higher satisfaction with care, and they were more likely to have decided on medical options over time [27]. Increased education and support for families around issues of end-of-life care decisions for PWAD are needed.

4.3 ACP and Quality of Care From Family Caregivers' Perspective

The ACP discussion can raise family caregivers' awareness of the impending final phase of dementia, which is often complicated by their intensified anticipatory grief and high care expectations. Therefore, focusing on current care is essential before advancing ACP discussion. Experts agree that ACP for PWD should cover available care options and their lifelong decision-making principles, as well as allowing time for them to come to terms with their diagnosis and prognosis [14].

A retrospective, cross-sectional analysis of data from nursing homes found no significant differences in family ratings of care between those with and without documentation of LST preferences [28]. This suggests that end-of-life care quality relies more on current care practices than solely on LST documentation or AMD.

A service evaluation revealed that about 80% of residents of care homes chose ACP, with most preferring to remain in the care home [29]. While ACP acceptance increased the likelihood of residents dying in their care homes, it did not reduce hospital admissions. Obstacles to effective ACP implementation include organizational policies, care home staff beliefs about ACP, and manpower shortages. Significant challenges remain to fully realize ACP in actual healthcare settings.