Developing Country-specific Questions about End-of-life Care for Nursing Home Residents with Advanced Dementia using the Nominal Group Technique with Family Caregivers

artificial nutrition and fluids Abstract Objective: We aimed to develop question prompt lists (QPLs) for family caregivers of nursing home residents with advanced dementia in the context of a study involving Canada, the Czech Republic, Italy, the Netherlands, the United Kingdom and Ireland, and to explore cross-national differences. QPLs can encourage family caregivers to ask questions about their relative’s end-of-life care. Methods: We used nominal group methods to create country-specific QPLs. Family caregivers read an information booklet about end-of-life care for people with dementia, and generated questions to ask healthcare professionals. They also selected questions from a shortlist. We analysed and compared the QPLs using content analysis. Results: Four to 20 family caregivers per country were involved. QPLs ranged from 15-24 questions. A quarter (24%) of the questions appeared in more than one country’s QPL. One question was included in all QPLs: “ Can you tell me more about palliative care in dementia?” . Conclusion: Family caregivers have many questions about dementia palliative care, but the local context may influence which questions specifically. Local end-user input is thus important to customize QPLs. Practice implications: Prompts for family caregivers should attend to the unique information preferences among different countries. Further research is needed to


Introduction
Healthcare professionals are increasingly adopting patient-centered care that is sensitive to the patient's preferences and needs [1]. This requires patients to be engaged in the decision-making process, to be informed about the positive and negative features of each option and to be aware of their own values and preferences that can guide the decision [2].
Patient engagement in the context of advanced dementia can be challenging. There are numerous quality and ethical issues, such as the cognitive ability of the person with dementia to understand and contribute to the conversation [3,4]. Consequently, family caregivers -for example partners or adult children-are expected to represent their relatives with dementia when they are no longer able to express own wishes [5]. However, family caregivers may lack a good understanding of dementia and end-of-life care [6,7]. To address this issue, Arcand and colleagues developed an informational booklet about comfort care for nursing home residents with advanced dementia: the Comfort Care Booklet [8]. The aim of the booklet is to prepare family caregivers for the end of life and reassure them about the patient's comfort. The Comfort Care Booklet has been translated and adapted for use in various countries [9][10][11][12][13]. The Comfort Care Booklets were developed by healthcare professionals with input from patients and family caregivers. Involving end-users can lead to more user-friendly information [14].
Although family caregivers welcome information about dementia and end-of-life care, information provision itself is not enough: there is a need for follow-up conversations with healthcare professionals [15]. As families will often not explicitly share their information and support needs with healthcare providers, they need assistance in expressing these needs so that they can be met [15]. An aid to increase patient or family caregiver engagement is a question prompt list (QPL) [16]. A QPL is a list of typical questions, that can encourage individuals to ask questions about their July 2021 4 relative's care. Thus, healthcare professionals can provide personally relevant information. Hyatt, Lipson-Smith and colleagues [17] found that patients with cancer experienced a QPL to be supportive to their care, but also recommended that QPLs should be culturally-tailored and patientdriven. For example, the development of a QPL for patients with cancer in the US and Australia led to two different versions as some aspects were deemed less acceptable or useful in the US compared with Australia [18].
Currently, available QPLs about illness and treatment to support decision making (e.g. [19][20][21][22]) are mostly tailored for patients with cancer. A QPL to support family caregivers of nursing home residents with advanced dementia is not internationally available. Therefore, the objective of this study was to develop a QPL for family caregivers of nursing home residents with advanced dementia.
The purpose of the new QPLs was to complement the Comfort Care Booklet [8][9][10][11][12][13] by encouraging question asking and aid conversations about care.
In this study, we consult current and bereaved family caregivers in five countries to ensure family-driven and culturally-tailored QPLs for dementia. We assessed differences between countries in the resulting QPLs to examine importance of the local context.

Methods
This study is part of an international EU Joint Programme -Neurodegenerative Disease Research (JPND) project to support family caregivers of nursing home residents with advanced dementia in decision making, called 'mySupport study', involving Canada, the Czech Republic, Italy, the Netherlands, the UK and Ireland. The QPLs were developed with the targeted consultation of family caregivers. The aim of the development process was to agree on the final QPLs with 20-25 questions.
This process took 13 months and was divided into three phases as described below and depicted in Figure 1. An evaluation of the QPLs in practice was outside the scope of the current study and will be addressed during the larger mySupport study [23].

Phase one
Phase one (July 2019 -September 2019) involved an elderly care physician, two researchers specialized in palliative care, a researcher trained in psychology and a researcher educated in neuroscience. This team selected a set of questions from three available QPLs (two about dementia care and one about palliative care). The QPLs originated in three different countries (Australia [24], Canada [25], the Netherlands [26]), providing a transnational overview of questions (248 questions in total, 206 without duplicates).
The team aimed to reduce the number of questions to < 50 and to include at least one question per topic discussed in the Comfort Care Booklets: dementia, end-of-life decision making, symptom relief, and end of life. Questions were deemed appropriate for inclusion if they were relevant to the nursing home setting and within the scope of the Comfort Care Booklets. 'Nursing home' is used to refer to a collective institutional setting in which care is provided to older adults 24 hours a day, including nurses and medical staff. Question selection took place in three consensus rounds and was informed by the inclusion criteria in Box 1. In the Netherlands, three researchers (LB, WPA, JTvdS) independently selected questions and discussed to reach consensus. In parallel, two researchers from the UK followed the same procedure (SM, KB). Next, the results of these two independent consensus rounds were discussed between the researchers from the two countries in a third consensus round. A preliminary selection of 39 questions was thus completed in preparation for Phase two.

Phase two
Phase two (November 2019 -July 2020) involved the targeted consultation of end-users of the QPLs.
End-users in each country independently compiled the final selection of questions for the QPL to address local needs. End-users were defined as current or bereaved family caregivers of people with dementia. The eligibility requirements were: over 18 years of age; sufficient capacity of the local July 2021 6 language to read the Comfort Care Booklet and participate in the discussion; and able to agree to terms and conditions of participation. The consultation protocol was reviewed by local ethics committees in Canada (Hamilton Integrated Research Ethics Board: 2019-5837-GRA) and the Netherlands (Medical Ethical Committee Leiden-Den Haag-Delft: N19.114) and written informed consent was obtained from all participants. In the Czech Republic, Ireland and the UK, the consultation was considered 'Public and Patient Involvement' and therefore the process was exempt from review by ethics committees and formal consent procedures. In Italy, the consultation could not be performed due to COVID-19.
Convenience recruitment strategies included local invitations and snowballing. In Canada, family caregivers involved at the nursing home that participates in the larger project were invited by the site's social service coordinator. The Czech team recruited participants using the Facebook page of the Center for Palliative Care and by using the snow-ball method, that is: participants invited others from their social networks to participate. In collaboration with Alzheimer Nederland, the Dutch team invited family caregivers who are part of the Alzheimer panel from a west and middle region over email. The team in Ireland posted flyers in public spaces and on Facebook, and invited potential participants via personal and professional networks. In the UK, researchers recruited family caregivers through their local patient and public involvement connections around Leicester and through the network of a family caregiver that was involved in previous projects in Lancaster.
The consultation process was structured in the four steps (see below) of a nominal group technique [27,28]. Although normally conducted in a group setting, the first two steps were adapted to allow family caregivers to complete them individually at home. This was done to ensure that participants were sufficiently prepared for the group discussion. The group discussion was virtually conducted in some cases using email, phone calls or videoconferencing (see Supplementary information Text S1). The steps were conducted as follows:

Silent-generation step (individually at home)
Upon reading the Comfort Care Booklet, family caregivers wrote down any questions that came to mind and that they would wish to discuss with a healthcare professional. They also wrote down their reasoning for posing these questions.

Round-robin step (individually at home)
Afterwards, the family caregivers selected approximately 15 questions of the 39 pre-selected questions that they felt may arise when a family member reviews the Comfort Care Booklet. Finally, the family caregivers could add any outstanding questions that were not yet addressed; once again, they wrote down their rationale for posing these questions.

Discussion/item clarification step
A moderator presented all individually generated and selected questions to the family caregivers at the group discussion. The moderator was not involved in the preliminary question selection to minimize bias in the facilitation of the discussion. Family caregivers read all the questions as a group and discussed each item's similarities, differences, and reasons for its inclusion in the final list.
Observers took notes on reasons for inclusion and exclusion of questions.

Voting step
After the group discussion, participants voted on which items to include in the QPL by raising hands or highlighting them on a collective list. The 20-25 questions that received the most votes were included in the final list. The research team mapped the questions from the final list onto the relevant sections of the Comfort Care Booklet to ensure that there was at least one question per topic.

Phase three
Phase three (June 2020 -August 2020) involved two researchers (MN, Canadian and LB, Dutch), who compared the final QPLs from each country to highlight similarities and differences across contexts using conventional content analysis [29]. First, LB and MN familiarized themselves with all of the questions. They then inductively derived a set of codes from the questions and labelled each July 2021 question with a code. Next, codes were sorted into themes when referring to a similar overarching topic. Finally, code names were adjusted after defining the themes, and themes were refined after adjusting the codes. Reasons for including questions that were available in field notes or participants' answer sheets were collected and also underwent a content analysis to explore any cross-national differences in the rationale for including questions in the QPLs.
To ensure validity and rigor [30], all steps of the analyses were independently performed by two individual coders (LB, MN). Codes, themes and interpretations were discussed at each step to reach consensus.

Phase one
Out of the 206 questions, the UK team selected 33 and the Dutch team 49. Ten questions were selected by both teams, totaling 72 questions. The two teams agreed on the in-and exclusion of 75% (186/248) of the questions. Cross-national differences were apparent during the consensus discussion. For example, questions about life termination and prognosis seemed inappropriate to the UK team because they were difficult to answer adequately, and end of life and euthanasia were not considered topics that people tend to discuss. They were pressing to the Dutch team because people will often ask about these issues and there was a concern of creating taboo when excluding such questions. Table 1 shows the 39 questions that were pre-selected by the research teams upon reaching consensus. Table 2 shows an overview of the nominal group for each country. Forty-three people participated, ranging from 4 to 20 per country. The ages ranged from 25 to 87 years, and the majority (72%) was female. The participants were current or bereaved family caregivers (n = 41) or living with dementia (n = 2). Six participants had professional experience with dementia. The group discussions took 40 to  (Table S1). The group discussion resulted in a final list of questions in each of the countries. Canada decided upon 15, the Czech Republic had 20, Ireland had 22, and the Netherlands and the UK each had 24 (Table 1 and Supplementary information Table S2).

Phase three
Questions were scanned for overlap and similarity, leading to a list of 76 distinct questions from the total of 105 selected questions. Almost a quarter (24%, 18/76 questions) were selected by more than one country. Question 2 from the pre-selected list of 39 questions was selected by all groups: "Can you tell me more about palliative care in dementia?" For 55/105 questions (52%) the reason for inclusion was clearly described in the field notes or participants' answer sheets. The Czech Republic (70%), Canada (67%) and the Netherlands (63%) had more information about the rationale available than the UK (37%) and Ireland (32%). The reasons that were provided for selecting questions were aggregated into the following six themes, in order of frequency: just obtaining information, preparation for end of life, reassurance, preparation for shared decision making, informing staff about the resident's needs and informing staff about the family caregiver's needs ( Table 3). The reason provided most often per country was just obtaining information in the UK, preparation for end of life in Ireland, and just obtaining information and preparation for shared decision making in the Czech Republic. In the Netherlands, informing staff about the resident's needs was directly followed by reassurance and just obtaining information.
Canada had mentioned reassurance, just obtaining information and preparation for end of life at equal frequency.
The inductive content analysis of the QPLs resulted in 18 codes (Table 4). The most common codes were communication with staff, care protocols, palliative care information, and roles and responsibilities. The codes were aggregated into seven broader themes. These include Request for  The cross-country differences may reflect socio-cultural differences and how well palliative care is established locally. The focus of the Dutch QPL on palliative care may portray that palliative care for people with dementia may be less well-known (and had not been part of Dutch dementia plans [31]). The Dutch participants indeed often provided 'just obtaining information' as reason for selecting questions, but reassurance and informing staff about the resident's needs were also frequently mentioned. The Irish QPL lacked questions about controlling death circumstances. This suggests that this topic felt less appropriate to the Irish family caregivers. However, the Irish family caregivers most often provided 'preparation for end of life' as a reason to include questions. The large number of questions about communication with staff in the UK QPL could indicate that patient engagement in healthcare is well established in the UK [32]. As the reason most often provided for including questions was 'just obtaining information', this could also suggest that family caregivers prefer being informed by staff members in personal conversations rather than searching for information themselves. The Canadian QPL focused on questions about specialist services and care protocols. One interpretation could be that the family caregivers had experienced that care coordination and accessibility to services can be improved [33]. The reasons for selecting questions varied.
The Czech QPL seemed to be the most different from the other countries, indicating concern about the quality of palliative care in dementia and questions regarding nutrition and hydration. The main reasons for question selection among Czech family caregivers included just obtaining information and preparation for shared decision making. Cross-national work indicated that artificial feeding and fluids is a sensitive topic [34] and recommendations on nutrition and hydration from the European Association for Palliative Care received only moderate consensus [35]. The concerns surrounding the quality of palliative care in dementia might relate to the poor resources for palliative care in the Czech Republic, where palliative care is not well known [36] and not yet officially acknowledged as necessary for people with dementia [37]. This could also explain the focus on curative treatment in the questions, rather than palliative options.
It is important to note some limitations of this study. Differences in how and when the group discussions were conducted, due to COVID-19 or local practice, may have influenced the results.
Adapting the group discussion from an onsite activity to a thread of emails [38], phone calls [39] or videoconferencing [40] could have impacted the engagement process. Furthermore, group sizes differed between the countries. The resulting QPLs are based on input from a convenience sample of family caregivers and may not be representative of general cross-country differences. Possibly, (larger) cultural differences within groups rather than between countries have affected the results.
The consultation process took place prior to the pandemic in Canada and the Netherlands, but during the pandemic in the Czech Republic, Ireland and the UK. This could have affected the type of questions that were pertinent to family caregivers. Also, we did not have clear information on the July 2021 rationale for selecting questions for almost half of the questions. Finally, the use and acceptability of the QPLs in real-time practice settings have not yet been evaluated.
However, we can explore the quality of the QPLs. According to the International Patient Decision Aids Standards (IPDAS) Collaboration [41], decision aids should contain the six key elements of shared decision making: (1) situation diagnosis, (2) choice awareness, (3) option clarification, (4) harms and benefits discussion, (5) patient preferences deliberation and (6) making the decision [42].
Interestingly, the themes and codes derived from the content analysis overlap strongly with these six elements of shared decision making, with two exceptions: questions about care staff and care setting are not included among the six elements. Care staff and care setting form the context of shared decision making, which is thus not well represented by the six elements. The importance of the relational context of shared decision making has been proposed before [43], but the care setting has not been included in shared decision making definitions yet [44]. Our results mirror the findings of Thompson and colleagues [15] that family caregivers need general information about nursing home life in addition to specific information about treatment and disease, to support them in decision making.

Conclusion
Given divergent preferences for sample questions, engaging end-users such as family caregivers of persons with advanced dementia in the development of materials is vital, especially when these materials aim to increase family caregiver engagement. Cultural differences may influence the information needs of family caregivers and should therefore be considered. The family caregivers in this study expressed a general need to be informed about palliative care in dementia, and about the process of shared decision making. The differences in questions generated between the countries underpin the value of cross-country exercises when developing materials for implementation into practice.

Practice implications
Person-centered care calls for patient and family engagement, that requires facilitation via communication tools such as QPLs. QPLs can be used by families to prepare for advance care planning conversations with healthcare professionals by reflecting on the questions, or provide possible topics to help start conversations about care and reassure families that their information needs will be met. Healthcare professionals need to be aware of the different informational needs of their patients and their families, possibly related to their cultural background. Service planning and training programs for healthcare professionals should have more attention for shared decision making and general information provision about dementia palliative care, as these subjects raised many questions across the countries. This study also highlighted areas where further guidance is needed, such as artificial hydration and nutrition in advanced dementia in the Czech Republic. More research is necessary to explore the use of QPLs in conversations in healthcare and possible crosscountry differences in these conversations.

Informed consent
Informed consent was obtained according to country-specific guidelines. I confirm all patient/personal identifiers have been removed or disguised so the patient/person(s) described are not identifiable and cannot be identified through the details of the story.  3. The question is of cultural, care practice or legislative relevance to at least one of the participating countries.
4. The question is not purely medical or medical-technical, and can thus be addressed by a long-term care staff member from the relevant occupation (for example, by a nurse or a social worker).
5. The question cannot be answered or on the contrary, is already covered by the Comfort Care Booklet, but is still pressing to family caregivers (according to literature and researchers' experience in practice) and bringing it to the table would be beneficial.    Outline of the three-phase process.pptx