INTRODUCTION AND BACKGROUND
In general, the goal in health care is to recommend treatments based upon research (i.e., the care should be evidence-based). The same applies for palliative care practice and medicine. Historically, the accumulation of evidence for palliative care has been largely based on clinical practice and experience of physicians/medical staff (Kaasa & De Conno, Reference Kaasa and De Conno2001; Kaasa et al., Reference Kaasa, Hjermstad and Loge2006). Most studies conducted in palliative care have been small and descriptive, often without the necessary quality to inform evidence based medicine (Kaasa & De Conno, Reference Kaasa and De Conno2001; Kaasa et al., Reference Kaasa, Hjermstad and Loge2006; Kaasa & Radbruch, Reference Kaasa and Radbruch2008; Bennett et al., Reference Bennett, Davies and Higginson2010). It has also been argued that palliative care research is not very well developed and that there is no consensus on how to measure the most important outcomes within palliative care study populations (Kaasa et al., Reference Kaasa, Hjermstad and Loge2006; Bennett, et al., Reference Bennett, Davies and Higginson2010).
Randomized trials are the gold standard method for evaluating treatments and services in health care, however, they are often difficult to complete in palliative care, and encounter poor recruitment (Higginson & Booth, Reference Higginson and Booth2011) due to perceived vulnerability, moral appropriateness, consent, gate-keeping and inclusion, and the research culture (Duke & Bennett, Reference Duke and Bennett2010). The quality of research in this field is dependent on complex ethical decisions in designing, conducting and reporting this research while addressing concerns for the protection of the safety and rights of people facing life-threatening illness (Duke & Bennett, Reference Duke and Bennett2010; Whiting & Vickers, Reference Whiting and Vicker2010).
It is also claimed that the palliative patient population is too ill and too vulnerable to allow meaningful scientific research and that the population is too heterogeneous for clinical research on a group level (Kaasa & De Conno, Reference Kaasa and De Conno2001). In the case of a dying person, who may not be alive at the end of the research, the judgments made by the researcher may be unchallenged and patients may be unaware of the extent to which the material they have consented to will be used (de Raeve, Reference de Raeve1994). Furthermore, there have been reports of adverse effects from what was considered a “low impact” survey of people with life-limiting illness where the content caused significant distress and anxiety among participants (Braithwaite et al., Reference Braithwaite, Philip, Finlayson, Tranberg, Gold, Kotsimbos and Wilson2009).
Despite the concerns and arguments discussed above, surveys of people with a life-limiting illness who have been asked about participation in clinical trials have demonstrated that they are as likely or more likely than the general patient population to want to be involved in clinical trials (White et al., Reference White, Gilshenan and Hardy2008; White & Hardy, Reference White and Hardy2010; Currow et al., Reference Currow, Agar and Abernethy2011). Palliative patients, even those very near death, have been found to be keen to take part in research, both qualitative and quantitative (Ross & Cornbleet, Reference Ross and Cornbleet2003; Henderson et al., Reference Henderson, Addington-Hall and Hotopf2005; Pautex et al., Reference Pautex, Herrmann and Zulian2005; Terry et al., Reference Terry, Olson, Ravenscroft, Wilss and Boulton-Lewis2006). Terry et al. (Reference Terry, Olson, Ravenscroft, Wilss and Boulton-Lewis2006) found the shared reasons patients wanted to participate in research were: altruism; enhancement of a sense of personal value; the assertion of persisting autonomy; and the value they placed on a commitment by clinicians to optimizing care through research. These reasons are consistent with the views of Berry (Reference Berry2004) when discussing guidelines for ethical research and vulnerable patients. Berry argues that to label palliative care patients as a “vulnerable” population is paternalistic, discriminatory, unjust, and a potential barrier to carrying out important research that can improve the quality of care for dying patients and their families.
PRIORITISING RESEARCH IN PALLIATIVE CARE
Attempts to prioritize palliative care research are not new. Bennett et al. (Reference Bennett, Davies and Higginson2010) reviewed challenges in the delivery of research in end-of-life care in the United Kingdom (UK) and internationally. The review drew on relevant policy and research examples from countries similar to the UK to define some of the issues. They concluded that there were problems of terminology in the field of palliative care and that there was a lack of emphasis on clinical studies relating to the medical aspects of palliative care. Future priorities identified by the authors included the need to build research capacity, make better use of existing research study outcomes, and develop sustained programs of research. They recommended that future research needs to focus on services and public health issues such as understanding care models, evaluating services, surveying social attitudes, and changing cultures (Bennett et al., Reference Bennett, Davies and Higginson2010). However, Bennett et al. (Reference Bennett, Davies and Higginson2010) point out that this is problematic in that such research is often conducted after policy initiatives have been put in place, making it difficult to understand the effectiveness and cost effectiveness of new models of care, or for innovation to lead to sustained changes in practice.
In a discussion paper on palliative care research, priorities and the way forward, Kaasa and Radbruch (Reference Kaasa and Radbruch2008) point out that the study population are described and defined differently across studies and there is no consensus on how to measure the most important outcomes, such as symptoms such as pain, fatigue and depression in palliative care research. These authors proposed that palliative care research should be conducted through the creation of permanent research posts and that these need to have sufficient and sustained funding and be collaborative with other centers of research. They also recommend that as the majority of palliative care patients come from oncology, it is important to establish a close collaboration between these two disciplines.
In summary, there has been considerable clinical research undertaken to improve care for palliative patients and bibliographic databases hold sizeable repositories of palliative care research that pose challenges for professionals in managing the complex evidence-base for this diverse field (Tieman et al., Reference Tieman, Sladek and Currow2009). The aim of this Delphi study was to: inform organizational decision-making and policy development regarding future research priorities; and inform the wider community of the research intentions for a hospice in New Zealand.
STUDY DESIGN
A modified three-round Delphi technique was used following the methodology described by Malcolm et al. (Reference Malcolm, Forbat, Knighting and Kearney2009). The Delphi technique has been used to identify research priorities in a number of fields, such as the care of infants, children and adolescents (Rudolph et al., Reference Rudolph, Hiscock, Price, Efron, Sewell, South and Wake2009; Wilson et al., Reference Wilson, Ramelet and Zuiderduyn2010), midwifery (Fenwick et al. Reference Fenwick, Butt, Downie, Monterosso and Wood2006), and palliative care (Malcolm et al., Reference Malcolm, Forbat, Knighting and Kearney2008; Reference Malcolm, Forbat, Knighting and Kearney2009; Steele et al., Reference Steele, Bosma, Johnston, Cadell, Davies, Siden and Straatman2008; Sigurdardottir et al., Reference Sigurdardottir, Haugen, van der Rijt, Sjøgren, Harding, Higginson and Kaasa2010). Delphi has also been used in a variety of areas, including government, medical, environmental and social studies, as well as business and industrial research.
The Delphi technique is a structured group communication method that uses a series of questionnaires or “rounds” to gather information or expert opinion from stakeholders about complex problems or novel ideas (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011). Rounds are held until group consensus is reached. Using this technique allows a large number of individuals across diverse locations and areas of expertise to be consulted simultaneously and anonymously, avoiding domination of the consensus process by one or a few experts (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011). In this research a modified Delphi procedure, with focus group interviews and meetings and questionnaire Delphi rounds using a nominal group process, was used (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011). Analysis was integrative to the data collection process. Ethical approval to conduct the study was gained from the Central Regional Ethics Committee.
SAMPLE
A purposive sample of staff (n = 10), volunteers (n = 10), patients and family members (n = 6) were invited from across the hospice and participated in Round 1 of the study. Recruitment of patients and family members was problematic due to the advanced disease of potential recruits and sensitivity to bereavement issues of family members. Consequently, the number of patient and family member participants was smaller than anticipated. This type of difficulty in recruitment has been demonstrated in other studies involving palliative patients and family members (e.g., Higginson & Booth, Reference Higginson and Booth2011). Community professionals were invited to the focus group interviews but none were able to attend. However, community linked professionals (n = 3) provided written lists of areas which they considered a priority for palliative care research.
DATA COLLECTION AND ANALYSIS PROCESSES
Focus Groups: Round 1
Round 1 of the Delphi consisted of focus group interviews. These were conducted in three separate groups: hospice staff (doctors, nurses, therapists, social workers); volunteers; and patients and family members together. Participants were provided with an information sheet detailing the study, and written informed consent was obtained from all participants. Questions were developed from the literature and used as an interview guide. In addition, free discussion was encouraged within the groups. These discussions and participant interactions were captured using audio-recording equipment, which were transcribed by a professional transcriber.
A process of repeated reading of the transcripts was carried out by three of the researchers. Themes identifying areas of research interest were classified under a coding system (Table 1).
Table 1. Areas of research interest
Focus Groups: Round 2
A large number (128 items) of research interests were identified and included in the Round 2 questionnaire. All respondents were asked to rate the relevance of each topic on a five-point Likert-type scale, with “1” indicating least important and “5” indicating most important. Round 2 included some individuals who were not in the Round 1 focus groups. Respondents were also invited to provide further comments on research topics. Each group (staff, volunteers, and patients and family members) received the same questionnaire. A unique identifier was used to establish responses by group as literature has clearly demonstrated that health related priorities are very different for professionals than for patients and families. We were uncertain as to whether there would be differences between professionals and volunteers, and sought to explore this in our analysis. Thirty questionnaires were given out to each group (n = 90) questionnaires in all. The response rate was: staff (n = 18, response rate 60%), volunteers (n = 12, response rate 40%), and patients and family members (n = 8, response rate 27%), with a total response rate of 42%.
All 38 Delphi responses could be used; however some respondents missed one or more questions. Responses for each question were entered into the Statistical Package of Social Sciences Version 19 and analyzed using descriptive statistics. For each question, the proportion of scores of four or more were calculated, following which all questions were ranked to identify the 48 most preferred topics. The ranking demonstrated that staff and volunteers prioritized similar topics. Consequently, ranking of topics by these two groups were collated for the Round 3 questionnaire.
Focus Groups: Round 3
The questionnaire developed for Round 3 was administered to staff and volunteers only. Patients and family members were not included in this round due to the limited response from Round 2 and to difficulty in accessing potential participants. Questionnaires (n = 30) were sent out. These were again separated and identifiers used for each of the groups, staff (n = 15) and volunteers (n = 15). The questionnaire contained the same topics for both groups but this time a seven-point Likert-type scale was used, with “1” indicating least important and “7” indicating most important. This broader scale was used to minimize the known propensity of Likert scale responses to center around the midpoint.
The response rate was: staff (n = 9, response rate 30%), and volunteers (n = 11, response rate 37%), overall response rate of 33%. The analysis followed a similar process to that in Round 2, with the exception that the focus was on scores of six or more. The final priority list was amalgamated for staff and volunteers.
FINDINGS
Research Priorities for Patients and Family Members
Priorities for patients and family members were markedly different to those of staff and volunteers. These were categorized into four topic areas: decision-making; bereavement and loss; symptom management; and recognition of need and response of service (Table 2). Within the decision-making category the priorities identified was to examine informed consent and decision-making processes in relation to care practices and to explore how patients were supported by the team in terms of personal preferences and planning for the future. Priorities for bereavement and loss focused on how patients experienced the impact of not being able to participate in activities and hobbies they had previously enjoyed. Research priorities in regard to symptom management were on examining side effects of medication when caring for dying patients at home.
Table 2. Topics identified by patients and family member/caregivers
For patients and family members, the most significant list of priorities was related to recognition of need and response of services. The list included: identify how responsive the service is to patients and family members/caregivers; examine what “dying well” means from the family and professional perspective; explore ways to increase the level of “continuity in the care” of patients and decrease the number of contacts with multiple staff; examine how well family members/caregivers are recognized and supported in their care for patients; explore the importance of maintaining independence and self-management from a patient's view; and identify things that patients look forward to and the goals or things they want to achieve. This list speaks to the emphasis that the family member/caregivers gave to maintaining their autonomy throughout the terminal illness trajectory.
Research Priorities for Staff and Volunteers
Research priorities for staff and volunteers were very similar and were considered together. These were categorized into six topic areas: symptom management; aged care; education; community; patient and family; and bereavement support and young people (Table 3). Symptom management was highly prioritized and focused on: development of assessment tools for anxiety associated with breathlessness and sleeplessness (as a symptom cluster); conducting systematic reviews on interventions, care and best practice for each of the main symptom issues in palliative care; find out from patients what level of information they receive about symptoms and ways to manage them, indication of priority to be responsive to family member/caregiver needs; and examine medicalization/pharmacology versus therapeutic interventions in management of symptoms in hospice practice.
Table 3. Topics identified by staff and volunteers
The topic of aged care focused on evaluation of the role of the hospice in providing education and clinical support to Aged Care facilities and exploring the preparedness of such facilities for the increase in palliative patients due to an ageing population, particularly in relation to people with dementia. This education theme continued with staff and volunteers prioritizing research into the experiences of staff providing family member/caregiver counselling and examining the education needs of those staff members, and to explore factors for retaining and building the volunteer workforce for the hospice.
Community topics were firmly focused on general practice activities. Priorities were to investigate what general practices and primary health nurses expect from a specialist palliative care program and what the support and education needs of these professionals are. Also included in the priorities of staff and volunteers was the examination of effective partnering with other providers and specialists in the care of palliative patients. Staff and volunteers did include the need to prioritize research on support needs of patient and family however their focus was towards how family members/caregivers manage side effects of medication when caring for palliative patients at home; finding out what it is like for family members/caregivers to have responsibility for monitoring patient changes and adjusting medications in the home; and exploring aspects of trust with health professionals and the hospice when there has been a misdiagnosis. The topic of bereavement was directed at evaluating services and supports for adolescents with palliative care needs.
DISCUSSION
Of interest, and consistent with the literature, this study has demonstrated differences between research priorities of professionals and volunteers and patients and their families. However, to some extend the broad context of these topics represented the “how” to achieve the “what” outlined in the patient family member/caregivers priorities above so were not completely unrelated but rather were overlapping. The study shows that professional research priorities are primarily focused on evaluation of clinical outcomes and education, upheld in research priority studies of professionals, particularly medical professionals (Sigurdardottir et al., Reference Sigurdardottir, Haugen, van der Rijt, Sjøgren, Harding, Higginson and Kaasa2010; Daveson et al., Reference Daveson, Harding, Vanden Berghe, Edwards and Higginson2011). For patients and family members, the research needs focused on research into supportive needs of this group and this is upheld by others (Hudson et al., Reference Hudson, Zordan and Trauer2011; Hudson & Payne, Reference Hudson and Payne2011). The small extant literature on research using Delphi or similar methods to investigate research priorities in palliative care, have generally separated the groups. However, a critical examination of this literature has consistently highlighted significant gaps with regard to palliative care research associated with family caregivers (Hudson et al., Reference Hudson, Remedios and Thomas2010, Reference Hudson, Zordan and Trauer2011). Hudson et al. (Reference Hudson, Zordan and Trauer2011) set out to determine research priorities in a web-based survey of 55 family caregivers from 12 countries (response rate, 70%) who had previously expressed interest in, or were members of The International Palliative Care Family Carer Research Collaboration. Similar to the topics identified in our Delphi study the key areas identified by the caregivers as warranting research attention were: under researched family member/caregiver subpopulations (e.g., young caregivers, ethnic minorities); bereavement experience and support, assessing unmet needs; and development of methods (e.g., assessment tools) to identify family caregivers at risk of poor psychosocial well-being (Hudson et al., Reference Hudson, Zordan and Trauer2011).
Grande et al. (Reference Grande, Stajduhar, Aoun, Toye, Funk, Addington-Hall, Payne and Todd2009) drew on a comprehensive review of the caregiver literature and a consensus meeting by experts in the field to conduct an overview of research on palliative caregivers needs. They considered evidence on: adverse effects of caregiving and interventions to address these; the gaps and weaknesses in the evidence; and how to improve the evidence base and interventions in the future. The authors concluded that there was a lack of longitudinal studies on psychological and bereavement issues and challenges that could have uncovered changes in needs over time, the precursors of increased need, appropriate timing of interventions and the consequences of a failure to meet needs.
In a two phase study, Perkins et al. (Reference Perkins, Barclay and Booth2007; Reference Perkins, Booth, Vowler and Barclay2008) set out to identify research priorities of palliative care patients through focus group interviews and a survey with day therapy and hospice in-patients with cancer whose estimated prognosis was six months or less. Of note, and supported by the findings of this study, particularly during the focus group interviews, the researchers found that patients spent most of the time in the groups talking about their own illnesses and how they had affected them and their families. Some participants were able to discuss prioritizing, however most wished to talk more about their own experiences and on what would have improved their own disease journey. They found that symptoms were mentioned infrequently and discussions were more on communication about symptoms, doctors listening more to patients, and patients finding the right form of words to express how they are feeling. These were also emphasized by the patients and family members who participated in this study, with a particular emphasis on wanting to remain in their own homes until death and have autonomy in relation to making decisions and choices about care.
Other groups of professionals and clinicians involved in palliative care practice have set out to determine research priorities. These have included palliative care social workers (Kramer et al., Reference Kramer, Bern- Klug and Francoeur2005), and providers of palliative care for children (Malcolm et al., Reference Malcolm, Forbat, Knighting and Kearney2008; Reference Malcolm, Forbat, Knighting and Kearney2009; Steele et al., Reference Steele, Bosma, Johnston, Cadell, Davies, Siden and Straatman2008). The priorities identified by these groups have generally taken a patient and caregiver focus. Social workers identified that there was a need to address mental health needs, grief, bereavement and psychosocial aspects of well-being, cultural issues and humane care, particularly for the most vulnerable and oppressed members of society (Kramer et al., Reference Kramer, Bern- Klug and Francoeur2005). Research priorities for children were heavily focused on increasing awareness of and improving access to children's hospice care, hospice and respite care needs of young people, and community/home care and issues related to supporting the wider family (Malcolm et al., Reference Malcolm, Forbat, Knighting and Kearney2008; Reference Malcolm, Forbat, Knighting and Kearney2009; Steele et al., Reference Steele, Bosma, Johnston, Cadell, Davies, Siden and Straatman2008). Research priorities identified by hospice organizations (Tolley & Payne, Reference Tolley and Payne2008) have been more mixed. Tolley and Payne (Reference Tolley and Payne2008) found that some of the lack of hospice organization participation in research could be attributed to philosophical differences between hospice care and mainstream medicine where the hospice (as a movement) promotes compassionate care for the dying while participation in research may be thought to impinge on this care.
In an attempt to address collaboration, Hagen et al. (Reference Hagen, Addington-Hall, Sharpe, Richardson and Cleeland2006) conducted a symposium with administrators from the United Kingdom National Cancer Research Institute, the Institute of Cancer Research of the Canadian Institutes of Health Research, and the USA National Cancer Institute. They discussed the feasibility of supporting multinational efforts to strengthen research activities in supportive, palliative, and end-of-life care. Their objective was to catalyze high-quality research that would improve this component of care for patients with cancer or other chronic illnesses. The symposium concluded that there was urgent need for research in: measuring symptoms and symptom clusters; study of the biologic bases of complex symptom clusters; enhancing research in symptom interventions; family and informal care-giving; the impact of poverty, societal beliefs and approaches to death and end-of-life care; measuring quality of life within different cultural milieus; performance of the health care system; understanding and responding to transitions in care objectives; and the use of technology to enhance care, and knowledge transfer (Hagen et al., Reference Hagen, Addington-Hall, Sharpe, Richardson and Cleeland2006). This comprehensive proposal reflects the priorities of both groups of participants in our study.
STUDY LIMITATIONS
This study is limited in two aspects. First, it was conducted at only one hospice site however 30 staff and volunteers participated in the study which is an acceptable number for a Delphi study. The second limitation is that patient and family member/caregiver recruitment was low with only eight participants. The reasons for this are consistent with the literature where patients and family member/caregivers declined to participate for reasons such as: fatigue; family member gate-keeping, for example in some cases patients were willing to participate but were prevented by family members.
IMPLICATIONS FOR PRACTICE AND FUTURE RESEARCH
There are a number of implications for practice and future research that emerge from this study, despite the limitations. Primarily, as palliative care services are steeped in an ethos of patient-centered care it is unsurprising to find that there is some overlap between the views of patients and family members and those of staff and volunteers. Nevertheless, this study demonstrates that important insights into opportunities for quality improvement can be gained through talking with members of the wider health care team (including patients and family members) and making sure that the voices and needs of patients and families are heard.
We were to some degree surprised that volunteers were more aligned to the priorities of professionals than of patients. However, there are multiple factors that may have influenced volunteers having similar research views as professional staff, notwithstanding the complex mix of people who volunteer in hospice and this is an area that requires further examination.
The small number of patient and family member recruits also meant that we could not separate the two into analyzable groupings. We suggest this as a fruitful area for future research.
CONCLUSION
The study demonstrated clearly that research priorities identified by staff and volunteers were similar in focus, while those of patients and carers are targeted toward different aspects of palliative and hospice care. Because of the differences in perspective we have kept the two lists of research priorities separate. However, while the two lists are not entirely separate, staff and volunteers focus on education and service provision and symptom management, while patients and family members were predominantly concerned with support with a strong emphasis on maintaining autonomy at all levels of engagement with the palliative care services.
It is entirely consistent with the literature that professional and volunteer workers have a different viewpoint from those for whom they care. It was not surprising that each group focuses on their own area of concern, and indeed we suggest that they may not each know what the gaps in knowledge of the other group are. Patients, for example, will not know what professionals do not know about symptom management, while those who work in the service cannot know exactly what concerns family and caregivers may have over and above those disclosed during care episodes consequentially both sets of priorities are important. A challenge for research funding bodies and palliative care services is to ensure that research that patients and caregivers deem important is also conducted.
ACKNOWLEDGEMENT
The authors would like to acknowledge Te Omanga Hospice, Upper Hutt, New Zealand for funding and supporting this project
