International palliative care research priorities: A systematic review

A systematic review of research priorities in palliative care was undertaken and guided by the PRISMA statement for reporting systematic reviews [38] .

A systematic search of databases from health sciences, medicine, and psychology was undertaken in August 2017 in conjunction with a subject librarian. Six databases were searched: Cumulative Index of Nursing and Allied Health Literature (CINAHL), Excerpta Medica database (EMBASE), PubMed, SCOPUS, Web of Science, and PsycINFO. Grey literature was identified via ProQuest Dissertations and Theses, CareSearch grey literature, James Lind Alliance Website, Lenus, and the Palliative Hub of the All Ireland Institute of Hospice and Palliative Care. A further search of the grey literature was conducted on the following sites in April and May 2019: OpenGrey, European Association of Palliative Care (EAPC) conferences, Australian and New Zealand Society of Palliative Medicine and Google.

Key words were identified through the titles, abstracts, and indexed phrases of relevant articles from a preliminary search of PubMed and CINAHL. Indexed terms from the selected databases were identified and included in the search terms for these specific databases (Table 1). Articles were limited by publication date (January 2008–June 2019) and language (English). Subsequently, indexed terms (e.g., CINAHL headings, mesh terms) from the selected databases were identified and included in the search terms for these databases

The search yielded 10,235 articles from which 2007 duplicates were removed (Fig. 1). Two reviewers (EN and FH) uploaded titles and abstracts of the remaining 8318 papers into Covidence for initial screening, of which 8252 were considered irrelevant and excluded. Four reviewers (EN, DM, FH and OB) screened the full texts of the remaining 66 articles against the review’s inclusion/exclusion criteria (Table 2).

Ten studies were eligible, and they were included for quality appraisal (see Fig. 1). Any conflict of opinion regarding which article to include or exclude was resolved through discussion among the reviewers. If no agreement could be reached between the reviewers, a fifth reviewer (SM) mediated.

Quality appraisal was undertaken to gain an understanding of the results and level of confidence in the findings. The quality, methodological rigour, and risk of bias of the ten studies included in the final review were assessed using the Joanna Briggs Institute (JBI) Critical Appraisal Tools [39]. JBI critical appraisal tools were selected based on their appropriateness to the methodology in the papers. The qualitative research tool [40] was utilised for papers using workshops, the systematic review tool [41] was used for evidence reviews, and the cross-sectional research tool [42] for research using surveys and questionnaires. To prevent the introduction of bias and enhance transparency, four studies [21, 43‐45], adopting multiple methods and consensus methods, were not subject to quality appraisal due to a lack of specific tools. No study was excluded based on quality.

Quality assessment for included studies was completed independently by EN & DM. Any unresolved variances were resolved by a third author (FH). These appraisals were summarised respectively and presented the grading using a range of coding systems. The scores were computed by counting the number of `Yes’ answers. A coding system of 0–10 was applied to three papers; 8–10 was considered high quality, 4–7 moderate and below 4 poor. Of the three papers one was considered high and two moderate quality. Two papers were accessed using a coding system of 0–8: 7–8 was considered high quality, 6–4 moderate and 3–1 poor. The two papers were each accessed as moderate quality. The remaining paper was accessed using coding system of 1–11; 8–11 was considered high quality, 4–7 moderate and below 4 poor. In terms of quality this paper was rated as high quality (see Table 3).

A data extraction form was developed on Microsoft Word to extract key data from the included studies. Data extracted included author, year, and aim of the study, geographical location, participants, method, data analysis and priorities identified. Three reviewers (EN, DM and OB) independently extracted data from the final ten papers using the data extraction form and any disagreements were resolved through discussion. A fourth reviewer (FH) was consulted if an agreement could not be reached.

Categorical data including year, country, participants, and method were extracted and analysed in Microsoft Excel. Qualitative data underwent a thematic synthesis [37] to integrate the findings of multiple studies and identify which priorities were the most common across the ten included papers and this allowed for the development of higher order themes. Synthesis included line by line coding of the findings of the primary studies, and the categorisation of codes into broad groups of research priorities followed by descriptive themes [37, 46]. The final stage in the analysis was the development of broad analytical themes.

The data extracted from the ten included studies [15, 21, 43‐50] are presented in Appendix. The largest group to shape the research priorities were academic, commissioners and healthcare professionals (HCPs) and these were included in eight of the ten studies [15, 21, 43‐45, 47, 49, 50]. Families and carers were the second largest group to contribute to the data and were included in two studies [15, 47] . Patients were the sole contributor in one study [48] and contributed as part of a group of patients and families in two other studies [21, 43]. However, these were the only studies in which patients shaped the research priorities. Six of the studies included researchers [15, 43‐45, 49, 50], two included researcher/clinicians [43, 50], one included members of the public [15] and two included palliative care volunteers [15, 43]. Details on the sample sizes for each participant group can be found in Appendix.

One study included a search of the international literature and as a result the data was not exclusive to one jurisdiction [46]. The geographical location of the studies was diverse. Three of studies were based in the United Kingdom [44, 47, 48], while one was conducted in both the United Kingdom and Ireland [20]. Each of the remaining five studies geographical location were conducted in New Zealand [21], United States [49], Canada [43], Australia [45] and Africa [50] respectively.

Consensus methodologies were the most commonly used method of developing research priorities. However, there was little consistency in how consensus was gained across the studies as methods were operationalised in different ways. For instance, two studies [21, 45] utilised the Delphi technique while others used workshops [15, 47, 49], a nominal group technique [44, 50] or a mixed methods approach involving literature review interviews and online surveys [43] respectively. One study used a systematic review methodology alongside an innovative analytical approach to synthesise evidence from review articles or consensus reports to develop a list of research priorities [46]. One study used a questionnaire as the only method for developing the priorities [48] while others incorporated surveys as one phase of a single priority-setting exercise [15].

Following the thematic synthesis [37], the data from the studies were organised into seven descriptive themes, which are described in more detail below.

The seven themes that emerged from the thematic synthesis were closely aligned to the data extracted from the ten included studies. However, it was important to further interpret this data in order to generate a higher-order explanation for the findings. The Donabedian Framework [36] is a model of assessing quality of care and provides a useful mechanism for displaying and analysing this data. This framework will help to provide a standardised model to summarise the research priority results according to quality indicators and identify gaps in the evidence.

The seven descriptive themes are reflective of and mapped onto the three interrelated categories contained within the Donabedian framework, where each category influences the one that succeeds it. Structure refers to the attributes of the settings in which palliative care is delivered. Process involves the activities, from both professionals and patients, that are carried out in giving and receiving care. Outcomes denote the impact of the care on patients and families [51]. Three of the themes (Service Models, Continuity of Care, Training and Education) are centred on structure and the physical and organisational features of palliative care service provision. Inequality of Access and Communication are elements of the process of care delivered to the patients While Patient Preference and Experience and Recognising the needs and importance of Family Carers are related to the outcomes of care for these populations (see Fig. 2).

The relative contributions of each of the ten papers to the final themes was examined and the data is displayed in Table 4. Each of the ten studies contributed at least three of descriptive themes and only one study [50] did not contribute to all three analytical themes (as outlined in the Donabedian Framework). Two of the studies [15, 49] contributed to all seven descriptive themes.

Seven priority areas were identified from the ten papers included in the systematic review; Service Models, Continuity of Care; Training and Education; Inequality of Access; Communication; Patient Preference and Experience; and Recognising the needs and importance of Family Carers. The themes were mapped to the Donabedian Framework [36], which highlighted that the priorities were associated with the setting, structure variables and the effects on patient outcomes. Despite research emphasising the inclusion of patients and caregivers in research, only five studies [15, 21, 43, 44, 48] included the patient and/or caregiver. The need to place the patient at the centre of this process in line with policy is advocated in many countries. This review also analysed the methods used for priority setting, indicating varied approaches were adopted. The majority were based upon consensus methodology however, the operationalisation of consensus and transparency of the process was lacking. This confirms previous work by Viergever and colleagues [28] which suggests there is no gold standard for setting research priorities emphasising the need to improve the rigorous and transparent reporting of methods.

Internationally there is a paucity of research considering the priorities for palliative care research originating from Western countries; only ten papers were identified that contribute to priority setting relevant across palliative care provision. While a proportion of the literature in this area focuses on palliative care provision for individual disease-groups and conditions, it is important to recognise the opportunities for shared learning and commonalities across all populations. This review is the first to synthesize international research priorities for palliative care that have been obtained from empirical research and that are not disease or population specific, onto a tangible framework within the broader healthcare context. The Donabedian framework, has been successfully applied across healthcare settings and contexts as a means of evaluating quality, and within the palliative care context, for example, as a systematic review framework [52], and as a proposed framework for intervention and evaluation studies [53, 54]. The utilisation of this framework transformed the thematic synthesis into higher order analytical themes that can be taken forward strategically to improve palliative care research. To the authors’ knowledge, this is the first systematic review to apply the Donabedian’s framework to mapping palliative care research priorities.

Palliative care is a field of healthcare in receipt of a historically small proportion of research funding [8, 9], therefore, it is essential to ensure that research value is maximised. In 2009, Chalmers and Glasziou [55] estimated that 85% of all health research is being avoidably “wasted”. Whilst it can be argued that progress has been made, over the last decade, Glasziou and Chalmers more recently [56] claimed that health service research still has a have a long way to go, with continued concerns over research design, conduct and reporting. When analysing the priority areas identified, it is important to be cognisant of these concerns and ensure that the palliative care research projects to address these priorities are of sufficient quality and rigour to address and mediate such concerns.

One of the central issues about priority setting is addressing the question around who should be involved in the process, and how can this be enabled. Findings from this review suggests the care providers are the dominant perspective, with the patient’s largely missing. This may explain why priorities are largely service orientated with notable gaps in the priorities relating to quality of life and symptom management. Given that patient perspectives routinely differ from those of other stakeholders [57], the need to elevate their voice to enhance the legitimacy in the identification of priorities is required. Doing so will help to ensure that future research addresses questions of relevance, helping decision-makers and service providers to be better equipped to design and deliver health services to meet patient/service user need [58]. Whilst researchers have acknowledged the need for greater patient involvement in research and planning [59], their inclusion is questioned by an array of ethical, practical and medical challenges [60], further complicated by researcher concerns about their roles and values [61, 62]. Nevertheless, the need to find new robust processes to support meaningful contribution of patients to enable inclusion in strategic directions in palliative care is required to ensure the inequality [2] is reduced and the research is not considered wasteful [35].

A second issue relates to how information is collected. Variations of methodological approaches exist, with little agreement on what constitutes reporting standards, guidelines or best practice [28, 63]. Findings in this study demonstrate great variation in the approaches used to organise research priority-setting exercises. Whilst consensus was the most common means to generate research priorities in the ten included studies, a myriad of approaches were used such as the Delphi technique, nominal group and consensus workshops, as well as a range of analysis techniques, for instance, ranking, statistical analyses and an immersion crystallisation framework. Within the ten studies, inconsistent reporting of priority development was noted and was a barrier when synthesizing the evidence across studies. Such variation and lack of consistency make it difficult to judge the validity and transferability of the priorities reported, creating a significant barrier to aggregating and reporting comparative findings across international contexts [63]. Given that credibility of consensus findings is influenced by the rigour application of the approach, we need to ensure the reporting and guidance by acceptable standards in palliative care is advocated [2, 64]. Moreover the application of the checklist to help standardise research priority setting in health could be used to inform priority setting exercises going forward [28].

A third issue relates to the context and landscape in which the priority setting is undertaken. In this review, the majority of priority setting exercises were conducted in countries with an preliminary (i.e., New Zealand) or advanced (i.e., UK, Ireland, U.S.A, Canada and Australia) level of palliative care programmes and integration into mainstream health services [65]. A gap in the knowledge of priorities representing those in lower and middle-income countries (LMIC) exists, echoing previous reviews of international palliative care research [66, 67]. Whilst there is evidence in this review of engagement, with researchers from high-income regions collaborating and undertaking research in Africa (Powell et al., 2014), a number of researchers [31, 67] suggest this raises multi-faceted challenges including the risk of imposing western norms in differing cultural contexts [68]. Therefore, the application of western research priority findings is limited, if not adapted to the specific economic, cultural and specific health care context and constraints of lower- and middle-income countries. Zaman et al. [31] suggests the need for LMIC to initially develop culturally and locally appropriate research, and then move towards international comparative research.

This review represents an initial step towards mapping international palliative care research which may help to inform policy and funding bodies on future action. However, it has several limitations for example, the search was limited to English language articles, which limits the generalizability of the findings. Moreover, it is recognised that this review excluded, disease specific empirical studies such as dementia and for specific populations such as intellectual disability however, the inclusion of such evidence would have resulted in greater heterogeneity between studies and limited the ability to synthesise the findings. The search was also limited that the exclusion of patient and public involvement which may have captured more caregiver and patient perspectives. Additionally, some of the studies lacked detailed information on the methodological analysis and procedures employed, questioning the rigour and validity.

This systematic review has called attention to the need for more end users in research priority setting exercises, therefore, researchers and funding bodies should develop new strategies to ensure meaningful participation of palliative care patients and families, building in structures and processes to account for the vulnerability often present within this population. Findings provide an initial blueprint for palliative care research funders and policymakers to contribute to the future research agenda for palliative care from a patient and HCP perspective. Given that funding resources are limited the importance of collaboration and international approaches to palliative care is growing, these findings may help to inform this debate.

Methodologically, a standardised approach and reporting for priority setting is advocated allowing for increased validity and comparability of findings from across palliative care settings. Due to varied methods and analytical techniques, an additional challenge was presented for the authors of this review when trying to compare and weight studies. Future attempts to set research priorities should involve a multi discipinary representation of stakeholders, such inclusion will provide credibility and enhance the feasibility of the developed priorities. Whilst it is outside the remit of this review to specific an appropriate prirotiy setting methodology, the conduct of any such exercises should be governed by methodological guidelines, clear objectives and defined criteria and concepts, for identifying and ranking priorities. Doing so, will aid the transparency of the process and credibility of the results.