Introduction

A bibliometric review is an appropriate approach utilised for mapping a research field providing a comprehensive picture of the development and current status of a research field over a long period of time.1 This approach uses ‘a set of quantitative methods to measure, track and analyse print-based scholarly literature’ published over a long period.2 Bibliometric approaches process a large number of research published over a long period with a limited investment of resources and time.3 Bibliometric analysis produces measures of productivity (eg, numbers of published papers); impact (eg, number of citations, journal impact factors) and cooperation across countries, institutions and authors.4 It provides an understanding of the growth of scientific research and their trends within a specific field and contributes to the development of health initiatives.5 A bibliometric analysis is recognised as a statistical evaluation of published articles, and it is an effective way to measure the influence of publications in the scientific community.3 It portrays a primary step towards global mapping and classifying palliative care (PC)-related research, which may assist in understanding the commonalities of research, identify research gaps, set research agenda and map future research directions. A bibliometric analysis is based on using encoded bibliographic information available in scholarly databases4 for examining publication trends, the contribution of researchers, countries and institutions to the development of research through generating and visualising maps based on network.3

Although the use of a bibliometric approach on mapping a research field is important, relatively little efforts have previously been directed to map and analyse PC-related research trends. For instance, Walshe and Ahmed6 assessed the relation between contributing authors and PC-related journals to highlight geographical bias that may affect access to evidence. They found that most of the authors were from North America and Europe. They also showed that American authors published their papers more commonly in American journals and European authors published in European journals, lacking cross-fertilisation of the available evidence. Cheong et al,7 mapped the PC-related articles published from inception until 2018 in the Asia Pacific Region. Findings revealed that out of 32 Asia Pacific countries, only 18 high-income countries (HICs) within this region published relevant research, and most of the published articles focused on cancer. Clark et al,8 mapped and synthesised the international PC-related research published from inception until 2013. They found that most research was from Europe. Most research focused on evaluation of services, policy-makers and policy-related issues. Liu et al,9 presented PC-related research trends from 2001 to 2016 and observed that the USA and UK took a leading contribution in PC research. Furthermore, Curiale10 mapped the global geriatric PC research in 2009. Results showed that most research was conducted in North America, and most of the published papers focused on advanced dementia and cancer. Despite the significant findings of previous studies, most of them limited their searches to articles in specific geographical settings. Moreover, the ‘sample’ included in the previous studies appeared to underrepresent the publications in the PC field as it was extracted from one database, being either Scopus or Web of Science (WOS) databases. Furthermore, all of the previous studies did not provide a comprehensive visualisation of PC-related research in the globe and did not use advanced programmes for mapping research that offers additional statistics such as degree of cooperation between countries, authors and institutions. Considering these limitations, previous studies did not offer a clear, representative and comprehensive picture of the state of research within the field of PC in the world.

Therefore, this bibliometric review is conducted to provide an updated overall picture and systematic mapping of the state of research trends within the field of PC from 2002 to 2020.The findings of this study indicate potential directions that the researchers, journals, institutions, and countries should consider for enhancing the PC field, and will highlight research fields that require further research attention.

MethodsStudy design

A bibliometric approach was employed to map the research literature on PC using metadata extracted from two databases over 18 years. Searching was conducted on 5 May 2020, and was updated on 6 May 2021.

Search sources and strategies

Searches were performed on Scopus and WOS, being the largest,11 most comprehensive and multidisciplinary databases utilised in this type of review. Both databases provide different citation impact metrics such as Field-Weighted Citation Impact to indicate the impact of a publication. They cover literature in the field of biomedical and social sciences. Scopus also covers 100% of Medline material.12

The search strategy of this review is based on the following terms, including ‘palliative care’ with its alternative search keywords (‘palliat*’, ‘palliative medicine’, ‘hospice care’, ‘terminal care’, ‘end-of-life care’, end of life care, ‘life-limiting’, ‘life-threatening’, and ‘incurable disease’) combined with the Boolean operator ‘OR’ for generating a large number of results, performed with the help of a university librarian. The terms were used for screening the title and the abstract to identify relevant papers. The search was limited to original articles published in English language between 1 January 2002 and 31 December 2020. The detailed search strategies and search results for both databases is presented in online supplemental table S1.

Inclusion criteria for considering papers in this reviewStudy selection and data extraction

The function ‘relevance’ in both databases was chosen to sort articles that most closely matched with the searched terms making easier to assess and screen articles. Titles and abstracts of articles were assessed by the first author, and those that did not meet the criteria were excluded after discussion with the second author. The final articles included from both databases were exported into ‘Comma- Separated Values’ file, which was subsequently utilised to remove duplicates based on each paper’s digital object identifier. Reasons for excluding articles were identified and documented in online supplemental figure S1.

Data analysis and visualisation

Before analysing the results, all inconsistencies in the bibliometric data, including names of journals and references, were adjusted and corrected. The analysis of the bibliometric metadata covers two main pillars, which include performance analysis and science mapping.13 Performance analysis aims to assess authors such as the year of publication, the number of citations, the productivity and impact of institution, country, journal and the impact of cited publications. Productivity is assessed by the number of publications, while the number of citations is an indicator of importance relevance14 and impact of publications. When studying the annual trend of PC-related publications, exponential smoothing with Brown’s linear trend was adopted to forecast the number of annual publications for the next 5 years. Included countries were categorised by income level, where low-income and middle-income economy countries (LMICs) refer to developing countries, and high income ones refer to developed countries.15 Identifying the geographic region of the journal is based on where the journal is hosted/based, acknowledging that the editorial boards of such journals and the content they publish is often wider or have more global coverage.

Gaining such information helps to find relevant literature and to support decision making regarding where to publish. Authors were converted to rank order by Standard competition ranking. The top 15 ranked authors were considered for further visualisation and presentation. Different values such as Impact Factor (IF) and SCImago Journal Rank were used for ranking and prioritising the authors.

Science mapping aims to explore the interrelation between institutions, disciplines, fields or specialities to draw content-related conclusions.13 Two software packages (VOSviewer (V.1.6.13) and Gephi (0.0.2)) were used for science mapping of PC research outputs. VOSviewer is a free software that allows researchers to generate and visualise maps easily.16 It is used to construct and visualise bibliometric maps according to network data.16 It was used for conducting a series of coauthorship analyses, including patterns of cooperation between countries and institutions by measuring occurrence/frequency and total link strength, which are used for measuring patterns of directions. For instance, in keywords analysis, total link strength reveals the publications’ number in which two keywords occur together. Cooperation across countries/institutions was categories into weak and strong, where weak cooperation means that the work between countries in PC related research is minimal (Have little collaboration). It can be determined by measuring the link between nodes such as countries. The thick link between nodes is representing strong cooperation across countries.

Gephi V.0.0.2 is another free software for visualisation and analysis of large network graphs.17 It was used in this study to estimate values such as centrality and betweenness centrality, which are commonly used criteria for analysing coauthorship.18 Degree centrality is the most straightforward approach which presents the number of relations a node has to other nodes.18 It is calculated by counting how many neighbours a node has.18 Betweenness centrality was calculated to estimate which nodes are more influenced when two or more nodes have the same equal value.17 18 It was used to measure the number of times a node lies on the shortest path between nodes.17 18

Discussion

This is the first bibliometric review that has provided an overall picture and systematic mapping of global PC-related research between 2002 and 2020. Our analysis revealed significant expansion of PC-related research, particularly from 2015 onwards. These findings are in line with past research.9 There are several possible explanations for this result. It might be related to the ageing world population and increasing demand for PC.23 It could be linked to responding countries to the WHO’s PC policy calls.24 It also might be associated with PC being formally recognised as a medical specialty from several countries more recently.25

It is not surprising that the most significant number of articles was from North America and Europe because more than half of the top 15 authors contributing to PC field are working in institutions in the USA and the UK, suggesting that American and British researchers have taken the leading contribution in the PC field. This result is congruent with Walshe and Ahmed recent study.6 The USA has a long history of PC development, and it has witnessed a rapidly growing provision of PC within its healthcare system more recently.26 For instance, in 2010, two-thirds of USA hospitals offered PC.26 High productivity of PC-related publications in USA, the UK, Australia and Canada might be attributed to the study’s inclusion criteria that limited analysis only on English language articles. Publications in other languages such as Chinese, Korean, German and Spanish would be more frequent and useful for scholars and other practitioners in Latin America.27 High productivity might be related to the country of reviewers, editors, and investigators, as most of them are from North America and Europe.6 Furthermore the congresses and meetings held by these three regional organisations (APCA, European Association for Palliative Care (EAPC) and Asia Pacific Hospice Palliative Care Network) are mostly in English while the congresses of the Latin American Association for Palliative Care are in Spanish and Portuguese. This may also be a contributing factor to more publications in English from these regions.

Although there is an increasing number of published articles in some countries, their research citations remain low. This finding is in agreement with the National Science Board28 report findings, which showed high frequency published papers with low citation. The increasing number of publications seen particularly in India may be attributed to India’s regulations that obligate post-graduate students to publish papers in indexed journals.29 Such regulations on conducting research help drive PC development which is observed in India30 that have witnessed successful growth of PC in recent years. Research is, therefore, an essential component of PC31 which plays a significant role in developing PC globally. Considerable efforts are required in focusing on innovation and building research capacity in the future, particularly in LMICs that seek on developing PC at a national level.

Study findings also highlighted that most PC journals’ IF was down in 2020 compared with previous years. IF reflects the importance of a journal by measuring the yearly average number of citations published in the last 2 years in a journal.32 The higher the journals' IF, the better the journal.32 Several factors influence journals' IF such as the number of papers published and an average number of citations.33 The lower IF of PC journals’ might be related to increasing publication output in recent years. Increasing publication output does not always translate to better quality papers published or more citations, which may be one of the reasons for a decreased IF in several PC journals last year.

The cooperation between scholars was limited. Minimal international cooperation was observed across countries and within institutions. International cooperation in health research is a quintessential mechanism for promoting knowledge, increasing research capacity and achieving breakthrough results.34 A possible explanation for these results might be the differing priorities of research between developed and LMICs. Another explanation is that most sponsoring/funding institutions are focusing on research related to their priorities in a single country.35 Thus, there is an urgent need to establish an international network, share ideas, and ensure transparency for increasing partnerships and setting research agendas. Such processes are led by international institutions such as the APCA and the EAPC. For instance, EAPC plays a pivotal role in supporting and organising global scientific and educational events/conferences, aiming at increasing cooperation between scholars globally, bringing together those who study and practice the PC-related issues, sharing ideas, and updating knowledge-related PC.36 Despite the focus of the EAPC in Europe, it also aims at enhancing PC in LMICs through the exchange of information and expertise.36 More collaboration on PC-related research between scholars in both within and across HICs and LMICs are required for enhancing PC services and quality of care. These can be established by setting appropriate policies and strategies that support cooperation and facilitate the exchange of ideas about PC.

Although PC is a comprehensive approach focusing on all patients with life-limiting conditions,37 the majority of PC-related research focuses on adult cancer-related issues within the study period (2002–2020). The results are consistent with past research.38 Adult cancer patients are the larger group of patients with life-limiting illnesses, often experiencing a wide range of symptoms that impact their needs.39 More attention to cancer was further observed after 2014. This might be attributed to countries' response to WHO’s call on implementing polices to ensure the development of PC services within countries' healthcare system, as cancer is often the focus of countries when developing their PC services.40 Our study findings also show that most PC-related research still focused on education, services provision, quality of life and pain-related issues. These results are in line with a systematic review that assessed priorities of the international PC research in 2020.41 Enhancing patients’ quality of life and comfort is PC’s primary goal.42 PC aims at relieving serious health related suffering.42 After 2015, major PC-related research shifted its focus to dementia-related areas. Transitional demographic changes and increasing the life expectancy in population might be the reasons for the increasing focus of PC in dementia more recently.43 It might be attributed to the growing number of people diagnosed with dementia.44 By 2050, the number of people diagnosed with dementia is expected to increase significantly, reaching 125 million globally.45 A small number of studies have focused on other life-limiting diseases such as respiratory diseases, kidney diseases, HIV, Parkinson’s, stroke and liver diseases. Further studies, which take these disease types into account, will need to be undertaken. While our search covered a period of a few early months into COVID-19 and indeed a small number of such papers was included in the list of identified studies, COVID-19 as a potentially life-threatening disease may be a new focus of PC research in the near future.

Pain-related issues are still one of the major focus of PC-related research. PC isused to ease pain and other associated symptoms such as physical, social, emotional and psychological symptoms. Availability and affordability of pain-relief medication is required while patients receiving PC, as pain management is critical component in PC services. Unavailability of and poor access to pain-relief medicationare unique challenges to the provision of PC in LMICs.46 Over 80% of the patients with PC needs in LMICs have limited or no access to such medication.47 This high percentage should be considered while developing PC in LMICs.

Findings also highlighted that 141 countries out of 195 countries globally had published papers related to PC. This means there are no PC publications from about one-quarter (54 countries) of the world countries, primarily LMICs where they may have significant PC needs and often outside the cancer context.

PC education has been recognised as the key step towards improving healthcare professionals’ knowledge, attitudes and skills.48 That is why this topic has frequently been reported in past studies, although most of the studies focused on educational assessments. Very little was found in the literature on interventional studies and curriculum evaluations, particularly for those non-cancer life-limiting diseases as reported in Meekin et al 49 study. Further work is required in this area.

Most patients with life-limiting diseases develop potentially devastating symptoms during their disease trajectory.50 In this study, most papers focused on managing specific symptoms, particularly anxiety, depression, pain and delirium; however, other crucial symptoms in a PC context such as constipation, cough, dyspnoea, xerostomia, nausea, vomiting and fever50 were not a key focus in the list of common keywords analysed. Future studies on many such symptoms are therefore recommended.

Survival and mortality topics have not received adequate research attention in PC too. However, other topic focus on general practitioners (GPs) and quality indicators need further research attention. GPs play a pivotal role in providing PC in the community.51 Many GPs are not confident and comfortable with PC when working in community settings and particularly when confronted with dying patients because of their inadequate level of skills and insufficient knowledge.52 Further research should be done in this area to deliver best practice PC such as the role of GPs in delivering PC to patients. Other areas such as assessing GP performance on working more effectively in the community, their knowledge and attitudes towards PC, and communication skills with patients are also necessary.

Measuring the quality of care which is provided to end-of-life patients is an important indicator enabling HCPs and policymakers to monitor and enhance care provision.53 Quality indicators can identify good care and potential problems.54 Although this topic is important; only limited attention has been paid to quality indicators for the care provided to patients with life-threatening conditions as reported in this study. Further research should focus on the different types of quality indicators, including process, structure and outcomes.

Strengths and limitations

This review is subject to several limitations such as limiting searches only on original papers available at Scopus and WOS databases. The search was restricted only to papers written in English, and this limits generalisability. In addition, the searches were based on specific keywords clearly stated in title and abstract of the paper; this might have not included all papers related to PC in this review.

Adopting the frequency of common keywords appearing on papers as a way to assess the direction of PC-related research, rather than indicating the research focus of articles, might be another limitation in this study which could mislead the future direction of PC-related research and be perhaps a linguistic indicator. Furthermore, assessing the quality of papers based only on their citations is also another limitation in this study as it is difficult to appraise a large number of papers and quality of each paper is not the purpose of a bibliometric analysis, which is focusing on outcome metrics. Meanwhile, it is not easy to assess in-depth all papers included in terms of their research type, sample size, which is another limitation in this study. Furthermore, another limitation may be the use of only two databases, as the analysis packages used require harmonised databases to be merged, a complex task to be done with multiple databases (and indeed most bibliographic analyses in the past tend to use a single database only because of this).