Ten years of research and publications about hospice and palliative care in Mainland China: a bibliometric study

Background Research is a fundamental element in the sustainable development of hospice and palliative care. Mainland China is facing increasing demand for hospice and palliative care and has launched related policies over the past decade. However, the state of research and publications in this eld in China remains largely unknown. This study aimed to provide an overall picture of hospice-and palliative care -related research and publications in Mainland China from 2010 to 2019. Methods We searched Web of Science, Scopus, PubMed, CINAHL, OVID, and China National Knowledge Infrastructure for hospice-and palliative care -related publications in English and Chinese for 2010–2019. We analyzed the production, citations and impacts, publishing journals, region and institution of origin, and themes and active topics. Results A total of 3224 publications were identied, and 636 of them were considered of high quality. The production and impacts showed a clear increase—especially after 2016. However, the regional disparity between East and Western China in production was conspicuous and closely linked to economic factors. Beijing and Shanghai were the most productive regions. The hospice providers in the rst ve pilot regions had no collaboration with leading universities in their publications. Hospice and palliative care for cancer patients was the most common publication topic; some essential themes were rarely explored. Conclusion Research and publications of hospice and palliative care in Mainland China is developing faster than before; however, it remains at an early stage and should be promoted in terms of interregional equity. Collaboration among different disciplines, institutes, and regions should be encouraged. including China’s largest academic database, the Chinese Journal Full-Text Database (CJFD). relevant articles published from the international databases if they contained “Chinese” or “China” in the aliation address and contained any of the following terms in the title, abstract or keywords: “palliative care,” “hospice,” “end-of-life care,” or “terminal care.” The searching strategy is shown in the additional le. When searching for Chinese articles in CNKI, we applied various Chinese versions of “HPC” as follows: “huan-he-yi-liao,” “shu-huan-yi-liao,” “shu-huan-liao-hu,” “gu-xi-yi-xue,” “gu-xi-zhi-liao,” “gu-xi-guan-huai,” “an-ning-liao-hu,” and “lin-zhong-guan-huai.” All the above terms have appeared in Chinese literature at different times and in different regions. We restricted the literature types to consensuses, guidelines, editorials, original research, and case studies. To eliminate the possibility of duplication with original research, we excluded conference proceedings, master’s theses, and doctoral dissertations. After completing the retrieval, we reviewed the titles, abstracts, and author aliations of the documents according to the following exclusion criteria: (1) The topic is curative medicine. (2) The article was in English and focused on Taiwan, Hongkong, Macao, and other places outside Mainland China. (3) The article was a translation of work that had already appeared in an international journal. We did not exclude articles if they were in Chinese and described the situation of HPC in places other than Mainland China: such articles play an essential role in distributing knowledge and reect people’s interest. Two reviewers (S.J. and S.L.)


Background
Hospice and palliative care (HPC) is a fundamental human right and a crucial part of an integrated, person-centered health service. The world as a whole has made rapid progress in this sector, but HPC in Mainland China has long been in a state of infancy [1]. According to the 2010 Quality of Death Index by the Economist Intelligence Unit, China ranked 37th among the 40 countries and regions [2]. Five years later, in 2015, the second edition of Quality of Death Index ranked China the 71st among the 80 countries and regions, revealing limited progress [3].
Though HPC has developed slowly in Mainland China, the demand for it has undoubtedly increased. China has a rapidly aging population: in 2019, 17,599 million Chinese were aged over 65 years (12.6% of the total population) [4]; that proportion is estimated to reach 26.9% in 2050 [5]. Meanwhile, chronic noncommunicable diseases (including stroke, ischemic heart disease, lung cancer, chronic obstructive pulmonary disease, and liver cancer) have become a major burden; they are the main causes of years of life lost [6]. According to the Serious Health-Related Suffering (SHS) Database in 2015, it was estimated that 10443 thousand people (7.55‰ of the population) in China were in need of palliative care [7]. Meanwhile, futile medication at the end of life is still prevalent in China [8]. According to a recent study, aggressive care and catastrophic health expenditure are common for cancer patients and families. More than 80% of the families of cancer patients were left below the poverty line after the patients' death [9]. Most patients are willing to learn more about HPC, but few have access to it [10].
In light of this situation, the Chinese government has promoted related policies over the last decade. In 2012, Shanghai began promoting a community-based hospice service [11]. In 2014, China, along with the other 194 countries in the 67th World Health Assembly, committed to enhancing HPC as a component of comprehensive care throughout the life course [12]. In 2017, the former National Health and Family Planning Commission of China issued a guideline for hospice care [13]; subsequently, ve regions were chosen for a national pilot on hospice implementation. In 2019, the number of pilot regions expanded to 71 [14].
Though China's policy making has become active in this regard, there has been no adequate promotion of research. The role of research has been identi ed as a fundamental element for sustaining the development of HPC [15]. The lessons learned from previous research have shaped and fueled the implementation of HPC within the standards of evidence-based medicine [16]. Thus, the state of research and publications should be regarded as a key measure when examining the development of HPC in a region. Previous reports have identi ed a growth in HPC-related publications, both globally and in the Asia-Paci c region [17][18][19]. However, documents in Chinese were not included owing to the language barrier. Recently, investigators from Hong Kong managed to avoid the "Tower of Babel" bias by searching in Chinese: they found a considerable knowledge gap related to HPC research in Mainland China [20]. However, the authors focused on original studies with a strict inclusion criterion and failed to provide a whole picture of the structural changes in the HPC-related literature that have occurred over recent years.
In the present study, we sought to remedy that de cit by analyzing HPC-related publications from Mainland China from 2010 to 2019, including both primary and secondary literature, in English and Chinese. Employing bibliometric tools, we investigated the production, regional distribution, impacts, and development related to this young discipline. We aimed to depict a panoramic, yet detailed, view of HPC-related research and publications in Mainland China: that is crucial to understanding the present situation and future direction for the sustainable development of HPC. We also hope that the study could provide bene cial information for promoting research in China and other low-and middle-income countries.

Data source and document retrieval
In this study, we searched the Web of Science (WoS), Scopus, PubMed, CINAHL, and OVID for publications in English. The China National Knowledge Infrastructure (CNKI) was searched for ones in Chinese. CNKI is the largest digital publishing platform in China, and it covers a whole category of academic resources, including China's largest academic database, the Chinese Journal Full-Text Database (CJFD). We retrieved relevant articles published from January 2010 to December 2019 from the international databases if they contained "Chinese" or "China" in the a liation address and contained any of the following terms in the title, abstract or keywords: "palliative care," "hospice," "end-of-life care," or "terminal care." The searching strategy is shown in the additional le. When searching for Chinese articles in CNKI, we applied various Chinese versions of "HPC" as follows: "huan-he-yi-liao," "shu-huan-yi-liao," "shu-huan-liao-hu," "gu-xi-yi-xue," "gu-xi-zhi-liao," "gu-xi-guan-huai," "an-ning-liao-hu," and "lin-zhong-guan-huai." All the above terms have appeared in Chinese literature at different times and in different regions. We restricted the literature types to consensuses, guidelines, editorials, original research, and case studies. To eliminate the possibility of duplication with original research, we excluded conference proceedings, master's theses, and doctoral dissertations. After completing the retrieval, we reviewed the titles, abstracts, and author a liations of the documents according to the following exclusion criteria: (1) The topic is curative medicine. (2) The article was in English and focused on Taiwan, Hongkong, Macao, and other places outside Mainland China. (3) The article was a translation of work that had already appeared in an international journal. We did not exclude articles if they were in Chinese and described the situation of HPC in places other than Mainland China: such articles play an essential role in distributing knowledge and re ect people's interest. Two reviewers (S.J. and S.L.) independently screened the documents for exclusion. Con icts were resolved by discussion and mutual agreement. The procedure is shown in gure 1. To conduct an additional impact assessment and content analysis with a focused data source, we created a subgroup of high-quality publications. The quality of a speci c article is very hard to determine. So, we used the quality of the publishing journal as a substitute for the quality of the article. If an article is published in a core journal according to the Peking University Library [21], or in a journal included in internationally recognized databases like WoS, Scopus, etc., it means that the article passed the rigorous peer review and can be considered as a high-quality article. The list of core journals of Peking University Library has been widely accepted in China for years. Searching for articles within core journals is a built-in function of CNKI. The core journal list is renewed periodically; thus, some journals may become newly included as their quality improve. We did not consider articles published in those journals before the inclusion year in the high-quality subgroup. The articles were assigned to a speci c province or municipal city according to the a liation address of the rst author.

Bibliometric indicators
Several bibliometric indicators were used here to evaluate scienti c activity in the eld of HPC. Production indicates the ability to create knowledge in a given period. It was assessed using the number of publications as well as that number per million inhabitants. We obtained socioeconomic parameters, such as the number of inhabitants and regional gross domestic product (GDP), from the National Bureau of Statistics of China on April 26, 2020 [4]. To evaluate the pattern of growth, we conducted a linear adjustment and exponential adjustment.
The impacts indicate the quality of the articles published by a speci c region. They were measured using the regional impact factor (RIF): that is similar to the impact factor of a journal and could be established each year based on the previous 2-year period. We determined the RIF of a given year as follows: the denominator signi ed the total number of articles published during the 2 years before the index year; the numerator signi ed the number of citations in the index year of the articles included in the denominator. Some regions produce few publications; thus, we introduced the 5-year RIF of 2019 as a more stable indicator for interregional comparison. The 5-year RIF was the same as the RIF, except it used a 5-year time window. To avoid excessive articles with no citations, the calculation of impact factors employed only the citation data of high-quality publications.

Content analysis
The articles in the high-quality subgroup were categorized according to literature types and themes. We regarded original research and case reports as primary literature and reviews, guidelines, and consensus studies as secondary literature. No editorial among our data followed an introduction-method-resultdiscussion format pattern, so we classi ed them as secondary literature. According to the subject and research theme described in the abstract, the two reviewers classi ed each article into the following categories with mutual agreement: We adopted those categories following the advice of the Asia Paci c Hospice Palliative Care Network with some modi cations [22]. Overlap was permitted, and so each article could fall into more than one category. There is an example of literature categorization in the additional le.
We analyzed the trend in research interests using CiteSpace (version 5.6.R3), a JAVA-based software that can detect the word bursts over a period of time [23]. A word burst is a surge in the frequency of the appearance of a word: it is a valuable indicator of active research topics [24]. Owing to language differences, we applied the burst detection to articles in English and Chinese separately.

Statistics
According to the nding of Derek John de Solla Price, scienti c knowledge grows exponentially [25]. We applied a regression analysis to test if the growth of HPC-related publications ful lls Price's law, that is to say, follows a pattern of normal growth. The regression analysis was performed using IBM SPSS Statistics 26. We conducted a linear adjustment of the total publications according to the equation y=41.794x-83871 and an exponential adjustment according to the equation y=4.562E-103e 0.12x . We undertook a similar analysis with the high-quality subgroup according to the following equations: y=11.188x-22474 versus y=1.183E-151e 0.175x . Difference were considered signi cant at a P-value below .05.

Growth of HPC-related publications
We identi ed 3224 documents as HPC-related publications for 2010-2019; among them, 636 were high-quality publications and 255 were in English.
Irrespective of quality, the annual number of publications grew notably during the 10-year period ( Figure 2). According to the regression analysis of total publications, the R 2 was 0.873 when adjusted to an exponential curve; it was 0.784 with a linear adjustment. This nding indicates that exponential growth better explained the pattern of increase, thereby ful lling Price's law [25]. A similar pattern was evident in the high-quality subgroup: the R 2 was 0.925 with an exponential adjustment and 0.830 with a linear adjustment. We also observed a sharp increase in total publications after 2016. A similar, though milder, trend was evident in high-quality publications after 2017. The impacts of scienti c production also rose during the study period.   1 The Chinese Core Journals. 2 International Journals in English.

Regional distribution of the publications
Although the overall growth of HPC-related publications during the 10-year period appears promising, the regional disparity was substantial ( Figure 3A). In decreasing order, the ve most productive regions were Beijing (11.8% of total production in Mainland China), Shanghai (11.1%), Jiangsu (8.2%), Guangdong (7.3%), and Sichuan (5.3%). By contrast, China's northern and western provinces lacked publications. When adjusted by the number of inhabitants, the prominence of Beijing and Shanghai became greater: during the 10-year period, they contributed 17.6 and 14.8 articles per million inhabitants, respectively.
They were followed at a considerable distance by Tianjin (5.5), the city with the rst-ever hospice center in Mainland China (Table 2). There was a strong correlation between annual production and regional economy ( Figure 3B). However, on that basis, Beijing and Shanghai were outliers and performed much better than could be expected. The regional difference in publication impacts is indicated by the 5-year RIF ( Table 2); there was no clear correlation between productivity and impacts (R 2 =0.02, P =0.42).
The names and locations of the top 10 productive institutes appear in Figure 3A: they are all universities and colleges with a liated hospitals. It is noteworthy that four of the 10 institutes were located in Beijing and Shanghai; only one (Sichuan University) was in Western China. We also examined the production of the 41 institutes that provided HPC in the rst ve pilot regions for hospice implementation. Those institutes published 51 articles and had no collaborations with the leading institutes listed in Figure 3.  Table 2 Regional indicators of production and impacts

Themes of the publications
We further analyzed high-quality publications concerning their literature type and content. In this subgroup, 411 of 636 articles were original research and 10 were case reports; thus, primary literature accounted for 66% of the total. Secondary literature comprised of 212 reviews or editorials and three consensuses. The articles were further categorized according to their themes. As shown in Table 3, HPC in cancer patients was the leading topic; that was followed by pain and symptom management, assessment of needs and implementation of HPC, and policy and program advocacy. The topics that received minor attention were bereavement, HPC screening and timing, rehabilitation and nutrition, self-care of the HPC team, and pediatric HPC. We examined trends in research interest in HPC with the word burst detection function of CiteSpace . Figure 4 shows the key words that began to burst from 2015. For example, one Chinese version of HPC (shu-huan-liao-hu) burst between 2015 to 2017, appearing in 100 articles. The terms "negative emotions" and "pain management" constituted the frontier of interest in 2019 and may continue to receive future attention.

Discussion
In this study, we attempted to provide an overall assessment of research and publications related to HPC in Mainland China over a 10-year period. During that time, both the production and impacts of the publications increased; however, regional disparities were conspicuous. There was also a lack of a discussion of some important areas of HPC.

Growing out of infancy
The rst hospice in Mainland China was founded in the municipal city of Tianjin in 1988 [26]. But the adoption of HPC in Mainland China has been slow. Even among health professionals, there is still insu cient knowledge of the subject [27]. However, the present study found that the number of HPC-related publications in 2019 was almost three times that in 2010. The exponential rise of the production and impacts of HPC-related publications is a sign of academic attention and discussion, which are vital for professional education as well as research innovation. In our opinion, this trend will lead HPC in Mainland China out of its infancy.
There are several reasons for the growth. According to the 2010 and 2015 Quality of Death Index by the Economist Intelligence Unit, China is among the countries with the lowest rank [2,3]. All three countries with a lower ranking than China in 2010 surpassed it in 2015. This was widely covered by Chinese media and elicited broad discussion, stimulating health-care workers and policy makers to reexamine the medical system and service. At the same time, a series of new policies prompted the promotion of knowledge about HPC. For example, community care, which burst as a key word in 2015, has been promoted by the government since 2012. The burst detection also showed a promising shift of key words from main concepts to practical issues, such as pain management and negative emotions, which may lead to more clinically relevant research.
Despite its growth, research into HPC in Mainland China remains at an early stage. First, the number of high-quality publications per million inhabitants is meager compared with other parts of Asia [19]. Second, a large number of those publications are secondary literature, which introduces ideas and knowledge but does not explore the unknown. Third, HPC for non-cancer patients is far less discussed than for cancer patients, and end-of-life care or hospice receives greater attention than integrated palliative care in an early stage. Thus, many professionals may be unaware of the latest concept of palliative care [28]. Fourth, there is still no specialized HPC journal in Mainland China. It hinders the development of the academic community of HPC professionals. Accordingly, policies should be designed to support original research, education, and publications in HPC.

Conspicuous regional disparity
Imbalanced development in China has long been recognized-especially between East and Western China [29]. This inequality is evident in many areas, including the economy, human capital, public facilities, and science and innovation. Studies have observed that economic factors are closely linked to a country's production of HPC-related publications [30]. We found that relationship also applies on a provincial scale. When residents in less prosperous regions have limited access to curative medicine [31], an insu cient HPC service will close the door to relief. Thus, the regional disparity in research and publications is both a result and a component of that imbalance, which would produce the Matthew effect and cause inequity in HPC.
China's great cultural richness warrants more research in different regions and populations. Chinese attitudes toward death are heterogeneous. One study among Christians in a northern rural region of China determined that 95% could freely talk about death at home; 85% felt joyful when thinking about death [32]. That nding is against the stereotype of Chinese people. Another study from urban areas showed that older people were more willing to die in hospital than younger ones [33]; that is also against the stereotype of the elderly. Owing to different cultural and social factors, patients' attitudes to HPC varied from hostility to embracing the approach [34,35]. Some regions with fewest publications (such as Tibet, Hainan, Xinjiang, and Inner Mongolia) are home to various ethnic groups. Thus, in a rapidly changing, culture-rich society, we believe that more research is needed to understand local needs and promote culturesensitive care.

Call for collaboration
The study identi ed a lack of collaboration at different levels. From the names of journals publishing HPC-related articles, nurses, general practitioners, and geriatricians are especially active in this eld. However, researchers from other specialties, such as emergency medicine and intensive care, were scarce.
Nurses have done excellent work in HPC research. However, in Mainland China, nurses are not allowed to prescribe medications, which compromises their ability to conduct intervention studies. More doctors from different specialties, nutritionists, physical therapists, psychologists, social workers, and other related professionals should be involved in multidisciplinary HPC research.
At the institutional level, the hospice providers in the rst ve pilot regions had no collaboration with leading universities in their publications. HPC is highly practical, so we strongly recommend collaboration between HPC providers and professional researchers. Future policies and funding should support this collaboration.
International collaboration is also important for HPC practitioners and researchers in Mainland China: they should seize opportunities to communicate with researchers from the rest of the world, especially other Asian regions. Overall, collaboration is the way to make more people heard.

Strengths and limitations
To our knowledge, this is the rst study that provides a whole picture of HPC-related research and publications in Mainland China. Our study has several limitations. First, the terminology for HPC in Mandarin Chinese is not uni ed, and so we included various versions. We may have missed articles that did not mention those terms. Second, it was di cult to determine the actual quality of an article. We used the quality of the publishing journal as a substitute for the quality of the article. However, a high-quality article can be published in a non-core journal [36]. Third, some excellent HPC centers dedicated themselves to delivering care rather than research. Their efforts were largely invisible in our study. Future investigations should aim to determine the factors that hinder or promote research in HPC.

Conclusion
China is a good example of a low-and middle-income country with limited, yet highly demanded, implementation of HPC. The 10-year study period witnessed a considerable increase in HPC-related publications in Mainland China. However, regional disparity in production was conspicuous, and HPC research is at an early stage. Research is essential for the sustainable development of HPC and should be promoted with interregional equity. Policies should be made to encourage collaboration among different disciplines, institutes, and regions. Future research is needed to identify the factors that hinder or promote research in HPC.    Figure 1 Flow chart of the bibliometric study Figure 2 Annual development of production and impacts of publications Figure 3 Regional distribution of publications Note: A, A heatmap of regional production, and locations and ranks of the top-10 productive institutes; B, an example of the correlation between annual regional GDP and number of publications in 2019. The GDP was based on a currency exchange rate of 689.85 Chinese Yuan