Bibliometric Study of the Comorbidity of Pain and Depression Research

Background Comorbid pain and depression occur with high prevalence in clinical observations, and published academic journals about them have been increasing in number over time. However, few studies used the bibliometric method to analyze the general aspects of scientific researches on the comorbidity of pain and depression. The aim of this study is to systematically provide global scientific research in the comorbidity of pain and depression from 1980 to 2018. Methods The published papers were searched between 1980 and 2018 in Web of Science. Publications related to comorbid pain and depression research were included. The language was restricted to English, and no species limitations were specified. Results A total of 2,519 papers met the inclusion criteria in our study. The results revealed that the publications had a significant growth over time in the comorbidity of pain and depression research (P < 0.001) by linear regression analyses. The United States had the largest number of publications and citations and the highest value of H-index. According to subject categories of Web of Science, research areas of the 2,519 papers mainly focused on clinical neurology (28.78%), neurosciences (22.9%), and psychiatry (22.23%). In accordance with types of pain, headache (19.09%) was the most popular topic in the included papers on comorbid pain and depression research. Conclusions The findings provide useful information for pain and depression researchers to detect new areas related to collaborators, cooperative institutions, popular topics, and research frontiers.


Introduction
Pain is the most pervasive physical symptom in primary care, which affects physical and emotive functions, deteriorates the quality of life, and reduces the ability to work [1][2][3][4]. The prevalence of chronic pain has been estimated to be as high as 40% in the United States [5]. The most common types of pain syndromes [6][7][8] include headache, low back pain, fibromyalgia, neuropathic pain, arthritis, cancer, and postsurgical pain. The total cost of pain has been calculated to be as high as 3% of the gross domestic product of European countries, which is larger than the cost of cancer or heart disease [9]. Depression is the most common psychological symptom and is a major public health concern in primary care [10,11]. The prevalence of depressive symptoms or depression among medical students from 43 countries has been reported to be 27.2% [12], and the lifetime prevalence of depression in 30 countries has been calculated to be 10.8% [13]. In Japan, the total cost of depression among adults was estimated to be ¥2.0 trillion in 2005 [14].
Comorbid pain and depression occur with high prevalence in clinical observations, as high as 60% [4]. A number of studies have reported that antidepressant drugs have been used to treat pain since the 1960s [4,15]. Pain intensity is aggravated by increasing levels of depression [16]. Although the neurobiological mechanisms of the interaction between pain and depression are unclear, human neuroimaging studies indicate that depression can influence changes in multiple brain regions including those related to pain processing and perception [17,18].
While research on comorbid pain and depression is widely available worldwide, relatively few studies use bibliometric methods to analyze the general aspects of this research topic. Bibliometric studies are extensively used to determine trends of scientific research and involve quantitative analysis of published studies [19,20]. This type of study could offer readers quantitative information on distribution by country, institution, author, and journal in a specific field. Over the last 10 years, bibliometric studies have been used to analyze scientific research in different areas, such as long noncoding RNA [21], paediatric pain [22], cancer rehabilitation [23], childhood immunization [24], eye disease [25], and drug delivery [26], worldwide.
To address the shortage of quantitative analysis of comorbid pain and depression research, the aim of this study is to systematically provide global scientific research in this field from 1980 to 2018. We used CiteSpace V (Drexel University, Philadelphia, United States) to perform a bibliometric study in Web of Science Core Collection, which is a tool frequently used to assess the trends in the global scientific research [27]. In this study, global scientific research on comorbid pain and depression includes the number of published papers, types of pain, distribution and collaborations between authors/institutions/countries, a citation burst analysis of keywords, cocitation analysis of authors and references.  Index Expanded (SCI-Expanded) of Web of Science. The following terms were searched: Title=(pain * or headache * or migraine * or head-ache * or "head ache * " or cephalalgi * or "abdominal ache * " or fibromyalg * or "tummy ache * " or "stomach ache * " or "belly ache * " or ear-ache * or earache * or tooth-ache * or toothache * or odontalgi * or neuralgi * or cervicodyn * or analg * or nocicept * or hyperalg * or hypoalg * or radiculalg * or colic or arthralg * or causalg * or maldyn * or eudyn * or ophthalmodyn * or cephalalg * or dysmenorrh * or sciatic * or otalg * or brachialg * ) and Title=(depression or depressions or depressed or despondent or gloomy or depressive).

Inclusion Criteria.
We included articles, reviews, letters, and editorial materials published in different academic journals. Meeting abstracts, conference presentations, book reviews, news items, and corrections were excluded. Publications related to comorbid pain and depression research were included. The language was restricted to English, and no species limitations were specified. The H-index is recognized as a measurement of the papers' citation frequency for academic journals or researchers. For example, the H-index is that an academic journal or researcher has H published papers which have at least H cita-tion times per paper (If the H-index of one researcher is 10, that means that he had published at least 10 papers and each of these 10 papers had been cited at least 10 times.) [28]. Furthermore, these included studies that were classified into the following categories: (1) single-or multiple-authored publications (authors ≥ 2); (2) Web of Science subject category. We would select and assess the top 20 subject categories of Web of Science categories according to the included studies. Every paper or journal is assigned to at least one Web of Science category in Web of Science core collection. Thus, the sum of publications could be more than 2,519. Research areas were classified into five broad categories in Web of Science categories, and the five broad categories included 225 subject categories [29], such as anesthesiology, neurosciences, and psychiatry; (3) types of pain. We would choose and analyze the top ten types of pain (such as headache and low back pain) according to the included studies. Because some papers exceeded one type of pain, the sum of publications could be more than 2,519. For example, the paper by Gesztelyi and Bereczki included headaches and low back pain [30].  subject categories of Web of Science, and types of pain. To assess whether the percentage statistically decreased or increased over time, the linear regression analyses were conducted with every respective category as the dependent variable and the year as the independent variable. The statistical analyses were performed with IBM SPSS Statistics 22.0 software (SPSS Inc., Chicago, USA). A P value less than 0.05 was considered as statistical significance.

Publication Outputs and Growth Trends.
A total of 2,519 papers met the inclusion criteria; 1,264 papers (meeting abstracts, proceedings papers, notes, corrections, book reviews, and news items) and 151 non-English papers were excluded (Supplementary Figure 1). The overall trend of publications increased from five publications in 1980 to 179 publications in 2018 (Figure 1(a)). The results of linear regression analyses showed that the percentages had a significant increase over time in the recent forty years (t = 11:37, P < 0:001). All papers related to comorbid pain and depression research were cited 74,746 times (29.67 times per year, H-index 121), and the percentages had a significant increase over time (t = 12:47, P < 0:001) (Figure 1(b)). Among the eight 5 years (1980-1984, 1985-1989, 1990-1994, 1995-1999, 2000-2004, 2005-2009 Table 1). Among the top 20 journals by the number of publications (Table 1) A dual-map overlay of journals is presented in Figure 3. In the dual-map, the map of the citing journals is on the left and the map of the cited journals is on the right. The disciplines included in the journals are shown as labels in the dual map. The lines are considered citation connections, beginning from the citing journals to the cited journals. The dual-map overlay shows that the majority of papers were published in neurology, sports, and ophthalmology journals and these journals mostly cited journals from psychology, education, and social areas. 3.4. Types of Pain. As shown in Figure 5, according to the number of publications, headache (19.09%) was the most popular topic in the included papers on comorbid pain and depression research, followed by low back pain, animal models of pain, and fibromyalgia. Among the top 10 types    In addition, the results showed that the percentages of publication count per year had a significant increase over time (P < 0:01) in headache, low back pain, animal models of pain, fibromyalgia, neuropathic pain, arthritis, cancer pain, postsurgical pain, visceral pain, and neck pain.

Distribution by Countries and
Institutions. The 2,519 papers on comorbid pain and depression research were contributed by 75 countries/territories (Supplementary Table 2). Figure 6 shows the top 10 countries/territories by quantity of papers, showed as sums of paper fractions. The United States had the largest number of publications (1,105), citations (44,490), and open access papers (342) and the highest value of H-index (102). The Netherlands had the largest number of citations per paper (40.37). In relation to the top 10 countries that contributed to comorbid pain and depression research, USA led the first research echelon, followed by the England (178), Canada (156), and Germany (139). There were extensive collaborations between countries/territories (Figure 7(a)). In accordance with the number of published papers, the overview of all countries was presented in world map (Figure 8).
A total of 1,884 institutions contributed to the publications on comorbid pain and depression research (Supplementary Table 3  3.6. Distribution by Authors. A total of 8,771 authors contributed to the total number of publications. The network map in Figure 9 outlines the cooperation between authors. Amongst the authors who had the most publications (   significantly increased and that with a single author significantly decreased (t = 11:76, P < 0:001).

Analysis of References.
Analysis of references was regarded as an important indicator in bibliometric study. The scientific relevance of the published papers was presented in the cocitation map of references ( Figure 11). The modularity Q score was 0.8897 (higher than 0.5), which indicated that the network was reasonably distributed to loosely coupled clusters. All clusters were traced by index terms extracted from the references. The "rehabilitation outcome" was labeled as the largest cluster #0, followed by "home care" as the second largest cluster #1, "pain comorbidity" as the third largest cluster #2, and "depressive disease." As shown in Figure 11, the top 23 clusters were showed in a timeline view.
3.8. Analysis of Keywords. The keywords of the 2,519 published papers were extracted by CiteSpace V. Figure 12 reveals the top 47 keywords with the strongest citation bursts. Keywords with the strongest citation bursts beginning in 1991 were as follows: "disease," "illness," "chronic pain," and "inventory.  Table 3 showed the top ten papers on comorbid pain and depression research with the most citation frequency. The top 10 papers contributed to 7.81% (5,831 citations) of the total number of citations. The article by Bair et al. [31] published in 2003 was the most cited (1,442 citations) paper, and the article is entitled "Depression and pain  comorbidity-a literature review." published in Archives of Internal Medicine. Among the top 10 papers, five [31][32][33][34][35] were published in journals with impact factor ≥ 10 (Archives of Internal Medicine, Brain, Psychological Bulletin and Neuron), two [36,37] in journals with 5 ≤ impact factor < 10 (Neurology, Arthritis & Rheumatology), two [38,39] in journals with 3 ≤ impact factor < 5 (Clinical Journal of Pain, Journal of Pain), and one [40] journal with 1 ≤ impact factor < 3 (The Journal of Nervous and Mental Disease). In accordance with the number of publications on comorbid pain and depression research, the top 20 journals contributed to 29.37% (740 publications) of the total number of publications. Pain contributed to the most publications on comorbid pain and depression research (5.39%) followed by Clinical Journal of Pain (2.26%), Headache (2.26%), and Cephalalgia (1.98%). According to the impact factors of the journals in the Journal Citation Reports (2017 edition), the impact factors were less than 10 among the top 20 journals. In the top 20 journals, the journals with 2 ≤ impact factor < 3 contribute to 40% of the top 20 jour-nals, the journals with 3 ≤ impact factor < 5 contribute to 50%, and the journals with 5 ≤ impact factor < 10 contribute to 10%. However, according to journal IF quartile of Web of Science, 55% of the top 20 journals were Q1 and 40% of the top 20 journals were Q2.

Discussion
In accordance with the number of publications, USA led the first research echelon (1,105), followed by England (178), Canada (156), and Germany (139). And the top 10 countries were from five European countries, three Asia-Pacific countries, and two American countries. As demonstrated in Figure 7, it had an extensive collaboration between countries. A total of 1,884 institutions contributed to the publications on comorbid pain and depression research. According to the number of publications, nine institutions were from USA and one from England. Compared with the cooperation of countries, the institutions were not significant.  cocitation map of references, the "rehabilitation outcome" was labeled as the largest cluster #0, followed by "home care" as the second largest cluster #1, "pain comorbidity" as the third largest cluster #2, and "depressive disease." According to the keywords with strongest citation bursts, the results found that the top 47 keywords that began in 1991 were as follows: "disease," "illness," "chronic pain," and "inventory." And the top 47 keywords by the end of 2018 were as follows:     1980-1984 1985-1989 1990-1994 1995-1999     and depression but also provided a potential therapeutic approach.

Strengths and Limitations.
This study is the first bibliometric analysis to assess the trends of the comorbid pain and depression research from SCI-Expanded of Web of Science in the recent forty years. Furthermore, to gain rich data, it was not restricted to one academic journal. The 2,519 papers on comorbid pain and depression research were published in 251 different academic journals. Moreover, bibliometric analysis of our study not only covered the number of publications, citations, journals, and collaborations between countries/institutions/authors but also included analysis of keywords, cocitation analysis on references and authors, Web of Science subject categories, and types of pain. This bibliometric study has some limitations. The electronic database is limited to SCI-Expanded of Web of Science, and other electronic databases are not searched and analyzed, for example, PubMed, Embase, Scopus, Cochrane Library. Furthermore, the non-English papers were excluded. Most included papers use English in this study; however, the limitation may induce a publication bias. The last limitation is that influential publications were not cited with high citation frequency, since some potential influential papers were published recently, which could be not cited with frequent times.

Conclusions
This study provides historical insights into the trends of comorbidity of pain and depression research. The number of published papers significantly increased over the last 40 years, and the overall trend of publications increased from 5 publications in 1980 to 179 publications in 2018. Although this study presents several limitations, it adequately exposes the trends of comorbidity of pain and depression research. Based on the type of pain, the research topics of the 2,519 papers included in this work mainly focused on headache, low back pain, and fibromyalgia. The results of our study could provide useful information for pain and depression researchers, funding agencies, and policy makers.

Data Availability
All research data used to support the findings of this study are included within the article and the supplementary information file.