Bibliometric Analysis of the Journal of Orthopaedic Research from 1991 To 2018

A chronological survey of papers was conducted to investigate the journal titled Journal of Orthopaedic Research in Science Citation Index Expanded (SCI-EXPANDED) between 1991 and 2018. Performances of countries, institutes, and authors, including total, single, collaborative, first author, and corresponding author publications were analyzed. Citation indicators including the total number of citations since publication to the end of 2018, the number of citations in 2018, total number of citations in the publication year, total number of citations per number of years, and citations per publication were applied to evaluate highly cited articles. Comparison of the articles contributing most to impact factor and highly cited articles were discussed. In addition, journal impact factor contributors and journal impact factor contributing articles were presented.


Introduction
The bibliometric analysis of publications can provide insight to improve the understanding by editors and readers of an academic journal's direction [1]. The Journal of Orthopaedic Research with ISO Abbreviation title J Orthop Res has been indexed in the Science Citation Index Expanded (SCI-EXPANDED) and classified in Web of Science category of orthopedics since 1985. In this category, bibliometric analyses of journals have been reported to investigate journals' development and impact in research fields, including journals Physical Therapy [2,3], Journal of Orthopaedic & Sports Physical Therapy [4], Acta Orthopaedica et Traumatologica Turcica [5], International Orthopaedics [6], Knee [7], Journal of Orthopaedic Trauma [8], and Spine [9]. Recently authorship trends in the Journal of Orthopaedic Research [10] and a bibliometric analysis of classic publications in Web of Science category of orthopedics [11] were also reported. Document types, languages, publication trends, and publications of country, institution, and author were generally revealed to provide basic information about a journal [1,12]. In addition, journals' impact factor, which was created in the early 1960s [13], has become a staple in many types of analyses of a journal's scientific impact [14]. Publication indicators such as total, independent, collaborative, first author, and corresponding author publications [15] as well as citation indicators including the total number citations from Web of Science Core Collection since publication to the end of the most recent year [16,17] and the number of citations in the most recent year [18] were generally applied to evaluate journals [1,19].
In this study, the bibliometric method was employed to obtain an overview of Journal of Orthopaedic Research from 1991 to 2018. General analysis including publication characteristics and trends were presented. Furthermore, relationship among journal's impact factor, its contributors, and highly cited publications were discussed.

Methodology
Documents used in this study were derived from the Science Citation Index Expanded (SCI-EXPANDED) of the Web of Science core Collection, Clarivate Analytics (formerly known as the Thomson Reuters and the Institute for Scientific Information). The keyword phrase "Journal of Orthopaedic Research" was searched as publication name based on SCI-EXPANDED (updated on 20 November 2019). All document information from SCI-EXPANDED was checked and downloaded into Microsoft Excel 2016 [20,21]. Collaboration type of country

Characteristics of document type
The distributions of document types at the Web of Science and their citations per publication (CPP year = TC year /TP) have been analyzed for journals by Ho's group, for example Polish Journal of Environmental Studies [25], Journal of Membrane Science [1], and Revista de Biología Tropical/International Journal of Tropical Biology and Conservation [12]. Journal of Orthopaedic Research published 5,589 documents in 11 Web of Science document types from 1991 to 2018 (Table 1). Article (5,329) was the dominant document type comprising 95% of 5,589 productions with the third ranking of APP (5.8), followed by letters (1.4%), proceedings papers (1.3%), and reviews (1.1%). Among the 11 document types, proceedings papers had the highest CPP 2018 of 50, followed by articles with CPP 2018 of 33, notes (17), and reviews (16). It should be pointed out that CPP 2018 of review is much lower than that of articles. Proceedings papers had the highest number of authors per publication (APP=8.5) followed by corrections (5.9). Only the articles were used for further study.

Characteristics of journal's Impact Factor (IF)
According to the Journal Citation Reports (JCR) in 2018, JCR indexed 9,258 journals with citation references across 178 Web of Science categories in SCI-EXPANDED. The journal impact factor (IF) is defined as all citations to the journal in the current JCR year to items published in the previous two years, divided by the total number of scholarly items (these comprise articles, reviews, and proceedings papers) published in the journal in the previous two years.
The journal impact factor has a simple formula: where IF year is journal impact factor in a specific JCR year, C year-2 : citations from JCR year to items in "year -2", C year-1 : citations from JCR year to items in "year -1", TP year-2 : citable items in "year -2", TP year-1 : citable items in "year -1". The denominator is comprised of three document types such as reviews, articles, and proceedings papers. An item with any other document type is excluded from the denominator.
The top 21 most impact factor contributing articles were listed in Table 2 However, only 14%, 14%, 24%, and 57% of the top 21 most impact factor contributing articles were ranked top 21 in terms of C 0 , total number of citations from Web of Science Core Collection in publication year [15]; TC 2018 , total number of citations from web of science core collection since publication to the end of 2018 [16,17]; Copyright © : Yuh-Shan Ho OPROJ.000632. 6(2).2019 C 2018 , total number of citations from web of science core collection in 2018 [18], and TCPY, total number of citations per number of year, TC 2018 /year [18] respectively. Furthermore, only 47% and 34% of the top 100 impact factor contributing articles were found in the top 100 articles in terms of C 2018 and TC 2018 , respectively. Article entitled "Post-traumatic osteoarthritis: Improved understanding and opportunities for early intervention" by Anderson et al. [26] from USA was not the top high impact factor contributing article (ranked 1 st ) but also the most impact article in the recent year with C 2018 of 35 (ranked 5 th ) in J Orthop Res However the article ranked 40 th in TC 2018 with 243. It was generally accepted that time is needed to accumulate total number of citations for an article. Similarly, article entitled "Distribution of in situ forces in the anterior cruciate ligament in response to rotatory loads" by Gabriel et al. [27] from University of Pittsburgh in USA was not the top high impact factor contributing article (ranked 6 th ) but also the highly cited with TC 2018 of 383 (ranked 11 th ). It can be concluded that the impact factor of journal is used to evaluate a journal's relative importance, especially when compared to others in the same field but not appropriate to be used when compared to individual research performance. It was also reported in web of science that the journal impact factor is a publication-level metric. It does not apply to individual papers or subgroups of papers that appeared in the publication. Additionally, it does not apply to authors of papers, research groups, institutions, or universities.   Figure 1 shows its journal impact factor and ranking within the subject category from 1997 to 2018. Journal impact factor (JIF) slightly increased with fluctuate trend was found. Ranking of J Orthop Res in the Web of Science category of orthopedics was decreased. How quickly recent publications are cited is an important factor that can affect IF [13]. In general, citations per publication for articles in a journal would have a sharp increase Copyright © : Yuh-Shan Ho OPROJ.000632. 6(2).2019 after publication and would reach a peak in a specific year. Figure  2 shows citations per publication for each year of article [28]. The peak year of citations per publication was found to be in the 5 th full year since its publication year. The peak year could be different, for example, the 7 th year for Revista de Biología Tropical [12]. A special case was reported that it does not show a peak, unlike previous findings, but shows an increasing trend without a peak after 4 th and 9 th year for Journal of Membrane Science [1] and Polish Journal of Environmental Studies [25] respectively. Since IF only considers citations within two years after publication, the IF of J Orthop Res would be higher if it was calculated after a couple of years. It was pointed that IF is not an unbiased criterion for all journals, since peak year citations per publication of each journal can be different from each other [25].

Trends of number of articles and citations per publication
Trends of number of articles shows a journal's development and citations per publication gives information about impact of a journal in research world. Trends of the number of annual articles (TP) and their citations per publication (CPP 2018 = TC 2018 /TP) has been proposed to discover the development of a journal [1,12]. Of In 1991, 101 articles had the highest CPP 2018 of 91, which can be attributed to the top ten most frequently cited articles by Caplan [29] with TC 2018 of 2,243 (ranked 1 st ), [30] with TC 2018 of 607 (ranked 2 nd ); and [31] with TC 2018 of 497 (ranked 7 th ).

Characteristics of countries, institutions, and authors
Excluding 16 articles without any author affiliation information on SCI-EXPANDED, the remaining 5,313 articles originated from 57 countries. Among those articles, 4,295 (81% of 5,313 articles) were single country articles, while 1,018 (19%) were internationally collaborative articles. Table 3 shows the top 10 most contributing countries with six publication indicators such as number total of articles (TP), country independent articles (IP), internationally collaborative articles (CP), first author articles (FP), corresponding author articles (RP), single author articles (SP), and citations per publication (CPP 2018 =TC 2018 /TP) [1,25]. The advantage of using CPP 2018 and TC 2018 is that it is invariant in comparison with the index of citations from the Web of Science Core Collection which has to be updated from time to time [24]. Four European countries, three Asian countries, two American countries, and one Oceania country were ranked on the top 10 of total articles. The most productive African country was Egypt with five articles ranked 37 th . USA    Of the total 5,313 articles with affiliation information in SCI-EXPANDED, 1,983 articles (37% of 5,313 articles) came from independent institutions with CPP 2018 of 35; 3,330 articles (63%) from inter-institutional collaborations with CPP 2018 of 32; included 1,018 articles (31% of 3,330 inter-institutionally collaborative articles) from international collaborations with CPP 2018 of 31 and 2,312 articles (69%) from national collaborations with CPP 2018 of 33.

Percentage of the inter-institutional collaboration rate of J Orthop
Res (63%) was found much higher than that of J Membr Sci (38%) [1] and Pol J Environ Stud (31%) [25]. Table 4  articles in J Orthop Res. It has been generally accepted that the first author and the corresponding author are the most significant author position [32,33]. Three bibliometric indicators such as total articles (TP), first author articles (FP), and corresponding author articles (RP) were applied for an analysis of authors' publications in a journal [18,34]. Table 5  Woo were journal impact factor contributors (JIF contributors) who published high journal impact factor contributing articles with TC year-1 + TC year-2 of 48 for articles by Murray et al. [35] and Gabriel et al. [36] respectively. Bias is always considered in authorship analysis because two or more authors may have the same name, or authors used different names in their publications (e.g. name changes due to marriage) [19].  Ho [18,22,37] proposed the Y-index which is related to the numbers of first author publications (FP) and corresponding author publications (RP). The Y-index with two parameters (j,h), assess both the publication potential and characteristics of contribution as a single index. The Y-index is defined as [22,37]: j indicates publication potential. It was calculated by using numbers of first authored articles and corresponding authored articles as the Eq. (1). When one had larger j, it means its Y-index located far away from origin of the polar coordinates. It indicates that one published more articles as "important author" [37]. h is a publication characteristic constant, that introduces the distribution of the numbers of the corresponding authored articles and the first authored articles. h could be calculated by using Eq. (2). When the number of the first authored articles and the number of the corresponding authored articles are the same, Y-index is located in the 45-degree (0.7854 rad) line with the same h of 0.7854. When j>0.7854, it means one published more corresponding author articles, and when j<0.7854, it means one published more first author articles. When h=0, j=number of first author articles, and h=π/2, j=number of corresponding author articles.
Of the 5,329 articles published in J Orthop Res, 4,889 articles (92% of 5,329 articles) with both first author and corresponding author's names in the SCI-EXPANDED were extensively analyzed by using the Y-index. A total of 4,889 articles were contributed by 15,911 authors. Only 1,803 authors (11% of the 15,911 authors) had both first-and corresponding-author articles in J Orthop Res In particular, 929 authors (5.8% of the 15,911 authors) had only corresponding-author articles (h=π/2); 334 (2.1%) authors had more corresponding-author articles than first-author articles (π/2>h>0.7854); 1,343 (8.4%) authors had the equivalent numbers of first-author and corresponding-author articles (h=0.7854); 126 (0.79%) authors had more first-author articles than correspondingauthor articles (0.7854>h>0); and 11,162 (70%) authors had neither first-nor corresponding-author articles with Y-index of (0, 0). Figure 4 shows distribution of the Y-index (j, h) of the top 18 contributing authors with j≥17 (j Cos h and j Sin h are chosen as the x and y coordinate axes). Each point has a coordinate (j, h) that could symbolize a single-author or multiple authors. j is contributed constant, an author with a higher j indicates more articles as first A bias would appear in authorship analysis because two or more authors may have the same name, or authors used different names in their publications [19].

Highly cited articles
The total citation count was obtained from web of science core collection, and this shows the total number of times that an article was cited by the journals listed in the Web of Science Core Collection database. The hotspots in a research field might be reflected by highly cited articles [38]. Highly cited publications in a journal were also investigated, for example JAMA-Journal of the American Medical Association [14], Water Research [19] and Polish Journal of Environmental Studies [25]. Articles with TC 2018 ≥100, were generally called highly cited articles [37,39]. In J Orthop Res, 333 of 5,329 articles were highly cited articles with TC 2018 ≥100. The top 10 highly cited articles were listed in Table 6. All the top 10 highly cited articles were published in the 1990s. The earliest highly cited was published in 1991 by Caplan [29], Schipplein [30] & Athanasiou et al. [31]. The latest highly cited articles were published in 1999 by Mckellop et al. [40] & Vunjak-Novakovic et al. [41]. Three highly cited articles with TC 2018 ≥100 such as articles by Caplan [29], Markolf et al. [42] & Schipplein [30] in J Orthop Res were also ranked top ten in C 2018 . The 10 highly cited articles were published by 45 highly cited authors from 18 institutions. Three of the top ten articles were published by authors from case Western Reserve University in USA. These top ten articles were published by USA while one was international collaboration with Switzerland. Caplan AI [29,43] and AJ Grodzinsky [41,44] published two the top cited articles respectively (Table 6). Citation histories of the top ten most frequently cited articles in J Orthop Res Were shown in Figure  5. The highly cited article about mesenchymal stem cells by Caplan [29] had different citation development trend. There is a sharply increased trend after publication year of 1991 for two decades to reach a plateau in recent years. Highly cited articles would not be always with high citations. Article by Buschmann et al. [44] with TC 2018 of 395 had low citation in the most recent year of 2018 with C 2018 of 7.
C 2018 : total number of citations from web of science core collection in 2018.

Conclusion
In total of 5,589 publications in 11 Web of Science document types were published in J Orthop Res from 1991 to 2018. Article was the dominant document type and had a CPP 2018 of 33 but review had only half of CPP 2018 with 16. The peak year of citations per publication was found to be in the 5 th full year since its publication year. USA published the most articles and had higher citations per publication. Harvard University contributed the most articles in J Orthop Res. Results of citation analysis may have provided some initial evidences that highly cited articles in a journal might not contribute a lot to the journal's impact factor. Articles related to osteoarthritis, biomechanics, cartilage, knee, and fracture healing were the most popular topics while intervertebral disc, tissue engineering, and apoptosis were new focuses in the journal.