Anesth Analg 2008; 106:1524-1533
© 2008 International Anesthesia Research Society
doi: 10.1213/ane.0b013e31816d18a1
GENERAL ARTICLE
Active Research Fields in Anesthesia: A Document Co-Citation Analysis of the Anesthetic Literature
Milan P. Jankovic, MD,
Mark Kaufmann, MD, and
Christoph H. Kindler, MD
Department of Anesthesia, University Hospital Basel, Basel, Switzerland.
Address correspondence and reprint requests to Christoph H. Kindler, MD, Professor and Chairman, Department of Anesthesia and Perioperative Medicine, Kantonsspital Aarau AG, CH-5001 Aarau, Switzerland. Address e-mail to christoph.kindler{at}ksa.ch.
 |
Abstract
|
|---|
BACKGROUND: The expansion of science has resulted in an increased information flow and in an exponentially growing number of connections between knowledge in different research fields. In this study, we used methods of scientometric analysis to obtain a conceptual network that forms the structure of active scientific research fields in anesthesia.
METHODS: We extracted from the Web of Science® (Institute for Scientific Information) all original articles (n = 3275) including their references (n = 79,972) that appeared in 2003 in all 23 journals listed in the Institute for Scientific Information Journal Citation Reports® under the subject heading "Anesthesiology." After identification of highly cited references (
5), pairs of co-cited references were created and grouped into uniformly structured clusters of documents using a single linkage and variable level clustering method. In addition, for each such cluster of documents, we identified corresponding front papers published in 2003, each of which co-cited at least two documents of the cluster core. Active anesthetic research fields were then named by examining the titles of the documents in both the established clusters and in their corresponding front papers. These research fields were sorted according to the proportion of recent documents in their cluster core (immediacy index) and were further analyzed.
RESULTS: Forty-six current anesthetic research fields were identified. The research field named "ProSeal laryngeal mask airway" showed the highest immediacy index (100%) whereas the research fields "Experimental models of neuropathic pain" and "Volatile anesthetic-induced cardioprotection" exhibited the highest level of co-citation strength (level 9). The research field with the largest cluster core, containing 12 homogeneous papers, was "Postoperative nausea and vomiting." The journal Anesthesia & Analgesia published most front papers while Anesthesiology published most of the fundamental documents used as references in the front papers.
CONCLUSIONS: Using co-citation analysis, we identified distinct homogenous clusters of highly cited documents representing 46 active current anesthetic research fields and determined multiple nets of knowledge among them.
|
Introduction
|
|---|
In the last century, there has been an explosive growth in science as reflected by the vast increase in research information. Internet access to indexing services was introduced in the 1970s and today full-text journal articles are routinely provided online. Readers can use these services to instantly access journal contents and to browse current and previous literature. The Institute for Scientific Information (ISI) is the main organization collecting research information for further analysis.1 Over the past four decades, the ISI cataloged scholarly publications together with their reference lists and developed software for information retrieval and analysis.2 Scientometrics is the science of such measuring and analyzing science or, in other words, the study of the quantitative aspects of science as a discipline.3,4 In practice, scientometrics is often done using bibliometrics, a measure of (scientific) publications.5 Vassily V. Nalimov coined the term scientometrics ("naukometriya") in 1969.6 Modern scientometric methods are based primarily on the work of Derek J. de Solla Price and Eugene Garfield.7,8 The importance of such scientometric methods is further illustrated by the search engine Google that uses co-citation algorithms to search the Internet.9 During the first half of the 20th century, the pioneers of bibliometric research described laws characterizing scholarly productivity and communication, and expressed these relationships with formulas as they were used in the natural sciences.10 The introduction of the Science Citation Index (SCI)8 allowed the unrestricted application of methods such as bibliographic coupling (documents are said to be biographically coupled if they share one or more bibliographic references) and co-citation analysis (two documents are said to be co-cited if they appear simultaneously in the reference list of a subsequent third document) to produce a clearer and more objective picture of the enormous, worldwide scientific activity. New publications are tightly connected to a small, selected part of the already existing scientific literature through their citations.7 The intellectual model of citation indexing is simple; the value of previous work is determined by the authors who cite it. Thus, the number of citations can be seen as a direct measure of the resonance or impact that a publication has had on the scientific community.11 In addition to such pure citation counting, a more meaningful approach to literature analysis uses document coupling or co-citation analysis12 that allows for the identification of active research fields, follows their development, and forecasts their future evolution.13–15
The scientometric analysis presented in this article allowed us to gain insight into the structure of existing and emerging anesthetic research fields by identifying and mapping interpretable and homogenous clusters of references from the various anesthetic research directions. Mapping of research fields is useful for the subject matter for established scientists and students/trainees alike. For students and trainees, mapping can provide an entry point into a domain, a means of gaining knowledge on both the macro and micro levels. For the scientist, mapping can help to validate perceptions and provide a means to quickly investigate trends and new information. Yet, even the expert can be surprised by developments peripheral to his/her perceptions. Mapping and interactive exploration provide context for such surprises; the unsuspected proximity of another area of research to his or her own area might suggest alternate routes of analysis that can yield new understandings. Co-citation analysis of documents, therefore, can provide inside knowledge of a field's structure, its dynamics, and potential paradigm developments.16,17 While those who evaluate research outcome require both quantitative and qualitative indicators of performance, others, such as research planners, investigators, policymakers, editors, and publishers, need to keep track of new developments in their own field and in related disciplines. This task is increasingly difficult because of the rapid accumulation of new knowledge. What is needed is a global view of research activities, a view that highlights the most prominent features of the ever-changing scientific landscape, one that reveals fast-moving or "hot" fields and newly emerging areas of a specialty.18
Information derived from defining and mapping active current research fields in anesthesia may be of interest to anesthesiologists, basic scientists, physicians, industry development managers, reviewers, editors, and policy makers from universities and funding institutions. The aim of the present study was to define the active research fields in anesthesia by implementing a document co-citation cluster analysis.
|
METHODS
|
|---|
To identify distinct homogenous clusters of highly cited documents representing active current research fields in anesthesia, we used all international anesthetic journals listed under the subject category "Anesthesiology," according to the ISI Journal Citation Reports® (JCR), as data sources for our analysis. Using the database of the SCI-EXPANDED (1945 to present), 3275 original articles and their 79,972 references published in these journals from 1 January 2003 until 31 December 2003 were identified and downloaded. Our search was restricted to "ARTICLES" in the field "Document Type" of the database SCI-EXPANDED and, therefore, editorials, reviews, letters, news items, and meeting abstracts are not included in the article counts. However, such items could appear later in the analysis in the reference lists of downloaded articles. The method used in our study has been described in detail and is called a document co-citation cluster analysis.12,19,20 The main steps are briefly summarized in Figure 1. When one author cites another author, a relationship is established. To identify important investigation trends and especially to select research fields with little initial impact, it is necessary to choose a citation frequency threshold as low as possible to avoid "noise," defined as the proportion of irrelevant (i.e., seldom cited) documents. A chosen citation threshold of five citations (i.e., documents that were cited 5 times) may separate the influential papers by maintaining the diversity of contemporary research themes. This threshold reduced our dataset to 2075 original articles containing 8243 highly cited documents (Fig. 1). A direct connection between articles and related references was maintained throughout our analysis. If two earlier documents were jointly cited by a later document, co-citation occurred and an intellectual relationship between both cited documents was ascertained, even though they did not directly cite each other. If two documents appear together in the reference list of many succeeding documents, they are strongly co-cited and related in subject matter carrying special patterns of information.
View larger version (25K):
[in this window]
[in a new window]
|
Figure 1. Data sources and basic steps of the co-citation analysis are chronologically presented combined with numerical results. Levels 1–9 refer to the various co-citation strength. Note that no uniform clusters of references were identified at level eight. ISI = Institute for Scientific Information.
|
|
For this reason, the next steps after the identification of highly cited documents were to create pairs of these documents and then to select frequently used pairs for further operationalization. For each article containing at least two or more highly cited documents (n), 1/2 x n x (n – 1) document pairs were generated (Fig. 1). Thus, for example, for the single paper by Struys et al.21 containing the highest number of highly cited documents in its reference list (n = 27), 351 pairs were created. In this way, 24,359 pairs of references (with repetition) were obtained for 1486 anesthetic articles published in 2003 and then reduced to 15,582 distinct pairs without repetition (Fig. 1). Only those pairs co-cited 3 times were further analyzed according to their co-citation strength (levels 3–9). Starting at level 3, co-cited documents were organized into clusters that have at least one document in common. This clustering algorithm, called the single-linkage method, was again used at higher levels of aggregation either to split or to concentrate larger, incoherent clusters (of previous levels) to smaller, homogeneous, and interpretable groups of documents. In other words, coherent and hence definitive clusters could appear at different levels of co-citation strength (Fig. 2). At increasingly higher levels, the degree to which this variable level clustering process will be continued depends on the chosen stopping rules. Such rules stop the clustering process when a desired structure of documents for each cluster is achieved. Stopping rules can be quantitative, qualitative, or both. Our stopping rules were initial size before further separation (5 documents) and homogeneity.14,15 Although a single highly co-cited document pair could identify per se a small research front, only a core of several analogous and interrelated documents can completely describe and represent research topics or specialties and, thus, reflect current research activity. Increasing the number of interrelated documents in a cluster will therefore broaden its scope. By analyzing clusters containing 5 documents, we selectively identified the most active anesthetic fields, filtering out small or narrowly focused fields, which may or may not be recognized as important in the future.
View larger version (34K):
[in this window]
[in a new window]
|
Figure 2. The path of transformation and the composition of clusters of references including the number of documents at different co-citation levels (levels 3–9) during the document co-citation analysis are shown. White ellipses and circles are clusters still under analysis because they are not yet homogeneous whereas shaded ellipses and circles represent final, homogeneous clusters. Dotted ellipses indicate clusters consisting of two closely related research fields. The concentrating process of the clusters through raising the co-citation level is shown by solid arrows. The fragmentation of clusters that separate into two related research fields is represented by dotted arrows. The italic numbers in the ellipses and circles refer to the number of documents in the clusters. The bold numbers next to the ellipses and circles refer to the rank number of the research fields shown in Table 1.
|
|
We emphasize that our study followed a single-blinded technique; from the beginning we were only aware of the full titles of the source articles published in 2003, while their references existed only as a bibliographic description sufficient for paper identification. As examples, we provide three retrieved reference items of the above-mentioned article by Struys et al.21:
- GLASS-PS-1997-ANESTHESIOLOGY-V86-P836 (cited 36 times)
- MINTO-CF-1997-ANESTHESIOLOGY-V86-P10 (cited 18 times)
- STRUYS-MMRF-2002-ANESTHESIOLOGY-V96-P803 (cited 16 times)
These descriptors were then converted into documents with clearly recognizable information with respect to already formed but yet unpredictable clusters of references. We used ISI Web of Science and PubMed as the sources for this document retrieving step.
By counting and clustering such highly cited related documents at variable co-citation strength (levels 3–9), we finally identified clearly separated and homogenous cluster cores of references (Figs. 1 and 2). Next, we selected for each cluster of references corresponding articles published in 2003 (i.e., front papers) that contained at least two or more documents of the respective cluster cores in their reference lists. These front papers represented the research fronts that provide the latest data and specifications of that topic. As a next step, we generated so-called research fields that are comprised of both the cluster core documents and corresponding front papers and gave each research field a name after our analysis of the titles of its cluster core documents and front papers (Table 1). Since each research field consists of a set of cited documents of varying age, an immediacy measure can be devised to reflect how recent its core literature is. The percentage of core documents published within the last 3 yr before being cited in the 2003 source papers (immediacy index) and the average age of documents in the cluster core (i.e., mean years since publication of the cluster core documents before being cited in the 2003 source papers) are two such previously described measures.15,17 Finally, we further analyzed the emerging research fields for their characteristics such as the immediacy index, average age of core documents, and the publishing journal.
|
RESULTS
|
|---|
We extracted 3275 original anesthetic articles published in 2003 from the ISI Web of Science including their 79,972 references. A total of 59,278 distinct documents were cited one or more times (average = 1.35 citations) by the original source articles (Fig. 1). Only a small proportion of these 59,278 documents were cited 5 times (1142 distinct documents or 1.9%; average = 7.22 citations). This small group of documents, regarded as useful by many authors for their scientific research, were cited by 2075 (63%) source articles, each citing on average four such highly cited documents. Almost all of these 1142 highly cited documents were primarily cited in different combinations; only three of them were exclusively cited individually thus staying unpaired throughout the analysis. From 1486 source articles, each of which co-cites 2 of the 1139 highly cited single documents in different combinations, 24,359 document pairs were retrieved, resulting in 15,582 individual, distinct pairs (at the co-citation level 1) (Fig. 1). Some document pairs were used more often (average = 1.56 citations) and 8173 pairs were co-cited 3 times (average = 3.8) resulting in 2136 distinct document pairs (at the co-citation level 3) consisting of 865 single documents (Fig. 1). Finally, these 2136 document pairs were used as the basis for our subsequent clustering, i.e., the grouping procedure, which aggregates co-cited pairs that have at least one document in common (single linkage clustering method). This procedure at first led to 63 clusters ranging in size from 2 to 428 documents. Small clusters containing <5 documents were eliminated, uniform clusters selected, and the remaining inhomogeneous clusters were further reduced in size or broken down into uniform clusters at a higher level of co-citation strength. Ultimately, we obtained 46 uniformly structured clusters of references at a variable co-citation strength (levels 3–9) (Figs. 1 and 2). Thus, the initial collection of 1139 highly cited distinct documents (level 1) was continuously reduced through the variable level clustering process to 284 cluster core documents composing the 46 final clusters. Together with 639 selected corresponding front papers each of which co-cited at least two of the cluster core documents, they represented the current and active anesthetic research fields (Table 1). These research fields were then ranked according to the percentage of their recently cited cluster core documents, i.e., according to their immediacy index.
The appearance and formation of the research fields as well as their size, linkage, and mutual proximity at six hierarchical levels of clustering are shown in detail in Figure 2. At each level, clusters are only composed of pairs admitted to this specific level, i.e., level 3 consists of pairs co-cited 3 or more times and level 4 only consists of document pairs co-cited 4 times or more. Thus, by raising the level of co-citation strength, documents initially linked to a number of different document groups will disappear, thereby separating nonuniform larger clusters into several smaller research areas (such as the large cluster containing 428 documents at level 3, Fig. 2). The finding that some areas only split at higher levels into two new subfields isolated from each other, suggests their close proximity (dotted ellipses and dotted arrows, Fig. 2). However, proximity does not necessarily mean similarity, but rather represents a close relationship between research areas. In this way, eight pairs of related research fields were detected (Table 1 and Fig. 2).
Using both the numerical data (Table 1) and the visual presentation (Fig. 2), it was possible to gain insight into the dynamics of the relationships between associated research fields. Starting at co-citation level 3, we identified, after elimination of unconnected document pairs and small clusters with <5 documents, 27 clusters ranging in size from 5 to 428 documents at that level. Of these, 12 clusters were already totally homogeneous, and therefore, were considered as final clusters containing 5–10 documents (Fig. 2, level 3, shaded ellipses or circles). The clustering at co-citation level 4 provided us with the most final homogenous clusters (n = 14).
In Figure 3, we show a detailed analysis of the research field with the highest immediacy index, i.e., the research field named "ProSeal laryngeal mask airway," including a co-citation map (Fig. 3a), the journal profile of its cluster core documents and front papers (Fig. 3b), a list of the main institutions contributing to current front papers in this field (Fig. 3c), and the papers published in 2003 that cited most of the cluster core references (Fig. 3d). Such an analysis could be generated for each individual research field listed in Table 1.
View larger version (29K):
[in this window]
[in a new window]
|
Figure 3. Detailed analysis of the research field "ProSeal laryngeal mask airway." (a) Co-citation map of all documents in the cluster core of this research field. Circle size is proportional to number of citations; distances between circles are inversely proportional to the co-citation strength. Lines show co-cited pairs. Times cited refer to citations during the year 2003. (b) Journal profile of the research field "ProSeal laryngeal mask airway." Anesthetic journals that published the cluster core documents and most of the front papers in this research field are shown. (c) Institutions that published the most front papers in the research field "ProSeal laryngeal mask airway" are listed. (d) Papers published in 2003 that cited the most documents from the cluster core of the research field "ProSeal laryngeal mask airway" including the number of cited documents are listed.
|
|
In Figure 4, we show the distribution of all 284 cluster core documents and corresponding 639 front papers of the pertinent research areas published in anesthetic journals. Interestingly, in the year 2003, the journal Anesthesia & Analgesia published most of the articles from the research fronts (137 or 21.4% front papers; 62 or 21.8% cluster core documents), whereas Anesthesiology published most of the cluster core documents (which were used later as references in the front papers) (75 or 26.4% cluster core documents; 111 or 17.4% front papers). The journal Pain, the leading journal in the JCR subject category "Anesthesiology" according to the impact factor, published 74 (11.6%) front papers and 24 (8.5%) cluster core documents.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 4. Distribution of cluster core documents and front papers for all 46 research fields in the international anesthetic journals which are available from the Journal Citation Reports® of the Institute for Scientific Information listed under the subject category "Anesthesiology." The numbers in parentheses refer to percent of total cluster core documents or front papers.
|
|
|
DISCUSSION
|
|---|
The use of standard indexes and traditional bibliometric indicators with respect to highly cited papers and authors can compromise the ability to recognize present research trends and small active research fields, although highly cited papers often stand the test of time and the core papers of the research fields, in general, remain regularly cited. Single scientific papers, however, usually are not cited until a few years after their publication and only reach a citation maximum after up to 10 yr.11 Therefore, the true impact and fame of a single article often cannot be accurately assessed for at least two decades.22 This normal life span of publications shows that evaluating the significance of recent publications and evolving research fields by pure citation counting is, at best, limited. In addition, documents often become highly cited because they contain descriptions of novel methodologies not referring to specific research fields. The results of our co-citation analysis, therefore, provide a more detailed and complete quantitative-qualitative overview of current research in anesthesia and evolving research fields than is available from pure citation counting alone. By sampling the highly co-cited anesthetic literature at every possible level of aggregation, we were able to examine the distribution of clusters of references, not only horizontally per level of co-citation with respect to number, size, and structure, but also vertically across all levels, thereby continuously following the transformation and changes in consistency and content of each cluster until homogeneity was reached; this method also allowed the detection of related research areas (Fig. 2).
In the present study, we describe 46 different current anesthetic research fields identified by co-citation analysis and variable level clustering. The results of our work indicate a strong correlation between the level of identification (co-citation strength) and the activity of specific research fields in anesthesia. The fields "Volatile anesthetic-induced cardioprotection" (Table 1, rank 10) and "Experimental models of neuropathic pain" (Table 1, rank 45) were both identified at the co-citation strength level 9 and displayed a large number of corresponding front papers co-citing their core documents (27 and 30, respectively). Similarly, the field "Difficult tracheal intubation" (Table 1, rank 44), emerging at the co-citation strength level 7, showed 27 corresponding current front papers. While the field "Volatile anesthetic-induced cardioprotection" displayed a young average age of its core documents (4.1 yr), the core documents of the other two active fields had an average age of 12–14 yr (Table 1).
A standardized method to rank research fields emerging from co-citation studies is according to their "immediacy" represented by the so-called immediacy index.15,17 Identified fields with a high immediacy index are rapidly developing areas of current research, often called highly dynamic or hot fields with a high rate of recently published and shortly afterwards often cited and co-cited documents. Examples of such highly dynamic or hot anesthetic research areas include the fields "ProSeal laryngeal mask airway," "Epidural/spinal analgesia and postoperative outcome," "A novel hydroxyethyl starch preparation," and "Local anesthetics and the inflammatory response," all of which showed an immediacy index of 80% (Table 1). Other research areas are not necessarily less important than such highly dynamic fields, but are momentarily progressing more slowly. A highly dynamic development of a research field is, therefore, apparent from its immediacy index. In general, the number of front and core documents as well as the number of received citations of core documents indicates the size and spectrum of a research field. New developments and changes in a field are also sometimes noticeable by simply comparing the words in the title of previous and current publications. Known expressions and conceptions, e.g., level of sedation or depth of anesthesia, will suddenly be substituted or complemented with new ones such as the AAI index,23 the BIS index,23 and Alaris AEP index24 as observed in the research field "Bispectral analysis" (Table 1, rank 25) containing the largest research front (44 front papers) of all fields.
The "ProSeal laryngeal mask airway" research field had the highest immediacy index (100%) (Table 1). The core of the "ProSeal laryngeal mask airway" research field consisted of six highly cited documents, each describing the ProSealTM mask as a new type of laryngeal mask airway. The much larger number of papers published in the corresponding research front (n = 21) together with a high immediacy index (100%) and low average age of documents in the cluster core (2.3 yr) indicated a rapid development and high research activity in this field. The co-citation map of the "ProSeal laryngeal mask airway" research field explicitly shows the relationships among the individual documents (Fig. 3a). The first document of this cluster core was published in May 2000 by Brain et al.25 This document, placed in the center of the map, is the most cited document of the cluster (20 citations in 2003) and the sole document that was co-cited with each other document of the cluster core. In describing a novel laryngeal mask airway, the authors opened a new research field in airway management. There is high research interest in this new mask airway in Australia (Fig. 3c). Papers published in 2003 that co-cited the most documents from the cluster core of this research field represent characteristic and representative papers of this field (Fig. 3d).
On the other hand, as mentioned above, we also observed research fields with a high activity but only a few recently cited documents in the cluster core. The research field "Experimental models of neuropathic pain" (Table 1, rank 45), for example, consisted of 30 current front papers that co-cite only five fairly old cluster core documents (average age 13.6 yr, Table 1). This illustrates a strong continuing interest for this specific research field by some contemporary scientists who use trusted methods and principles for their present studies previously established in only a small number of papers that were already relevant years ago. Two of these cluster core documents have been cited 1521 and 1967 times, respectively, through July 2007.26,27 The complete absence of recent documents in the cluster core, as seen in this research field with the lowest possible immediacy index of 0%, only reflects the paucity of important authorative theoretical or practical current scientific input. Having a low immediacy index is, therefore, by no means a parameter of obsolescence as shown by two front papers clearly demonstrating novel ideas in this long-established research field.28,29
Another interesting research field is "Postoperative nausea and vomiting" (Table 1, rank 5) with the largest cluster core containing 12 homogeneous papers. Surprisingly, this field also displayed a high rate of recent papers in its core (immediacy index 75%). This large number of papers in the research "base" indicates a well founded research field the establishment of which usually takes many years. Obviously, a current innovation in this field can be observed leading to a renewal of its core literature despite the fact that the problem of postoperative nausea and vomiting is not really new. In addition, a relatively large number of current front papers (n = 19, Table 1) confirms the importance and coherence of this literature and displays the fast movement in the field dealing with new products of anesthetic pharmacology and its successful use for prevention or treatment of postoperative nausea and vomiting such as dexamethasone, granisetron, ondansetron, dolasetron, and cyclizine. Another part of this front literature defines new guidelines for the management, prevention, and treatment of postoperative nausea and vomiting.
Interestingly, the journal Anesthesia & Analgesia published not only the most original articles in 2003 but also contributed more research front papers from the detected active research fields than journals such as Pain and Anesthesiology, implying that Anesthesia & Analgesia recognizes the active research fields and allocates adequate space for their publication (Fig. 4). In contrast, Anesthesiology published most of the important fundamental work, i.e., highly cited cluster core documents, thereby establishing the greatest share of the intellectual "base" of the detected research fields.
Several limitations of our study must be noted including a potential database bias.30,31 Biased citing of articles, including self citations, may also apply to co-citations thereby affecting the co-citation analysis as well.32–34 However, in our example of the work by Struys et al.,21 there were only two self-citations, which we consider not to influence the total number of co-cited cluster documents by other investigators. Furthermore, the attribution of names to the various research fields was subjective and dependent on the decision of the authors. Although other investigators might have proposed other research field names, the content of each research field with its cluster core documents and corresponding front papers would still be the same.
In conclusion, the co-citation analysis presented in this article identified 46 active anesthetic research fields. Most of the front papers and the fundamental work described in the cluster core documents were published in two anesthetic journals, namely Anesthesia & Analgesia and Anesthesiology. This study provides information about the main structure of current anesthetic research reflecting its cognitive environment, participants, publications, and connections.
|
Footnotes
|
|---|
Accepted for publication January 31, 2008.
|
REFERENCES
|
|---|
- van Raan AF. Measuring science. Capita selecta of current main issues. In: Moed HF, Glänzel W, Schmoch U, eds. Handbook of quantitative science and technology research. Dordrecht: Kluwer Academic Publishers, 2004:19–50
- Adam D. The counting house. Nature 2002;415:726–9[Medline]
- Tague-Sutcliffe JM. An introduction to informetrics. Inform Process Manage 1992;28:1–3
- van Raan AF. Scientometrics: State-of-the-art. Scientometrics 1997;38:205–18[Web of Science]
- Wikipedia, The Free Encyclopedia, article, scientometrics. Available from http://en.wikipedia.org/wiki/Scientometrics. Accessed 25 September 2007
- Hood WW, Wilson CS. The literature of bibliometrics, scientometrics, and informetrics. Scientometrics 2001;52:291–314[Web of Science]
- de Solla Price DJ. Networks of scientific papers. The pattern of bibliographic references indicates the nature of the scientific research front. Science 1965;149:510–5[Free Full Text]
- Garfield E. "Science Citation Index" – a new dimension in indexing. Science 1964;144:649–54[Free Full Text]
- Schöch V. Die Suchmaschine Google. Algorithmen für das WWW. Institut für Informatik, Freie Universität Berlin, Germany. Available from http://www.inf.fu-berlin.de/lehre/WS01/netbasedIS/uebungen/PageRank_vortrag_schoech.pdf. Accessed 25 September 2007
- Lotka AJ. The frequency distribution of scientific productivity. J Washington Acad Sci 1926;16:317–23
- Marx W, Schier H, Wanitschek M. Citation analysis using online databases: feasibilities and shortcomings. Scientometrics 2001;52:59–82[Web of Science]
- Small H. Co-citation in the scientific literature: a new measure of the relationship between two documents. J Am Soc Inform Sci 1973;24:265–9[Web of Science]
- Crawford JW, Crawford S. Research in psychiatry: a co-citation analysis. Am J Psychiatry 1980;137:52–5[Abstract/Free Full Text]
- Jarneving B. The cognitive structure of current cardiovascular research. Scientometrics 2001;50:365–89[Web of Science]
- Schwechheimer H, Winterhager M. Highly dynamic specialities in climate research. Scientometrics 1999;44:547–60[Web of Science]
- Boyack KW. Mapping knowledge domains: characterizing PNAS. Proc Natl Acad Sci 2004;101(Suppl 1):5192–9[Abstract/Free Full Text]
- Small H, Garfield E. The geography of science: disciplinary and national mappings. J Inform Sci 1985;11:147–59
- Aversa E. Research on research: customized citation analysis for governmental, industrial, and academic clients. In: Essays of an information scientist: of nobel class, women in science, citation classics, and other Essays 1992–1993;15:77–83
- Small H, Griffith BC. The structure of scientific literatures. I. Identifying and graphing specialties. Sci Stud 1974;4:17–40[Web of Science]
- Small H, Sweeney E. Clustering the science citation index using co-citations. I. A comparison of methods. Scientometrics 1985;7:391–409[Web of Science]
- Struys MM, Vereecke H, Moerman A, Jensen EW, Verhaeghen D, De Neve N, Dumortier FJ, Mortier EP. Ability of the bispectral index, autoregressive modelling with exogenous input-derived auditory evoked potentials, and predicted propofol concentrations to measure patient responsiveness during anesthesia with propofol and remifentanil. Anesthesiology 2003;99:802–12[Web of Science][Medline]
- Garfield E. 100 citation classics from the Journal of the American Medical Association. JAMA 1987;257:52–9[Abstract/Free Full Text]
- Anderson RE, Barr G, Assareh H, Jakobsson J. The AAI index, the BIS index and end-tidal concentration during wash in and wash out of sevoflurane. Anaesthesia 2003;58:531–5[Web of Science][Medline]
- Kreuer S, Bruhn J, Larsen R, Hoepstein M, Wilhelm W. Comparison of Alaris AEP index and bispectral index during propofol-remifentanil anaesthesia. Br J Anaesth 2003;91:336–40[Abstract/Free Full Text]
- Brain AI, Verghese C, Strube PJ. The LMA ‘ProSeal’ - a laryngeal mask with an oesophageal vent. Br J Anaesth 2000;84:650–4[Abstract/Free Full Text]
- Bennett GJ, Xie YK. A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man. Pain 1988;33:87–107[Web of Science][Medline]
- Zimmermann M. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 1983;16:109–10[Web of Science][Medline]
- Kayser V, Farre A, Hamon M, Bourgoin S. Effects of the novel analgesic, cizolirtine, in a rat model of neuropathic pain. Pain 2003;104:169–77[Web of Science][Medline]
- Suter MR, Papaloizos M, Berde CB, Woolf CJ, Gilliard N, Spahn DR, Decosterd I. Development of neuropathic pain in the rat spared nerve injury model is not prevented by a peripheral nerve block. Anesthesiology 2003;99:1402–8[Web of Science][Medline]
- Elia N, Junod JD, Tramèr MR. Medline is increasingly biased in favour of journals published in the United States. A retrospective cohort study of anaesthesia literature (Abstract). Swiss Medical Wkly 2006;136(Suppl 153):15S
- Moed HF. The impact-factors debate: the ISI's uses and limits. Nature 2002;415:731–2[Medline]
- Cole S. Citations and the evaluation of individual scientists. Trends Biochem Sci 1989;14:9–13[Medline]
- Martyn J. Bibliographic coupling. J Documen 1964;20:236
- van Raan AF. Performance-related differences of bibliometric statistical properties of research groups: cumulative advantages and hierarchically layered networks. J Am Soc Inform Sci Technol 2006;57:1919–35
Top
Abstract
Introduction
