Synthetic Knowledge Synthesis in Hospital Libraries

ABSTRACT

Knowledge explosion is associated with the exponential growth of research literature production, triggering the need for new approaches to structure and synthesize knowledge. Traditional knowledge synthesis approaches have weaknesses like requiring many resources, being time and labor consuming, and being executed manually. To overcome these weaknesses, we developed a new semi-automated synthesis approach, synthetic knowledge synthesis (SKS). The new approach is a triangulation of bibliometrics, bibliometric mapping, and content analysis and can be used by hospital librarians to advance their role as partners in health research. In this paper, we present SKS and demonstrate its use.

KEYWORDS:

• Bibliometrics
• content analysis
• knowledge explosion
• new library services
• synthetic knowledge synthesis

Introduction

The phenomenon of a knowledge explosion gained prominence in the second half of the twentieth century. This indicates an almost momentous increase in the amount of available knowledge resulting in exponential growth as noted by Lipmann in 1962 (1). Indeed, until the end of the nineteenth century, scientific knowledge doubled approximately every century, and then, until the end of World War II, every 25 years (2,3). Recently Bornmann and Mutz estimated that the doubling of global scientific output occurs roughly every 9 years, however, Schilling speculated that due to the exponential development of information technology human knowledge will double every 12 h in the not too distant future (4,5). A more realistic knowledge-doubling period was estimated by Densen, namely that medical knowledge will double in 73 days beyond 2020 (6). Kurtzweil, on the other hand, expressed knowledge growth in what he called The Law of Accelerating Returns and speculated that in the 21^st century, humans will experience scientific 20,000 years of progress rather than the 100 years experienced today (7).

Knowledge explosion is associated with the exponential growth of research literature production, which has positive and challenging implications. On the one hand, available knowledge in fields such as health care improves our understanding of diseases, treatments, health systems, and processes and enables us to enhance the quality of life. On the other hand, it makes it difficult to keep up with the latest developments and navigate the vast amount of evidence available. This results in the need for new knowledge synthesis approaches (8). Knowledge synthesis dates back more than 120 years; however, it became more popular in the 1960s, and even more so toward the end of the millennia with the introduction of evidence-based practices (9–11). Knowledge synthesis enables researchers to structure knowledge and overcome the problem of isolated findings that might be incomplete or lack overlap with other possible solutions (12). However, traditional knowledge synthesis approaches like reviews, meta-analysis/synthesis, and similar have some weaknesses. Primarily, they require many resources, are time and labor consuming, and are conducted manually. As a result, they are performed on a small number of publications and only provide sampling insights into certain research (13). Markoulli et al. point out that these samplings are not reproducible as they broadly explore the trees, without considering the forest (14).

Academic and hospital librarians started to support knowledge synthesis with the shift in the role of academic librarians which began in the late 1980s by co-operating in the preparation of different types of reviews, wherein they performed various tasks\such as expert evidence searching, protocol and research plan development, research question formulation, reporting, and assistance with technological and analytical tools (15–21). Indeed, in a recent case report, Demetres et al, assessing a decade of the Weill Cornell Medicine Systematic Review Service, found that this service has been highly in demand and has significantly evolved with the automation of screening and data extraction processes and the introduction of machine learning and natural language processing in knowledge synthesis (22).

In line with the above trend, the objective of this paper is to present and demonstrate the use of Synthetic Knowledge Synthesis (SKS), an approach which partially automates big data level knowledge synthesis by triangulation of bibliometric mapping, text mining, and content analysis that hospital librarians can use to advance their services (12).

Synthetic knowledge synthesis

Bibliometrics supports the analyses of vast numbers of publications on macroscopic and microscopic levels and is domain-independent (23,24). Bibliometrics was defined by Pritchard as the application of mathematical and statistical methods to books and other media of communication and by Hawkins as the quantitative analysis of the bibliographic features of a body of literature (25,26).

Bibliometric mapping is one of the bibliometrics tools which uses text mining, co-word analysis, and clustering algorithms to capture associations between pairs of units of analysis (words, phrases, country, institution or author names, or other terms found in scientific documents) and visualizes them in the form of bibliometric maps. A bibliometric map is a network of nodes where nodes represent units and the links between nodes represent relationships. The proximity of nodes can be interpreted as an indicator of their similarity, while the node size might indicate the popularity of a unit. Clusters of nodes, usually indicated by different colors, represent units that are strongly associated. Frequently used bibliometric mapping software tools are VOSViewer (Leiden University, the Netherlands) which we used in our demonstration and Bibliometrix (27,28).

Content analysis is a method that is commonly used in both quantitative and qualitative research. It enables one to systematically and objectively describe phenomena and can be applied to various types of documents – in our case research publications. Concept analysis is used to create concepts, categories, and themes (29). Synthetic knowledge synthesis follows the steps below:

1. Harvest the research publications concerning the topic of interest from the selected bibliographic databases using an appropriate search string representing the research question(s) to be answered by knowledge synthesis.

2. Select Author Keywords as meaningful units of information. Author keywords are used because they most accurately convey the content the authors want to communicate to the scientific community and are semi-formalized.

3. Perform a bibliometric mapping of an Author’s keywords into a cluster bibliometric map using selected bibliometric software.

4. Analyze the links and proximity between author keywords in individual clusters to form categories.

5. Condense categories into themes.

Demonstration of SKS use

COVID-19 forced academic and hospital libraries to suddenly and dramatically adjust their operations and service (30). Beglou and Akhshik analyzed the response of more than 300 world top university libraries (31). They found that academic and hospital libraries created new research and educational services, changed the existing ones, introduced COVID-19 research support and provided medical information about COVID-19 and how to cope with its consequences. To demonstrate the use of SKS, we posed the following research questions:

How did COVID-19 pandemic affect library services?

Consequently, the search string used was (librarian* OR “library service*) AND covid. The search was performed on the 17th of June 2023 in the Scopus bibliographic database. The search resulted in 627 publications, containing 1514 authors assigned keywords. VOSviewer was used for bibliometric mapping. The resulting authors’ keywords cluster map was comprised of 36 keywords emerging in at least six publications. It contains four clusters (resolution in VOSViewer was set to 0.70) and is presented in Figure 1.

Figure 1. The author keyword landscape.

A sample single cluster is shown in Figure 2. The representative keywords are academic libraries occurring 78 times, pandemics occurring 58 times, library service occurring 46 times, references services occurring nine times, lockdown occurring eight times, and virtual services occurring seven times.

Figure 2. A simple cluster of author keywords.

Analyzing the links between representative keywords we identified three categories. The first one was named Online library services in academic libraries during lockdown (Figure 3), and the second one Virtual reference and loan services during pandemics (Figure 4). Finally, the theme was named Virtual academic library services during pandemics and lockdowns.

Figure 3. Link analysis for the first sample category.

Figure 4. Link analysis for the second sample category.

We apply the same approach to the remaining three clusters as shown in Figure 1. The outcome of the complete SKS approach is summarized in Table 1.

Table 1. The Outcome of the SKS Approach (The Numbers in Parentheses Represent the Number of Publications in which a Specific Author Keyword Occurred).

Cluster color	Representative author keywords	Categories	Theme
Red (11 keywords)	Academic libraries (n = 78), Pandemics (n = 56), Library services (n = 46), Reference services (n=9), Lockdown (n = 8), Virtual services (n = 7)	Online library services in academic libraries during lock down; Academic library services during lockdown	Virtual and online academic libraries during pandemics and lockdowns
Blue (10 keywords)	Covid-19 (n = 261), Public libraries (n=30), Librarians (n=22), Social media (n=13), Telemedicine (n=8), Public Health (n=7), Information dissemination (n=6)	Information dissemination through public libraries and social media; Librarians and information services; Public health and telemedicine	Information dissemination through libraries
Green (10 keywords)	Information literacy (n = 38), Libraries (n = 36), Higher education (n = 15), Infodemic (n = 8), Digital literacy (n = 8)	Libraries and information literacy in association with misinformation, Higher education and information and digital literacy; Infodemic, information literacy, and fake news	Information literacy and infodemics
Yellow color (5 keywords)	On line learning (n = 19), Library instruction (n = 8)	Library instructions for online learning through digital library services	Library instructions for online learning through digital library services

Conclusion

The advantage of the SKS approach is that it first enables librarians to synthesize large corpuses of publications. The second strength is that the use of bibliometrics software visualizes the relationships and associations between units of analysis and that significantly simplifies the naming of categories and which consequently, reduces bias and increases the replicability of the synthesis. Third, it can be performed by the researcher or supported by skilled hospital librarians. This leads us to the fourth strength, namely that performing SKS can be offered as a new and innovative service in hospital libraries. The SKS approach is a semi-automatic process and can thus be executed in a shorter time with fewer resources. Last, but not least, it can be extended to other forms of textual evidence such as reports, interview transcripts, comments, and so forth. In comparison to content analysis tools like Atlas.ti or MAXDQDA, bibliometrics tools used in the SKS approach are free, simpler to use, and can perform several different kinds of mapping functions which can support the further quantitative and qualitative synthesis of evidence (32,33).

Disclosure statement

No potential conflict of interest was reported by the author(s).