Analyzing the Chinese landscape in anti-diabetic drug research: leading knowledge production institutions and thematic communities

Background: The discovery of anti-diabetic drugs is an active Chinese medicine research area. This study aims to map out anti-diabetic drug research in China using a network-based systemic approach based on co-authorship of academic publications. We focused on identifying leading knowledge production institutions, analyzing interactions among them, detecting communities with high internal associations, and exploring future research directions. Methods: Target articles published in 2009-2013 under the topic "diabetes" and subject category "pharmacology & pharmacy," with "China," "Taiwan," "Hong Kong," or "Macao" (or "Macau") in the authors' address field were retrieved from the science citation index expanded database and their bibliographic information (e.g., article title, authors, keywords, and authors' affiliation addresses) analyzed. A social network approach was used to construct an institutional collaboration network based on co-publications. Gephi software was used to visualize the network and relationships among institutes were analyzed using centrality measurements. Thematic analysis based on article keywords and R-sc value was applied to reveal the research hotspots and directions of network communities. Results: The top 50 institutions were identified; these included Shanghai Jiao Tong University, National Taiwan University, Peking University, and China Pharmaceutical University. Institutes from Taiwan tended to cooperate with institutes outside Taiwan, but those from mainland China showed low interest in external collaboration. Fourteen thematic communities were detected with the Louvain algorithm and further labeled by their high-frequency and characteristic keywords, such as Chinese medicines, diabetic complications, oxidative stress, pharmacokinetics, and insulin resistance. The keyword Chinese medicines comprised a range of Chinese medicine-related topics, including berberine, flavonoids, Astragalus polysaccharide, emodin, and ginsenoside. These keywords suggest potential fields for further antidiabetic drug research. The correlation of -0.641 (P = 0.013) between degree centrality and the Rsc value of non-core keywords indicates that communities concentrating on rare research fields are usually isolated by others and have a lower chance of collaboration. Conclusion: With a better understanding of the Chinese landscape in anti-diabetic drug research, researchers and scholars looking for experts and institutions in a specific research area can rapidly spot their target community, then select the most appropriate potential collaborator and suggest preferential research directions for future studies.