Non-pharmacological traditional Chinese medicine interventions for depression: bibliometric mapping of Chinese databases and WoSCC, 2000–2025
Abstract
Objective The aim of the study was to characterize publication activity, collaboration structures, keyword themes, and temporal changes in the literature on non-pharmacological traditional Chinese medicine interventions for depression. Methods Journal records published from 2000 to 2025 were obtained from CNKI, Wanfang, VIP, and the Web of Science Core Collection (WoSCC). Eligibility combined mechanical rules with independent paired assessment of 2,089 boundary records and reference-based reconciliation. The final corpora contained 5,385 Chinese and 721 WoSCC records. Full-counting collaboration and keyword networks were analyzed using weighted Louvain communities, cosine-distance silhouettes, two-state burst intervals, and resolution, seed, Leiden, and terminology sensitivity analyses. Selected graph quantities were cross-checked in R/igraph. Additional checks evaluated record-identity consistency, linkage sensitivity, and annual keyword availability denominators. Results Chinese output peaked at 390 records in 2021, while WoSCC output reached 118 records in 2025. Complete author networks contained 12,141 Chinese and 2,933 WoSCC nodes and were sparse (densities 0.000410 and 0.003572). Eligible keyword networks contained 569 Chinese and 81 WoSCC terms. Weighted Louvain yielded five Chinese communities (Q = 0.194124; mean silhouette = 0.005173) and four WoSCC communities (Q = 0.134767; mean silhouette = 0.029369). Stability analyses showed only moderate agreement across seeds and alternative algorithms, indicating that the communities are exploratory summaries rather than discrete topic classes. Keyword availability denominators materially altered some WoSCC community time profiles. Conclusion The literature expanded substantially and centered on acupuncture-related interventions, poststroke depression, evidence synthesis terminology, exercise-based techniques, sleep, and experimental themes. Sparse collaboration networks and overlapping thematic interpretations indicate opportunities for broader collaboration and more consistent terminology. Bibliometric prominence does not establish efficacy, safety, mechanism, or causality.