Beyond Rankings: Tree Plots for Co-word Analysis of Mobile Health Research in JMIR Mhealth Uhealth Using Cohen’s d Effect Sizes (2017–2026): Bibliometric Analysis (Preprint)

BACKGROUND Bibliometric rankings summarize research output but may obscure the strongest network links and the degree of concentration among leading contributors or topics. OBJECTIVE To map publication trends, collaboration, and themes in mHealth research published in JMIR mHealth and uHealth, and assess FLCA tree plots and the Absolute Advantage Coefficient (AAC) as complementary summaries. METHODS We analyzed English-language Web of Science Core Collection records published from 2017 through 2026. We summarized annual output and citation patterns and constructed article-level co-occurrence networks for countries, institutions, departments, authors, journal/year, and keywords. The follower-leading clustering algorithm (FLCA) retained each follower’s strongest weighted co-occurrence degree link to a leader. AAC was calculated from the top three values and interpreted using cutoffs calibrated to conventional Cohen’s d benchmarks. Career-level coauthorship for the most productive author was described using AAC, structural hole capacity (SHC), collaboration leadership index (CLI), and middle-author rate (MAR). RESULTS The analysis included 1,878 records: 1,579 articles and 299 reviews. Annual output peaked at 388 records in 2020. The United States led country output (641 records; h-index, 64), followed by the United Kingdom (238) and China (209); country AAC was 0.703, below the 0.712 medium-effect boundary. The University of California System led institutional output (89; AAC, 0.504). Table rankings tied Christine Jacob and Puhong Zhang at six publications each. The leading keywords were HEALTH (227), CARE (199), and INTERVENTIONS (163); their AAC of 0.483 was within the negligible-effect range (0.410–0.590). In the plotted author tree, AAC was 0.600; the department tree AAC was 0.586. In Jacob’s career network, her 19 publications compared with 3 for each of the next two coauthors, yielding an AAC of 0.864; SHC and CLI were 0.667, and MAR was 50%. FLCA trees displayed strongest links across the analyzed domains. CONCLUSIONS FLCA tree plots made strongest-link structures visible, while AAC described concentration beyond rank order. Keyword concentration was s small effect size, country concentration was small and near the medium boundary, and Jacob’s career network showed a large leading-author gap. Findings represent one journal’s WoSCC records and may not generalize to all mHealth research.