Thematic Structure of Clinical Rehabilitation Research in the Early Post-Pandemic Period: An LDA-Based Analysis of...

Background/Objective: Recent advances in scientometric topic modeling have enhanced the capacity to rapidly assess evolving research priorities. This study aimed to characterize the thematic reorganization of clinical rehabilitation research during the early post-pandemic period and to compare how an Latent Dirichlet Allocation (LDA)-derived topic structure, SciVal, and OpenAlex represent this interdisciplinary research domain. Methods: Titles of 55,711 publications on clinical rehabilitation (2022-2025) were analyzed using an LDA-based topic modeling pipeline to identify predominant thematic directions. For comparison, topics automatically generated by Scopus and OpenAlex were examined. Results: The analysis revealed ten major thematic clusters encompassing the entire publication corpus. These clusters represented both medical and non-medical dimensions of rehabilitation research. The principal directions included post-COVID-19 recovery, psychological and quality-of-life outcomes, technological innovations such as virtual reality and telemedicine, and the growing emphasis on multidisciplinary and integrative rehabilitation frameworks. Conclusions: The proposed approach provided a field-wide, two-level representation of rehabilitation research and enabled a direct structural comparison of alternative topic-classification systems. The analysis showed systematic differences in aggregation, fragmentation, and corpus coverage across the compared approaches, without implying the universal superiority of any single method.