The Ebbs and Flows of Vaccine Equity Research: Revealing the Hidden Architecture of Knowledge Evolution Through Multilayer Networks and Natural Language Processing

Global vaccine equity is a persistent challenge that goes beyond national borders and healthcare systems. The COVID-19 pandemic has highlighted significant dis parities in access to life-saving interventions. Yet we still lack an integrated under- standing of how vaccine equity research has reorganized thematically and methodologically over time, and how scholarly attention has shifted across populations, geographies, and analytic strategies in response to external disruptions. This study conducts a systematic analysis of vaccine equity research from 2001 to 2024, examining a total of 3,939 publications from the Web of Science and Scopus databases. A multilayer analytical framework is employed, which includes natural language processing (NLP) for topic modeling, fine-tuned large language models (LLMs) for systematic content extraction, and multilayer network analysis to capture complex, interdisciplinary dynamics across fourteen distinct analytical layers. Our findings reveal that systematic shocks, such as the COVID-19 pandemic, reconfigure the structure of scientific inquiry by amplifying equity-centered agendas and redistributing methodological prominence across research layers. While convergence increased in procedural dimensions, conceptual novelty and cross-dimensional integration remained uneven. The study introduces a multilayer network framework for tracing scientific evolution in interdisciplinary domains. It offers design principles for research infrastructures that balance rapid response coordination with long-term epistemic diversity.