Synthetic Knowledge Synthesis in Pediatric Endocrinology: Comparing WoS and Scopus Bibliographic Databases

Background: The integration of Artificial Intelligence (AI) into the management of pediatric metabolic diseases offers unprecedented opportunities for precision medicine. However, the explosive growth of research literature production has led to a fragmented research landscape, often skewed by the indexing biases of major academic databases. Objective: This study aims to conduct a comparative thematic analysis of the Web of Science and Scopus databases to uncover the distinct research paradigms governing AI in pediatric metabolic diseases. Methods: We employed the Synthetic Knowledge Synthesis methodology, integrating automated bibliometric mapping (co-word analysis via VOSviewer) with qualitative content analysis. Metadata was extracted from both databases and author keywords were clustered to evaluate underlying thematic structures. Results: The comparative analysis revealed a significant thematic divergence. Literature indexed in WoS predominantly emphasizes algorithmic novelty and methodological advancement, highlighting the use of Deep Learning, Large Language Models (LLMs), and complex metabolomic integrations. Conversely, Scopus encapsulates a distinctly clinical and translational paradigm, prioritizing Explainable AI (XAI), the integration of Natural Language Processing (NLP) with Electronic Health Records (EHR), and the application of clinical decision support systems like Continuous Glucose Monitoring (CGM). Conclusion: Relying on a singular bibliographic database provides an incomplete view of the field, creating a disconnect between algorithmic development and clinical implementation. To successfully bridge the "algorithm-to-clinic" gap in pediatric endocrinology, researchers must adopt a holistic approach that synthesizes the predictive power emphasized in WoS with the clinical transparency and applicability highlighted in Scopus.