Computational Biomedical Knowledge and the Language of Discovery: A Cross-Repository Metadata Evaluation [Poster]

Abstract

Patient cohorts in real world data analysis (RWD) are defined by computable, machine-executable definitions, or computable phenotypes. Access to well-described, reusable computable phenotypes, can reduce redundancy and improve reproducibility in RWD research. Therefore, systems to store and share phenotypes in accordance with FAIR principles is crucial. Despite broad consensus on the importance of sharing cohort-defining CBK, a key barrier to interoperability remains: the absence of a shared descriptive language. Phenotype and codeset repositories often operate in isolation and apply heterogeneous terminology and metadata, even for similar artifacts. A lack of standardized vocabularies limits discovery, comparison, and reuse across systems. Poster designed and presented by Annabel Pinkney on May 7, 2026 at the Medical Library Association Research Training Institute Virtual Poster Session.