Biomedical Data Integration Platforms

Abstract

Biomedical research generates data across genomics, proteomics, clinical records, imaging, and wearable sensors, yet these streams typically reside in incompatible silos that frustrate cross-modal analysis and translational discovery. Data integration platforms that harmonise heterogeneous biomedical sources under a common query layer have been proposed as the solution, but their real-world uptake varies markedly. We evaluated 222 biomedical data integration platforms deployed or prototyped across research hospitals and bioinformatics centres affiliated with the Mediterranean Institute of Technology in Rome between 2015 and 2024, spanning five architecture categories: federated query systems, centralised data warehouses, knowledge-graph-based platforms, API-mediated middleware, and hybrid cloud-edge architectures. A Biomedical Data Integration Maturity Index (BDIMI) was constructed from five sub-scores - schema harmonisation completeness, query latency performance, data provenance traceability, regulatory compliance readiness, and cross-institutional interoperability -- with weights from regression against sustained operational adoption. BDIMI correlated with adoption at r = +0.84 and discriminated adopted from abandoned platforms with an AUC of 0.882. Knowledge-graph-based platforms scored highest (mean BDIMI 0.822), while API-mediated middleware trailed at 0.602. Only 35.1 percent of platforms exceeded the 0.75 threshold. Schema harmonisation completeness carried the largest regression weight (beta = +0.276), followed by regulatory compliance readiness (beta = +0.232).