From electronic health records to real-world evidence: Savana Next-Generation Registry digital solution

Abstract

Background The secondary use of electronic health records (EHRs) offers a major opportunity to generate real-world evidence (RWE) to complement clinical trials and support regulatory and clinical decision making. However, a large proportion of clinically relevant information remains embedded in unstructured text, limiting its systematic reuse, particularly in complex areas such as oncology. This article presents the Savana Next-Generation Registry (SNGR), a methodology for building oncology-focused registries through the extraction and structuring of unstructured EHR data. Materials and methods SNGR uses EHRead®, a clinical natural language processing engine trained on multilingual EHRs, to transform free-text data into standardized clinical variables. By extracting clinically relevant information from unstructured clinical narratives, the approach enriches existing structured EHR data, enabling more complete and clinically meaningful datasets. The solution integrates medical expert input, terminology curation, and a multilayer quality assurance framework designed to support data validity, reproducibility, and traceability. Results SNGR is aligned with European regulatory frameworks, including the European Health Data Space, General Data Protection Regulation, and Data Act, enabling compliant and interoperable secondary data use. To date, it has supported over 65 real-world use cases across multiple therapeutic areas, including oncology, generating >30 peer-reviewed publications and analysis-ready datasets in multicenter and international collaborations. Conclusions SNGR demonstrates how unstructured EHR data can be transformed into analysis-ready datasets for RWE generation. It provides a scalable and methodologically robust infrastructure for next-generation registries, particularly suited for oncology and other data-rich clinical domains, supporting research, drug development, and health care decision making. As the framework evolves, future studies will continue to expand its evidence base.