Building Federated Data Networks With Common Data Models to Generate Insights Through Real-World Evidence Observational Studies in Oncology
Author(s)
Oeste C1, Hens D1, Fovel I2, Bringas C1, Encarnacion E1, De Feu G1
1LynxCare, Leuven, Flemish Brabant, Belgium, 2LynxCare, Everberg, Belgium
Presentation Documents
OBJECTIVES: Accessing and standardizing raw clinical data across multiple hospitals presents a challenge in Oncology. However, it is crucial to use real-world data sources such as electronic health records (EHR) to leverage untapped information. We are building a federated data network to facilitate GDPR-compliant data exchange of large datasets, with hospitals as owners. This network, governed by a common data model (CDM), is aimed at fostering multicenter, observational, real-world evidence (RWE) studies in Oncology, with breast cancer, lung cancer, and immunotherapy as therapeutic areas of focus.
METHODS: Oncology studies are ongoing in participating Belgian hospitals using the LynxCare data processing technology over 4 different EHR systems to automatically process structured and unstructured data, using natural language processing for the latter, to generate OMOP-CDM databases. Variables (n = 1056) include demographics, comorbidities, cancer diagnosis, tumor staging, performance status, oncology treatments and procedures (including immune checkpoint inhibitors), anatomical pathology data (genomics), and adverse events. Collaborating with key opinion leaders (KOLs), we established an Oncology research group in Belgium that is redefining data dictionaries, establishing a shared comprehension per variable and enhancing data granularity.
RESULTS: To date, the Oncology data network comprises 71572 patients, for which more than 5 million unstructured records were processed and there were 27149 mappings from structured data sources (e.g., administered and prescribed medication, laboratory parameters, multidisciplinary Oncology consult, and mortality). More than 1000 quality-controlled variables are measured continuously to identify patterns, trends, and associations between datapoints and patient outcomes.
CONCLUSIONS: Clinicians can leverage the resulting RWE to develop personalized treatment plans based on patients’ specific characteristics, disease progression, and prognostic factors. These insights can also inform evidence-based guidelines and regulatory decisions. This groundbreaking initiative will expand to other European countries, providing a sandbox of federated data networks for multicenter RWE studies in Oncology, but also paving the way towards precision medicine.
Conference/Value in Health Info
Value in Health, Volume 26, Issue 11, S2 (December 2023)
Code
RWD84
Topic
Clinical Outcomes, Methodological & Statistical Research, Real World Data & Information Systems, Study Approaches
Topic Subcategory
Artificial Intelligence, Machine Learning, Predictive Analytics, Clinical Outcomes Assessment, Distributed Data & Research Networks, Electronic Medical & Health Records
Disease
Drugs, Oncology, Personalized & Precision Medicine