Densifying the Patient Journey by Integrating Data of Oncology Patients From Multiple Data Sources Through Confidential Computing

BACKGROUND: The patient journey in oncology can only be mapped authentically and with high information density if data from different sources is used. The concept presented here uses record linkage procedures to seamlessly merge clinical, registry and real-world data from different sources. A particular focus is applying the principles of Confidential Computing (CC), which aims to protect data during processing by keeping it encrypted in a secure processing environment (Trusted Execution Environment).

OBJECTIVES: The project’s main objective was to create a compliant data infrastructure through which different data sets can be automatically linked, to enable common accessibility for outcomes research analyses, market access issues and the development of new treatment methods in the light of EHDS. Forward-looking technologies, such as CC, should be used to ensure the highest possible security of data integration with maximum efficiency. The entire record linkage process is illustrated along a practical use case from oncological care.

METHODS: After the cross-sector standardization of patient data using a hashing process, the data is stored in a Data Clean Room (DCR). The DCR provides a secure environment in which uploaded patient data can be jointly analysed without the data holders involved being able to access the raw data of the other contributing parties.

RESULTS: Successful proofs of concept were provided for linkage of data within the University Hospital of Cologne and cross-sectoral, with further partners.

CONCLUSIONS: In addition to the time saved by automated record linkage and the cross-sector integration of patient data from different sources, the applied concept of CC holds enormous potential for medical research and healthcare practice. This application, implemented in the form of a DCR, improves the availability, authenticity, informative value and quality of oncological patient data by helping to close information gaps (e.g. on clinical status, staging and grading, or patient-reported outcomes) and promote cross-sector outcomes research.