DPPH Key Enablers
The envisioned platform for DPPH relies on cutting-edge cryptographic technology and open-source frameworks to provide a usable, flexible and easily maintainable system that can be seamlessly deployed at hospitals and enable efficient and scalable data sharing and processing across clinical data warehouses in a privacy-conscious way, and enforcing the privacy policies of each hospital. These are the sought properties and enablers leveraged to achieve them in DPPH:
Privacy-conscious sensitive data processing
The developed framework comprises state-of-the-art protection of the sensitive patient data by relying on solid cryptographic protocols developed at LCA1 and DEDIS, such as UnLynx, and optimizing them to work on large-scale *omics data, while also protecting the aggregated results against inference attacks.
Accessibility and Usability
DPPH foresees the integration of widespread tools such as i2b2, TranSMART and SHRINE as the interfaces to the end user for accessing and operating on data; furthermore, these tools will be endowed with strong privacy and security guarantees and be part of an accessible platform for medical data sharing.
Scalability and Elasticity
The framework and architecture brought about by the Swiss Data Science Center is built on top of open source platforms for scalable and elastic data processing like Apache Spark, Hbase or Flink. DPPH extensively leverages on these platforms to provide scalable, elastic and efficient processing that can cope with large amounts of *omics data with high throughput.
Data Traceability and Management
DPPH relies on distributed ledger technologies (a.k.a. blockchains) to provide integral traceability of the results and enable effective audits, while respecting the ownership, confidentiality and privacy of the databases, therefore properly addressing the governance of data and results in bioinformatics-related research.
Security and Accountability
DPPH relies on collective authority (cothority) architectures developed and maintained by DEDIS, featuring a robust security level of the stored and processed data, and trust decentralization that prevents data leakage even in worst-case scenarios where all servers, but one get compromised. Blockchain-based immutable databases also guarantee accountability of the performed actions.
DPPH accounts for mHealth technologies, key enablers for P4 medicine, and analyzes and addresses the privacy enforcement in mobile platforms and the use and leakage of personal information in mHealth apps in connected devices.
The DPPH architecture enables efficient and scalable privacy-conscious data sharing across a federation of hospitals, comprising all steps from the encrypted ETL (Extract, Transform, Load) at the hospitals, to the sharing and processing of the encrypted and protected data in the federation. The figure below depicts an example instantiation integrating i2b2 and SHRINE.