Calculating effectiveness of eHealth without sharing any patient information


The effectiveness of eHealth solutions is hard to prove in a cost-efficient way. Part of this problem is due to data-sharing issues and privacy concerns. What if we could make statistical calculations on separate datasets without revealing or sharing any personal information?

In the Techruption use-case “privacy-preserving analytics” (PPA) we are able to mathematically  prove the effectiveness of eHealth solutions through privacy-preserving statistical analysis on private datasets from 3 different organizations: CZ (a health insurance company), Zuyderland (a hospital) and CBS (Statistics Netherlands). With the help of advanced cryptography, secure multi-party computation (MPC), the individual data-items remain encrypted at all times during the analysis; only the final, aggregated result is revealed, and no patient information is shared.

We use a blockchain and a smart contract to be able to control and govern who can do analyses (queries), what datasets can be used, what the minimal anonymity set should be etc. In addition, the blockchain provides an audit trail of all analyses that have been performed. A prototype of this approach has been developed & evaluated.

We are now planning a pilot with real patient data to use this technology to evaluate the effectiveness of an eHealth coaching app for inflammatory bowel disease.

We are eager to apply this technology on other eHealth solutions and would like to hear your feedback on our vision of a scalable platform for automated privacy preserving cost-effectiveness analysis.