MedCo is the first operational system that makes sensitive medical-data available for research in a simple, privacy-conscious and secure way. It enables a consortium of clinical sites to collectively protect their data and to securely share them with investigators, without single points of failure.
MedCo applies advanced privacy and security techniques, such as:
- Collective homomorphic encryption;
- Secure multi-party computation;
- Differential Privacy.
DISTRIBUTED DATA EXPLORATION AND ANALYSIS
Authorized investigators can quickly explore and analyse data distributed across several clinical sites.
Local data control
Each clinical site keeps full control over its data and can decide to store the data locally or to use an external storage provider such as a private/public cloud.
There is no need of any central authority as, thanks to secure multi-party computation, trust is distributed across all clinical sites.
End-to-end Data Protection
Thanks to homomorphic encryption, clinical sites’ sensitive data is ALWAYS encrypted: at rest, in transit and during computation.
MedCo ensures unlinkability between query end-results and the clinical sites having generated them.
Replicated audit Log
MedCo keeps trace of all the operations performed within the system for full transparency and auditability.
DISTRIBUTED DATA EXPLORATION
Learn the number of individuals who match a set of provided inclusion/exclusion clinical and/or genomic criteria at the participating institutions, without leaking individual data.
Fully integrated with widespread cohort exploration tools and data models, with a modern graphical user interface.
e.g. “How many living patients included in the database are more than 50 years old and have melanoma and a given genetic mutation?”
DISTRIBUTED DATA ANALYSIS
Perform statistical analyses and train machine learning models on a group of individuals’ data, in order to generate new discoveries (e.g., linear and logistic regression, survival analysis, neural network operations).
Post-quantum protection for distributed and federated machine learning operations. Highly scalable and efficient secure data processing.
e.g. “Compute the overall survival rates within the selected cohort for different types of treatment and/or other stratification factors (e.g., age, gender, initial diagnosis).”