MedCo2: Privacy-Preserving Cohort Exploration and Analysis

Area of Science:

• Medical informatics
• Distributed systems
• Privacy-preserving technologies

Background:

• Medical studies, especially for rare diseases, are costly, time-consuming, and face challenges in patient recruitment and data sharing due to varying regulations.
• Existing methods struggle with international data collaboration and maintaining patient privacy.

Purpose of the Study:

• To present MedCo2, a novel distributed system designed to streamline medical study processes.
• To enable secure data discovery and analysis across multiple databases while ensuring data confidentiality and patient privacy.

Main Methods:

• MedCo2 utilizes interactive protocols, homomorphic encryption, and differential privacy.
• The system facilitates privacy-preserving computations of statistics (e.g., cosine similarity, variance) and machine learning model training.
• It allows for oblivious selection of patients based on specific criteria across distributed databases.

Main Results:

• MedCo2 effectively bridges data discovery and analysis among disparate databases.
• The system ensures patient privacy and data confidentiality during computations.
• Enables privacy-preserving statistical analysis and machine learning model training on distributed datasets.

Conclusions:

• MedCo2 offers a robust solution for conducting medical studies, particularly for rare diseases, by addressing data sharing and privacy concerns.
• The system facilitates collaborative research through secure, distributed data analysis.
• It represents a significant advancement in privacy-preserving medical informatics.