Record linkage based patient intersection cardinality for rare disease studies using Mainzelliste and secure...

Area of Science:

• Medical Informatics
• Computational Biology
• Epidemiology

Background:

• Rare disease research faces challenges due to fragmented patient data across multiple institutions.
• Legal and technical barriers hinder data sharing and joint analysis of sensitive medical records.
• Existing solutions often lack robust privacy guarantees or are difficult to implement.

Purpose of the Study:

• To develop and evaluate the Mainzelliste Secure EpiLinker (MainSEL) for privacy-preserving record linkage.
• To enable secure calculation of common patients between institutions for feasibility studies.
• To facilitate data consolidation and rare disease research through enhanced security and usability.

Main Methods:

• Utilized a secure Multi-Party Computation (MPC) based application with trusted-third-party-less cryptographic protocols.
• Extended MainSEL to perform record linkage for calculating the number of common patients.
• Employed microservice containerization and CI/CD for easy software deployment.
• Conducted real-world testing with medical researchers using synthetic patient data.

Main Results:

• MainSEL demonstrated practical runtimes, processing 10,000 comparisons in approximately 5 minutes.
• The framework proved feasible across diverse network settings and use cases.
• Real-world testing highlighted the need for explicit support for analysis pipeline integration and ad-hoc analysis.

Conclusions:

• MainSEL is a valuable tool for rare disease and general medical research, offering practical runtimes and enhanced security.
• The software is easily deployable in strict clinical IT environments.
• MainSEL's design, informed by real-world testing, aims to meet diverse researcher needs using modern privacy-preserving technologies.