Privacy-Preserving Search on Medical Data.

Area of Science:

• Health Informatics
• Medical Data Security
• Bioinformatics

Background:

• Facilitating medical research requires secure and accessible data sharing spaces.
• Current data sharing mechanisms may not adequately balance researcher access with patient privacy.
• The European Union is actively promoting privacy-aware health data sharing initiatives.

Purpose of the Study:

• To propose a novel, privacy-preserving algorithm for medical data sharing.
• To enable researchers to access relevant data for specific hypotheses while minimizing re-identification risks.
• To support the development of secure health data sharing environments.

Main Methods:

• A three-party, two-stage search algorithm was developed.
• The algorithm operates on centrally stored, yet technically and organizationally separated data.
• Stage one retrieves data IDs matching cohort criteria; stage two downloads data if a minimum cohort size is met.

Main Results:

• The proposed algorithm minimizes patient re-identification risks.
• Data minimization is achieved through a staged access approach.
• The researcher uniquely connects medical and demographic data, preventing access by malicious third parties.

Conclusions:

• The developed algorithm offers a privacy-preserving solution for medical data sharing.
• This approach supports hypothesis-driven research by providing controlled access to sensitive information.
• It paves the way for future privacy-aware health data sharing spaces, aligning with EU proposals.