Health Data Re-Identification: Assessing Adversaries and Potential Harms

Area of Science:

• Biomedical data science
• Health informatics
• Privacy engineering

Background:

• Biomedical data sharing accelerates disease understanding and therapeutic development.
• Balancing data utility with robust privacy protection is essential.
• Identifying potential adversaries and their motivations is key to mitigating re-identification risks.

Purpose of the Study:

• To comprehensively list potential adversaries in biomedical data sharing.
• To categorize adversaries by their motivations.
• To identify potential harms to individuals resulting from data re-identification.

Main Methods:

• A three-step workshop process involving 13 researchers.
• Identification of potential adversaries.
• Categorization of adversaries by motivation and deduction of associated harms.

Main Results:

• A collection of 28 adversary suggestions.
• Categorization into six distinct adversary types.
• Association of each type with several of six distinct potential harms.

Conclusions:

• The findings provide a structured framework for evaluating re-identification risks.
• The study offers a foundation for developing effective protection measures in health data sharing.
• Understanding threat actors and outcomes is vital for secure biomedical data exchange.