Enabling Health Data Sharing with Fine-Grained Privacy

Area of Science:

• Health Informatics
• Data Privacy
• Medical Research

Background:

• Sharing health data accelerates medical research and clinical practice.
• Existing privacy methods like anonymization offer dataset-level protection.
• Individual privacy preferences and policies present unique challenges in health data sharing.

Purpose of the Study:

• To address fine-grained privacy challenges in health data sharing.
• To propose a solution enabling data curators to protect sensitive data elements.
• To balance privacy protection with data usefulness for research.

Main Methods:

• Developed a novel privacy solution using randomized techniques for sensitive data elements.
• Employed graphical models to mitigate privacy leakage from dependent data elements.
• Incorporated domain knowledge and a block-structured design to preserve data utility.

Main Results:

• The proposed approach provides rigorous privacy protection for sensitive health data.
• The method effectively mitigates privacy leakage from dependent data elements.
• Evaluations show the shared data retains high usefulness for health applications.

Conclusions:

• The novel privacy solution successfully protects sensitive health data at a fine-grained level.
• The approach balances robust privacy with essential data utility for research.
• This method enhances secure health data sharing for advancing medical science.