Efficient Protection of Health Data from Sensitive Attribute Disclosure

Area of Science:

• Biomedical Informatics
• Data Privacy
• Health Data Security

Background:

• Biomedical research increasingly relies on large datasets, necessitating health data sharing and reuse.
• Sharing health data raises significant privacy concerns, requiring robust anonymization techniques.
• Scalability and efficiency are critical for implementing privacy-preserving methods in practice.

Purpose of the Study:

• To address the scalability challenges in anonymizing numeric biomedical data using the t-closeness privacy model.
• To develop and present a series of optimizations for efficient production use of t-closeness anonymization.
• To improve the performance of anonymization processes for sensitive health information.

Main Methods:

• Focus on optimizing the algorithmic representation of the t-closeness model for totally ordered attribute values.
• Development of a series of practical optimizations to enhance computational efficiency.
• Experimental evaluation of the proposed optimizations on anonymization processes.

Main Results:

• The proposed optimizations significantly improve the efficiency of t-closeness anonymization.
• Execution times for anonymization processes involving t-closeness were reduced by up to a factor of two.
• Demonstrated the practical applicability and performance gains of the optimized approach.

Conclusions:

• The developed optimizations provide a scalable and efficient solution for t-closeness anonymization of biomedical data.
• This work facilitates more secure and widespread sharing of health data for research purposes.
• The findings contribute to advancing privacy-preserving techniques in the era of big data in medicine.