Data Anonymization for Pervasive Health Care: Systematic Literature Mapping Study.

Area of Science:

• Digital Health
• Data Science
• Health Informatics

Background:

• Data science in healthcare presents privacy, transparency, and trustworthiness challenges.
• Regulations like GDPR and the UK Data Protection Act mandate legal bases for data processing and patient consent.
• Existing data anonymization tools lack standardized privacy functionalities and efficacy in risk assessment.

Purpose of the Study:

• To systematically map the landscape of data anonymization techniques for digital healthcare.
• To address challenges in privacy, transparency, and trustworthiness of data science applications in healthcare.

Main Methods:

• Systematic literature mapping using major academic databases (Google Scholar, Web of Science, Scopus, PubMed) up to June 2020.
• Focused on five aspects: anonymization operations, privacy models, reidentification risk/usability metrics, anonymization tools, and lawful basis for EHR data anonymization.

Main Results:

• Identified 239 eligible studies, analyzing anonymization operations, 72 privacy models (conventional and ML-based), and 15 usability metrics.
• Reviewed 20 data anonymization software tools and evaluated practical feasibility for EHR data.
• Summarized the lawful basis for practical EHR data anonymization.

Conclusions:

• Anonymization of electronic health record (EHR) data is achievable in theory.
• Further research is needed for practical implementation to balance privacy preservation with data usability.
• Enhanced practical anonymization methods are essential for reliable healthcare applications.