Novel use of a hash-based tokenization for sharing sickle cell disease data without sharing protected health...

Area of Science:

• Biomedical Informatics
• Hematology
• Data Science

Background:

• Sickle cell disease (SCD) care improvements are hindered by fragmented, uncoordinated registries.
• Existing data aggregation is limited by poor interoperability and inconsistent data elements.
• Optimizing current data resources is crucial to avoid data loss and advance SCD research.

Purpose of the Study:

• To develop a privacy-preserving method for securely linking multiple SCD data collection efforts.
• To create a unified, longitudinal dataset for comprehensive SCD research.

Main Methods:

• Leveraged IRB-approved access to three major US SCD registries.
• Generated and hashed identity tokens using SHA-256 for secure, privacy-preserving linkage.
• Utilized deterministic matching of hashed tokens to identify unique individuals across datasets.

Main Results:

• Successfully linked data from the Sickle Cell Data Collection project, ASH Research Collaborative Data Hub, and GRNDaD.
• Identified 1,080 unique individuals across at least two of the three registries.
• Demonstrated the first privacy-preserving linkage of multiple SCD registries.

Conclusions:

• Secure data integration enhances interoperability for SCD research.
• This method enables richer longitudinal analyses for advancing SCD treatment.
• Optimized data linkage is vital for understanding lifelong SCD progression.