Implementing a Cloud Based Method for Protected Clinical Trial Data Sharing.

Area of Science:

• Clinical data science
• Biostatistics
• Health informatics

Background:

• Clinical trials produce vast datasets that are often underutilized due to data sharing challenges.
• Obstacles include patient privacy risks, potential data misrepresentation, and invalid secondary analyses.

Purpose of the Study:

• To develop a novel, flexible, and robust data sharing method for clinical trial data.
• To ensure patient privacy while safeguarding clinical trial investigator interests.

Main Methods:

• Implementation of a cloud-based querying language for hypothesis testing without direct data access.
• Generation of corresponding synthetic data for exploratory research and model development.
• Utilizing a permutation-based synthetic data generator coupled with the querying language.

Main Results:

• The developed system effectively addresses obstacles in clinical trial data sharing.
• Identified significant relationships among variables within a realistic clinical trial dataset using synthetic data.
• Demonstrated the system's capability for secure data exploration and analysis.

Conclusions:

• The novel approach enhances clinical trial data utility while maintaining privacy and investigator control.
• This method facilitates secure secondary analyses and model development from clinical trial data.
• The system offers a flexible and robust solution for modern clinical data sharing needs.