Synthetic Data for the Get With The Guidelines-Stroke Registry.

Area of Science:

• Cardiovascular research
• Health informatics
• Data science

Background:

• The American Heart Association's Get With The Guidelines-Quality Improvement registry contains millions of real-world cardiovascular and stroke records.
• Access to this data is crucial for research but limited due to privacy concerns.

Purpose of the Study:

• To generate a privacy-preserving synthetic dataset from the Get With The Guidelines-Stroke registry.
• To facilitate broader access and exploration of real-world stroke data for research.

Main Methods:

• Randomly sampled 1000 records from 7.8 million registry records (2005-2021).
• Shifted time variables and replaced patient identifiers to ensure privacy.
• Compared demographic and clinical variable distributions between original and synthetic data.

Main Results:

• The synthetic dataset closely mirrored the original data's distributions for demographics (age, race, sex).
• Key stroke metrics, including door-to-needle time and time to thrombolytic therapy, showed similar distributions.
• The synthetic data successfully preserved critical data characteristics.

Conclusions:

• Generated synthetic data provides a viable, privacy-protected alternative for accessing real-world stroke data.
• This open-access dataset empowers researchers to explore cardiovascular and stroke data more freely.
• Facilitates wider research and potential advancements in stroke care.