LB06 Panel - Demonstrations in Synthetic Data and the National COVID Cohort Collaborative (N3C)

Abstract

This panel discusses the use of synthetic data for various real analyses related to the COVID-19 pandemic. Using a synthetic dataset derived from N3C data, panelists have performed studies of data characterization, epidemic measurement, prevalence prediction, and inpatient severity prediction, and compared the data to results drawn from the original dataset. The panel provides a comprehensive review of how synthetic data can be used for analysis. While each of the studies could be considered separately, the combination of studies together creates an opportunity to clearly recognize the status of synthetic data validation, which can better advance the appropriate application of the technology in informatics broadly.

Jason A. Thomas

PhD
Medical Data & AI Scientist | Strategist | Informatician | Tech lead - Senior Data & AI Scientist - Philips

My research interests include 1) building foundational layers (data, infrastructure, knowledge representation, talent, culture) to support biomedical data science and 2) applying data science & AI methods on data to drive business value and improve patient outcomes.