Three Pilot Projects Ensure Synthetic Data Accuracy and Maintained Patient Privacy: Washington University

Washington University School of Medicine's Institute for Informatics needed to make patient data available to researchers at scale without the weeks-to-months regulatory delays and data-fidelity loss of traditional de-identification. In 2018 it became the first U.S. medical school to adopt MDClone's synthetic-data platform, generating data that mimics real patient populations without traceable identifying information. Three pilots - head trauma severity prediction, sepsis prediction, and STI analysis - validated the approach; researcher Dr. Randi Foraker said synthetic data were "scientifically valid representations of the original data" and that a model "trained on the synthetic data performed well when applied to the original data." The platform is now also used for grant analysis, conference abstracts and cardiovascular drug-development research.

MDClone is a healthcare synthetic-data platform that converts real EHR data into privacy-safe synthetic datasets for self-service research and analytics.

This sits in the Design & Stress-Test stage of the AI use-case lifecycle: Named examples from the tools pharma teams use to scope, prototype and benchmark an AI use case before committing to build it — data-catalog, sandbox and evaluation platforms.