In 2022, Delta Dental of Michigan (DDMI), through their commercial arm Roosevelt Innovations, sought to use their sensitive data to reduce FWA (fraud, waste, and abuse) and streamline their pre-authorization process. DDMI's claims data, however, are owned by their customers and contains PHI (Protected Health Information), so are governed by HIPAA.
The challenge? Moving the data to their preferred ML vendor's cloud, H2O.ai, or to researchers would take 6-12 months due to regulatory and contractual consent impediments.
Subsalt's generative database enables DDMI to programmatically create and manage fully synthetic, de-identified data on demand and at scale.
The data are both usable for AI/ML projects and meet the HIPAA standard for being treated as non-PHI, eliminating HIPAA related privacy and security constraints.
Because the data are fully synthetic and contain no customer-owned records, they also resolve the contractual consent challenges.
DDMI is able to safely and easily provide production quality, de-identified data to H2O.ai for model development and training. This data will power various AI/ML model training needs
95% Collapse in Time to Data = 7 Figure ROI
Time to access DDMI's sensitive data was reduced from approximately nine months to less than a week. This resulted in ROI acceleration of over $1M in year 1.
Real Performance, Synthetic Data
Performance of predictive models was maintained using synthetic data as compared to models trained on sensitive source data.
All the benefits, none of the risk.