Saving €2.5M Annually for Public Health Funds
In a national healthcare system, atypical prescribing was draining public funds — but no individual case raised a flag in isolation. The problem only emerged through statistical testing and clustering across the full physician population: a disproportionate share of atypical volume was concentrated among a small number of providers, a pattern invisible to manual review.
We developed an algorithmic detection system to audit national prescription data:
Statistical Modeling: Built robust baselines for distinct provider specialties.
Outlier Detection: Automatically flagged prescriptions that deviated significantly from the peer-group norm.
€2.5M
Annual Savings
€200k
Monthly reduction
€350k
EU Grant Secured
EU Recognition
The project methodology helped secure a €350,000 grant from the European Commission for AI innovation in public administration.
