OpenMedicare – 10 years of Medicare physician data analyzed for fraud patterns (www.openmedicare.us)

OpenMedicare has leveraged a decade's worth of Medicare physician data, analyzing 10 years of payments totaling $854.8 billion to uncover patterns of fraud using advanced AI techniques. By training a machine learning model on the records of over 8,300 convicted fraudsters, the initiative flagged 500 providers who exhibit billing behaviors indicative of criminal activity yet remain undetected. Notably, the analysis revealed staggering statistics, such as one nurse practitioner averaging 413 Medicare services per day and a Beverly Hills surgeon charging 59 times the standard Medicare rate for wound care. This project is significant for the AI and healthcare oversight communities as it highlights the effectiveness of AI in identifying fraudulent medical billing more efficiently than traditional audits. The model's ability to uncover complex patterns of abuse, such as inflated billing in the fast-growing category of physician-administered drugs, demonstrates the potential for machine learning to transform fraud detection in healthcare. OpenMedicare not only provides a comprehensive resource of payment histories across 30,000+ providers and 130+ specialties but also emphasizes the value of data-driven investigations in enhancing accountability within Medicare.