Electronic Health Record Anomaly Detection System

Create a sophisticated Python anomaly detection pipeline for electronic health records using advanced statistical techniques and machine learning. Develop a system that can identify potential data entry errors, unusual patient patterns, and potential fraud indicators across multiple medical databases. Utilize numpy for numerical processing, implement multiple detection algorithms (isolation forests, clustering-based methods), and create a real-time alerting mechanism with configurable sensitivity thresholds.