Real-Time Patient Data Anomaly Detection System

Design a complex SQL-based anomaly detection system using window functions and statistical aggregation techniques to identify potential medical data inconsistencies in real-time. Create a stored procedure that compares patient records against multiple statistical thresholds, implementing Z-score and interquartile range methods. The system must generate alerts with confidence levels, preserve data privacy, and log potential anomalies without compromising patient confidentiality.