PATDISCOVER: Privacy-Preserving Patient Discovery Framework

This document is a technical report that presents PATDISCOVER, a privacy-preserving framework designed for the cross-institutional discovery of patients based on shared attributes while ensuring data security. The report outlines the challenges faced in sourcing medical experience for effective patient treatment strategies, particularly in the context of rare diseases where patient data is scarce. It discusses the limitations of existing approaches to patient data sharing and highlights the need for a solution that maintains patient privacy. PATDISCOVER employs fully homomorphic encryption to allow institutions to submit complex queries without exposing sensitive patient information. The report evaluates the framework using real-world patient data from nuclear medicine, demonstrating its performance, scalability, and security. The contributions of the work include the development of efficient homomorphic encryption operators and the validation of the framework's feasibility for indexing and discovering patients with rare diseases. The document also emphasizes the importance of ethical considerations in handling patient data.