Secured Multi-Party Computations for Privacy-Preserved Medical Data Analysis

Abstract

The "Internet + Intelligent Medical" advancement lets users detect common ailments online. The procedure for
diagnosis reveals numerous significant privacy issues concerning individuals' sensitive health information. This work
introduces a novel privacy-preserving self-service medical diagnosis method utilizing Secure Multi-Party
Computation (SMC) to address these issues. In the protocol, a registered client first encrypts their personal health
information and then transmits it to the facility's computer, which computes the similarity of the individual's medical
health information and the hospital's illness trait vectors. The medical facility's server ultimately identifies the disease
corresponding to the patient based on the computed similarity value. It transmits the treatment protocol for this
condition to the individual in question. The self-service healthcare diagnosis system, grounded in Homomorphic
Encryption (HE) and privacy-preserving access control, ensures the privacy of individual patient data and the secrecy
of hospital diagnostic methods. The comprehensive security research demonstrates that the technique can withstand
numerous recognized security risks. The system not only diminishes treatment costs for patients and alleviates the
substantial burden on healthcare facilities during evaluation but also forecasts additional illnesses, thereby providing
patients with a more transparent comprehension of how they are doing now and enabling them to receive the most
suitable therapy.