Rapid screening of influenza patients with severe clinical signs using an extensible neural-network-based infection screening system

Abstract

The outbreak of emerging infectious diseases frequently becomes severe are threatening people on global basis. To classify people into severe-influenza patients, mildly-influenza patients and healthy people at places of mass gathering, we developed an influenza screening system. The system conducts screening within 10 s from vital-signs, i.e., respiration rate, heart rate, and facial temperature. A neural network based discriminant function was implemented into the system to predict the infected individuals. We conducted influenza screening for 35 seasonal influenza patients at the Japan Self-defense Central Hospital. To assess the clustering performance of this system, SpO₂ was measured as a reference. The system classified 34/35 influenza patients to be mildly or severe influenza. The 10/17 severe-influenza group indicated SpO₂ less than 96%, while, only 2/17 mildly-influenza group showed SpO₂ less than 96%. This result indicates that the system has the potential to serve as a helpful tool for rapid screening of influenza in clinical settings at places of mass gathering.