Visual Analysis of Probabilistic Infection Contagion in Hospitals

Abstract

Clinicians and hygienists need to know how an infection of one patient could be transmitted among other patients in the hospital (e.g., to prevent outbreaks). They need to analyze how many and which patients will possibly be infected, how fast the infection could spread, and which contacts are likely to transfer the infections within the hospital. Currently, infection contagion is modeled and visualized for populations only on an aggregate level, without identification and exploration of possible infection between individuals. We present a novel visual analytics approach that simulates the contagion in a contact graph of patients in a hospital. We propose a clustering approach to identify probable contagion scenarios in the simulation ensemble. Furthermore, our novel visual design for detailed assessment of transmission shows the temporal development of contagion per patient in one view. We demonstrate the capability of our approach to a real-world use case in a German hospital.