A visual approach for analyzing readmissions in intensive care medicine

Intensive care units (ICUs) are under constant pressure to balance capacity. ICUs have a limited number of resources and therefore effective monitoring of planned and unplanned transfers is crucial. Transfers can result in critical readmissions, i.e. "down" transfers from ICUs to a normal ward or an intermediate care unit (IMC) and back "up" to an ICU within a short time span. In this work, we present a tool to visually analyze such readmissions. Patient transfer data is extracted from a clinical data warehouse via HL7-FHIR. The interactive prototype consists of a timeline of readmission cases, an aggregated view of transfer flows between wards, and histograms and calender heatmaps to show a set of key performance indicators. The aim of our tool is to support identifying peaks, discovering temporal patterns, comparing wards, and investigating potential causes. We report on our user centered approach, describe the data pipeline, present the visualization and interaction techniques of the functional prototype, and discuss initial feedback.