Hospital Online Uncertainty and Stability Estimation

Abstract

Monitoring the real-time physiological status of patients in intensive care following surgery is essential to predict and prevent deterioration. Most of the scoring systems currently available are offline and static, cannot detect multi-dimensional trends, rely on arbitrary thresholds to generate alarms and cannot be customised for each patient. Waveform data such as electrocardiograms carry a vast amount of information about the underlying health state of an individual; most of these data is constantly recorded but most of the acquired information is never processed or analysed. Using a Bayesian statistical modelling approach, we present a real-time adaptive dynamic monitoring system sensitive to internal variability.

Publication
In: Bissell J, Caiado C, Goldstein M, Straughan B, Curtis SE (Eds.), Tipping Points: Modelling Social Problems and Health, Part II: Mathematical Modelling in Healthcare. Wiley: London. Chapter 4.
Date