The Value of Data Fusion for Predicting Alarms in Critical Care

Studies show that patients who have inhospital cardiac arrests or unexpectedly require admission to an Intensive Care Unit frequently show physiological signs of deterioration prior to the event. This deterioration frequently goes unnoticed and hence is not acted on. To combat this there has been increased use of mandated vital sign measurement and Medical Emergency Teams (MET) - groups of clinical experts who are called according to criteria relating to changes in physiological parameters. An automated system for detecting patient deterioration through data fusion of heart rate, breathing rate, oxygen saturation, temperature, and blood pressure has been developed. Its performance is tested against current techniques for generating MET calls and early warning of such events is demonstrated.