Mauro Dala Santos
St Hilda's College
CDT Postgraduate Studentship in Healthcare Innovation (RCUK Digital Economy Programme grant number EP/G036861/1)
Vital-Sign datafusion to detect patient deterioration in the Emergency Department
Acute hospitals in the UK are required to use “track-and-trigger” (T&T) systems in which vital-sign data is collected periodically from patients and then scored according to their abnormality. If the total score exceeds a pre-defined threshold, the care of the patient is escalated. The workload of the Emergency Department (ED), in which clinical observations are taken more frequently (less than 1 hour) than in other wards, affects the completion rate of T&T charts. This has the consequence that patient deterioration may be missed between observations. Currently a clinical trial is being conducted in the Jonh Radcliff ED, to study if the combination of an electronic T&T system and a continuous datafusion system connected to bedside monitors will help decrease patients’ length of hospital stay and short-term mortality. My current research consists of validating and analysing the database that will result from this trial that will include anonymised clinical information and vital-sign data from 9000 patients and developing machine learning techniques based on novelty detection and Gaussian processes to improve the detection of patient deterioration in the ED.