Early detection of critical illness in hospital wards may help prevent the occurrence of adverse clinical events such as cardiac arrest, unexpected admission to the intensive care unit, and death. The current standard of clinical practice for patient monitoring on surgical wards in the UK is regular manual observations of the main vital signs by nursing staff and the usage of early-warning-score systems, in which scores are assigned to the manual observations, and care escalated if the scores exceed some predefined threshold. However, the performance of these scoring system are constrained by the large time intervals between consecutive measurements. The use of continuous vital-sign monitoring technology has been proposed to overcome this limitation by providing acute-care patients with continuous surveillance for the duration of their stay on the ward. Machine learning algorithms for early warning of physiological deterioration will be developed for continuous vital-sign signals, using datasets containing data from post-operative patients taking part in a clinical trial at the Oxford Cancer Centre.
Our research group have developed a novel means of extracting continuous vital-sign signals from video camera footage. Development continues in this area with the Digital Health in a Connected Hospital (DHiCH) study, in which non-contact vital-sign monitoring data will be collected in a post-surgical setting. Using this dataset, machine learning algorithms for detecting physiological deterioration from non-contact vital-sign monitoring signals will be derived and tested at a proof-of-concept level.