Real-time Adaptive & Predictive Indicator of Deterioration (RAPID)

1.5 million UK children are admitted to hospital every year. Approximately 35,000 children require higher dependency care, 18,000 require intensive care, 650 suffer in-hospital cardiac arrest and 3000 die in hospital. 57% of in-hospital deaths are potentially avoidable and 78% of life-threatening deteriorations have early warning signs that are unreliably detected. Current aggregate early warning systems can detect deterioration, trigger escalation and response, reducing avoidable life-threatening deterioration in hospital by 25%. However, we need earlier and reliable detection without false or missed alarms. We are developing an accurate, patient specific, continuous, early warning system by running advanced, predictive algorithms to provide real-time learning and detection of change of an individual's vital sign patterns. The software, based on McLaren F1 technology, Aston University and Birmingham Children's Hospital algorithms extracts changing patterns in physiological data from commercially available medical sensors. Physiological risk will be embedded in a ward management system with real-time presentation of patient status and alerts. This will be used to optimise identification and response to facilitate the best patient care. This system will be developed for children in a hospital, but will be highly scalable adult and pre-hospital patient populations with high specificity and sensitivity for reliable detection.