PROSPECT
PROSPECT: Predicting deterioration following critical illness using continuous vital sign monitoring and machine learning
Every year in the United Kingdom 1 in 10 (14,000) people discharged from an Intensive Care Unit (ICU) die unexpectedly, or urgently return to an ICU before hospital discharge. Of those who leave hospital, a quarter are unexpectedly readmitted within 3 months. It is not known whether additional electronic health information, or wearable monitoring after ICU, may help staff to identify which patients are at highest risk and might benefit from closer follow-up.
The PROgnosiS Prediction after Enhanced or CriTical care (PROSPECT) study aims to develop a prediction model for deterioration after ICU discharge using routinely collected hospital data, and to explore whether adding short-term wearable monitoring (already established in other hospital settings) can improve risk prediction. We will also work with patients and staff to understand how this information could be used in practice.
To do this, we will:
-
use existing research and expert input to identify the most useful patient information (for example, certain blood tests) for predicting risk after ICU discharge,
-
extract relevant data from historical patient records to build and validate a prediction model for deterioration,
-
ask patients to wear a monitoring system for up to 14 days after leaving ICU, and for up to 14 days after leaving hospital,
-
explore whether wearable data adds value to routine information in predicting deterioration,
-
work with patients, families and clinical staff through interviews and questionnaires to assess acceptability and refine how monitoring could be incorporated into care, and
-
carry out a pilot phase to test feasibility and inform how such approaches might be used in future studies and clinical practice.