Emergency departments(EDs)are facing unprecedented levels of overcrowding which inevitably lead to increased delays inpatient care. Without the ability to increase the size of an ED or the number of clinical staff, the only way to minimize delays in care is to use available resources more effciently.
We propose to take advantage of the data collected by EDs at all patient visits to develop probabilistic machine learning models to predict patient outcomes(e.g., discharged or admitted to hospital) at the point of departure from an EDs. These models will predict the patient outcome as early as possible in a patient's visit, potentially enabling more efficient running of EDs and the hospitals they admit patients to. Notably, the earlier a hospital gets an indication of an incoming admission the more proactive preparations they can make, ultimately improving patient flow.
We will use the database (containing approximately one million patient episodes) maintained by University Hospitals Southampton (UHS) ED to develop models which perform retrospective patient outcome predictions. Successful demonstration of patient outcome prediction models will allow us to apply for additional funding to implement the developed prediction tools into front line clinical practice. As an NHS Global Digital Exemplar, UHS has a responsibility to disseminate successful research in IT solutions to enable other trusts to follow in their footsteps, utilizing this network would allow software solutions we develop to have national impact.
Principal Investigator: Professor Neil White (University of Southampton)
Co-Investigators: Dr Zlatko Zlatev (University of Southampton) and Dr Michael Kiuber and Dr Thomas Daniels (University Hospital Southampton)