Utility of Deep Learning Model to Prioritize the A&E Patients Admission Criteria

Overcrowding in hospital emergency departments is a rudimentary issue due to patients who are presenting for treatment, but do not require admission or could be treated by their own general practitioner or over-the-counter remedies. This research work analyses the existing process of patient triage admission in accident and emergency departments and attempts to apply deep learning techniques to automate, improve and evaluate the triage process. This research proposed to utilize a deep learning model for efficiency and reducing the requirement for specialized triage professionals when evaluating and determining admission, treatment in accident and emergency departments. Automating the triage process could potentially be developed into an online application which a patient or less specialized medical practitioner could potentially perform prior to presenting at emergency departments, reducing the overall inflow to emergency departments and freeing up resources to better treat those who do require admission or treatment. The core areas to be considered are the use of emergency health records (EHR), as a suitable data source for performing the triage process in emergency departments and the application of deep learning methods using the said EHR dataset(s).