Predicting readmission risk
Funded Twin Fellowship projectPredicting patient readmission risk to support early discharge decisions
Twin Fellows
Project core team
Dr Rachel Denholm
Dr Rebecca Winterborn
Dr Richard Wood
Hub Research theme
Service and resource planning
Background
Readmission to hospitals leads to adverse consequences for patients and an increased pressure on healthcare systems that are currently stretched beyond what they have ever been. Virtual Wards can be used to prevent readmission. They are an innovative new healthcare setting, designed to support patients who would otherwise be in hospital.
Virtual Wards are often used in Early Supported Discharge (ESD), where a patient is discharged from hospital sooner than they would otherwise be, whilst still receiving professional medical care from their home. This can benefit the individual, as well as relieving pressure on beds within a healthcare system.
Intended outcomes
Dr James’ and Mr Shaw’s project will develop a machine learning model to predict a patient’s risk of readmission, that is optimised for South West England. Applying the model to ESD patients, they hope to quantify the effect of Virtual Wards on readmission risk, and model an optimal number of days in hospital before ESD, depending on patient characteristics.
If successful, the model can be used to inform operational decision making within hospitals about which patients are most appropriate to receive ESD. Knowing which patients will benefit most from ESD will ensure the value of Virtual Wards, for patients and healthcare systems, is maximised.