University of Edinburgh

Mr Konstantin Georgiev

The Sir Jules Charitable Trust PhD Scholarship award

DL_2022_001

An ageing population is a major success of modern healthcare, but this challenges the NHS to better support an increasingly frail hospital population. One third of older people acquire new disability by discharge, leaving hospital with less independence than before getting ill. Rehabilitation attempts to maximise recovery, but this is not well targeted to people at highest risk of disability, as the risk factors are not well understood. However, electronic health records now routinely hold information about rehabilitation progress. In this project, we will use methods like machine learning to find patterns from previous admissions of older patients. The aim is to build a tool that uses data to predict the rehabilitation needs of an older patient at the point of hospital admission, and display this in an understandable way for healthcare professionals and patients themselves. A future trial could then test this tool to help target hospital rehabilitation better.

Frailty and disability after hospital care are a growing challenge to the NHS, which will only increase as our population ages. More than one third of older people leave hospital with worsening disability, leading to greater care needs and loss of independence. Healthcare resources to support rehabilitation are finite, but there is insufficient evidence to target this care efficiently. Current approaches do not support personalized care for complex older patients, many of whom have multiple chronic health conditions (multimorbidity). There is a clear need to target rehabilitation better for older people in hospital, to maximise function and limit disability. As well as benefiting patients, this would have major benefits for NHS efficiency and wider society. With the recent introduction of Electronic Health Records, clinical pathways and outcomes from rehabilitation can be studied at scale, allowing novel data-driven approaches including Artificial Intelligence (AI) capable of learning patterns from these data. This can identify key factors that drive improvement or decline in the health and function of older people during recovery from acute illness. An explainable AI tool to assist healthcare professionals to treat patients at risk of hospital-acquired disability would be of major benefit and would help target individualised patient rehabilitation.

Public Health Research

19/06/2023

De-Personalised

TRE