Using ECGs in an AI Model to Predict Clinical Outcomes in CHD Patients
Research type
Research Study
Full title
Can ECG’s be used in an Artificial Intelligence Model to Predict Clinical Outcomes in Adults with Congenital Heart Disease?
IRAS ID
335717
Contact name
Gruschen Veldtman
Contact email
Sponsor organisation
Golden Jubilee National University Hospital
Clinicaltrials.gov Identifier
N/A, N/A
Duration of Study in the UK
5 years, 0 months, 1 days
Research summary
Congenital heart disease (CHD) results from developmental structural abnormalities of the heart or great vessels, and is among the most common congenital disorders. There has been a steady reduction in CHD deaths over the past century, resulting in a growing adult CHD population requiring lifetime care. To provide optimal care while managing the demand on services, risk stratification is used to identify and predict the risk level of patients. Currently, it is hard to quantify risk within the CHD population because of limitations in generally available risk stratification tools. Research has produced promising support for the use of artificial intelligence (AI) and electrocardiograms (ECG) in risk stratification. AI is a computer program that performs at the level one would expect of a human mind. An ECG is a test that records the hearts electrical rhythm: it is simple to conduct, is used widely across healthcare, and is a powerful tool to identify poor cardiovascular outcomes. However, the interpretation of ECGs require expertise - that is where support tools arise. In non-CHD populations, AI has identified and predicted outcomes from ECGs more accurately than current support tools. This study aims to explore the use of AI for predicting cardiovascular outcomes in the complex field of CHD. This will be a retrospective study in which relevant clinical data of all CHD patients under follow up care at the Golden Jubilee National Hospital will be collected. The cohort will be randomly split into training and test data, with the training data used to train an AI model to predict outcomes, and the remaining test data used to test the AIs prediction abilities. This research could contribute towards revolutionising the care of CHD patients by providing clinicians with a more efficient and accurate tool to support the diagnosis, prognosis, treatment and management of CHD.
REC name
South Central - Berkshire B Research Ethics Committee
REC reference
23/SC/0436
Date of REC Opinion
5 Feb 2024
REC opinion
Further Information Favourable Opinion