Cardiovascular diseases (CVD) are one of the main drivers for death world-wide. The population within the UK is ageing and as a result, atrial fibrillation and/or heart failure are frequently becoming an additional factor for patients with other current diseases.
Currently computational approaches designed to calculate risk of further CVD hospitalisation, have not translated into clinical practice as readily as desired by clinicians. This is predominantly due to the lack of incorporating multiple different data types into model based approaches. Clinicians rely on a combination of data sources to arrive at a risk assessment for patients’ long term health. In order for computational models to be used and trusted by clinicians, there is a need for incorporation of multiple different data sources.
Many works have been produced for high quality assessment of individual data types but research into an incorporated approach is infrequent. The hopeful impact of this project will be wider spread usage of artificial intelligence (specifically computer based programmes designed to aid in clinical decision making) within the healthcare system, leading to improved identification of patients at risk of long term CVD hospitalisation as well as improvement to clinical workflow. Benefit to patients will be through targeted care, specific to their individual condition allowing for a reduction in risk of adverse CVD related events.
In order to identify patients at future risk of CVD, we will be using ultrasound scans of the heart in combination with clinical data as well as electrocardiogram data (signals indicating the electrical pulse of the heart). These will be supplied from the 4 hospitals as part of the UHB trust: Good Hope, Heartlands, Solihull and Queen Elizabeth Hospital. We will be developing novel computational approaches that exploit the type of data discussed above to identify groups of patients with similar features.
We anticipate that this will allow us to calculate additional risks related to long term CVD hospitalisation for these groups and hence stratify patients into priority groups for treatment and long term care.