Cardiovascular disease (CVD), comprising mostly heart attacks and strokes, is one of the UK’s leading causes of death and disability. It is far better and cheaper to prevent CVD than to treat patients after they get sick. Consequently, doctors aim to identify patients at high risk of future CVD and offer healthy lifestyle advice and medication, such as statins.
However, studies have found that such advice is poorly communicated to patients. This has resulted in low numbers of patients choosing to make healthy lifestyle changes and take medication, especially amongst patients living in more deprived areas. It is important to improve the way CVD risk is communicated to patients across the whole of society so that more people benefit from advice and medication, and also to reduce inequalities in CVD.
Currently, doctors use risk prediction calculators to help decide whether a patient is at high risk of future CVD. The most commonly used risk prediction calculators in England were developed using data from 2004-2016 from 2.3 million patients in England. More current data is now available from over 56 million patients in England, as well as 10 million patients from Wales, Scotland and Northern Ireland. These datasets also include information about whether a patient has had COVID-19 and any related complications. Patients with COVID-19 complications may be at higher risk of future CVD than patients without COVID-19 complications. Therefore, COVID-19 information may be useful information for doctors in helping to identify the right patients at high risk of CVD.
We plan to validate existing risk prediction models and, where necessary, develop new risk prediction tools to assess and help communicate a patients’ future risk of CVD. Our aims are (i) to better identify patients at high risk of CVD in the UK before, during and post the COVID-19 pandemic and (ii) to improve communication of CVD risk to patients across the whole of society. Ultimately, this should reduce CVD in the UK.
Outputs
A nationwide deep learning pipeline to predict stroke and COVID-19 death in atrial fibrillation