Cardiovascular disease (CVD) is a broad term for disorders that affect the heart and circulatory systems, including major debilitating conditions such as heart attacks, stroke, and certain types of dementia. While changes can be made to lower the risk of developing these disorders, whether these be via medication (e.g. to reduce levels of fat in the blood) or behaviour (e.g. encouraging people to quit smoking) these may not result in the same level of risk reduction across individuals or apply equally to all types of disease.
In recent years, it has become apparent that while many of these illnesses share common risk factors, risk of disease is dynamic, not necessarily following a standard course. Some individuals with the same characteristics may develop a condition earlier in life than others or experience more than one disease, and amongst those with a specific diagnosis there will be differences in how well they respond to treatment. All of these clinical scenarios are likely influenced by risk factor trajectories (including how these are managed before disease develops) as well as the timing and order of events (e.g. experiencing a mini-stroke before a stroke).
Consideration of the entire patient journey from disease-free to diseased, and ultimately death, will open up opportunities for us to examine what set of conditions hasten or accelerate disease onset and progression (including any aspects of clinical care implemented), as well as whether these cluster, which may eventually lead to substantial improvements in patient wellbeing and population health.
We will examine associations of known and novel risk factors, plus their interaction, with the onset and prognosis of cardiovascular diseases (CVD) and dementia, evaluate the efficacy of existing medications (and any medications that have potential to be used off-label based on external bioinformatics analyses) to treat these conditions, and develop and evaluate (dynamic) models to enhance clinical prediction for risk and evolution of CVD and dementia. This includes within specific currently underserved patient groups (e.g. cancer survivors) and as yet unidentified subgroups. Statistical approaches that we will use include Cox proportional-hazards regression, linear and logistic regression models (including conditional or Prentice-weighted methods for case-cohort analyses), as well as more complex manifestations of these within a machine-learning framework (supervised and unsupervised methods, e.g. support vector machines, random forests, regression trees, ensemble/"super" learning, mixture models/latent class analysis, artificial neural networks and deep learning methods) to explore symptom/disease clustering and develop/refine risk prediction algorithms. Alongside primary care data obtained from CPRD, Hospital Episode Statistics (HES) data and Office of National Statistics (ONS) mortality data will be used to determine hospitalisations and deaths attributed to each outcome. We will use previously published code lists for products and diagnoses.
Abdominal aortic aneurysm; Atrial fibrillation; Atrioventricular block, complete; Atrioventricular block, first degree; Atrioventricular block, second degree; Cardiac arrest; Coronary heart disease not otherwise specified; Dementia; Dilated cardiomyopathy; Heart failure; Hypertrophic cardiomyopathy; Intracerebral haemorrhage; Ischaemic stroke; Myocardial infarction; Other cardiomyopathy; Peripheral arterial disease; Primary pulmonary hypertension; Pulmonary embolism; Secondary pulmonary hypertension; Stable angina; Stroke not otherwise specified; Subarachnoid haemorrhage; Subdural haematoma - nontraumatic; Supraventricular tachycardia; Thrombophilia; Transient ischaemic attack; Trifascicular block; Unstable angina; Venous thromboembolic disease (excluding pulmonary embolism); Ventricular tachycardia.
Hugh Markus - Chief Investigator - University of Cambridge
Steven Bell - Corresponding Applicant - University of Cambridge
Bernard Cho - Collaborator - University of Cambridge
Christopher Osuafor - Collaborator - University of Cambridge
Duncan Edwards - Collaborator - University of Cambridge
Elias Allara - Collaborator - University of Cambridge
Eric Harshfield - Collaborator - University of Cambridge
Joel Gibson - Collaborator - University of Cambridge
Luanluan Sun - Collaborator - University of Cambridge
Robin Brown - Collaborator - University of Cambridge
HES Accident and Emergency;HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation