In both research and healthcare, we often consider and treat single diseases. However, as populations age across countries, multimorbidity, where two or more diseases occur together, is becoming more common. Understanding the patterns of multimorbidity, particularly in older individuals, for specific diseases will help to prevent and manage ill health at individual and population levels. We will investigate the risk factors and trajectories of individuals with five common diseases with different risk factors, treatments and approaches at health system level: hip fracture, acute myocardial infarction, ischaemic stroke, elective aortic aneurysm repair, and heart failure using electronic health records. We will use both traditional approaches like epidemiology (describing risk factors and patterns of disease using statistics) as well as machine learning, which uses artificial intelligence to learn patterns from the data with or without the need to define them by labelling information. We will describe differences in risk factors, hospitalisation and death rates across the five diseases, looking at the impact of socioeconomic status and multimorbidity. We will use machine learning to try to understand how if there are particular patterns of multimorbidity in each of the five diseases. By looking across diseases, rather than at diseases individually, we will be able to better describe the burden of multimorbidity in older populations. This research may improve prediction of the risk of particular diseases or clusters of diseases, as well as providing insights into possible prevention strategies.
The single disease model is not suited to deal with ageing populations and multimorbidity. New characterisations of diseases are required across disease silos to plan appropriate public health and policy responses. We have selected five diseases (hip fracture, acute myocardial infarction; ischaemic stroke, elective aortic aneurysm repair, and heart failure) to reflect a range of different prevention, treatment and health system approaches to gain a more complete picture across the health system. The project intends to work on complex data to better understand how to define risk factors, trajectories, outcomes and discover sub-phenotypes of multimorbidity across hip fracture, acute myocardial infarction; ischaemic stroke, elective aortic aneurysm repair, and heart failure. In addition to traditional epidemiologic analyses, unsupervised machine learning approaches such as clustering have been adopted with the intention of recognising subtypes for single diseases and for multimorbidity involving these five index diseases. Therefore, we aim:
1: To describe differences in the epidemiology of hip fracture, acute myocardial infarction; ischaemic stroke, elective aortic aneurysm repair, and heart failure in individuals age > 65 years.
2: To compare in-hospital, 90-day, and 1-year mortality across diseases and in multimorbidity.
3: To compare treatment intensity (e.g., length-of-stay, readmissions, post-acute-care utilisation) and types of treatment (e.g., consultations, outpatient visits, imaging, procedure use) for each condition.
4: Compare epidemiology, outcomes, and utilisation by socioeconomic status and multimorbidity.
Hip fracture
Myocardial infarction
Ischaemic stroke
Elective aortic aneurysm repair
Heart failure
Cardiovascular disease
Death (all cause, HF-related, cardiovascular-related composite endpoint)
GP appointments
Outpatient appointments (all-cause)
Emergency department attendances (all-cause)
Hospital admissions and readmission (all-cause, HF-related, cardiovascular-related composite endpoint)
Hospital admission length (all cause, HF-related, cardiovascular-related composite endpoint)
Days alive out of hospital
Percutaneous coronary interventions
Coronary artery bypass graft surgery
Heart transplant surgery
Amitava Banerjee - Chief Investigator - University College London ( UCL )
Laura Pasea - Corresponding Applicant - University College London ( UCL )
Arturo Gonzalez-Izquierdo - Collaborator - University College London ( UCL )
Folkert Asselbergs - Collaborator - University College London ( UCL )
Harry Hemingway - Collaborator - University College London ( UCL )
Kenan Direk - Collaborator - University College London ( UCL )
Mohamed Mohamed - Collaborator - University College London ( UCL )
Muhammad Qummer ul Arfeen - Collaborator - University College London ( UCL )
Spiros Denaxas - Collaborator - University College London ( UCL )
Suliang Chen - Collaborator - University College London ( UCL )
HES Accident and Emergency;HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation