Heart failure, diabetes mellitus and chronic kidney disease are major causes of global disease and death, both individually and in combination. They frequently coexist, but are studied separately in terms of what increases their risks and outcomes following diagnosis. Better understanding of the effects of combinations of these three diseases may aid earlier diagnosis and prevention.
We will investigate the characteristics and outcomes of individuals with heart failure, diabetes mellitus and chronic kidney disease using electronic health records, by both traditional epidemiologic approaches as well as machine learning. Machine learning uses artificial intelligence to learn patterns from the data with or without the need to define them by labelling information. It has been used to identify novel disease definitions, but usually restricted to single diseases and smaller samples which may be unrepresentative of the whole population. We will use machine learning to see if new subtypes of the diseases can be discovered in patients with overlapping heart failure, diabetes mellitus and chronic disease. Machine learning will be used to assign patients into groups according to their clinical features (e.g. diagnoses, lab results). This research may improve prediction of the risk of particular outcomes, as well as providing novel insights into mechanism of action of current and new therapies.
The project intends to work on complex data to better understand how to define phenotypes and discover sub-phenotypes of three common diseases: heart failure (HF), diabetes mellitus (DM) and chronic kidney disease (CKD). In addition to traditional epidemiologic analyses, unsupervised machine learning approaches such as clustering have been adopted with the intention of recognising disease subtypes. However, those analyses are mostly confined to individual diseases and specific patient sub-groups, which may lead to inaccurate categorisation and consequently inaccurate diagnostics. Furthermore, these three diseases are frequently risk factors for each other, and sometimes co-exist but the overlap syndrome of all three is understudied. Therefore, we aim:
- To study baseline characteristics and outcomes of this overlap syndrome and outcomes, and compare with the individual diseases (HF, DM, and CKD).
- To describe the missed opportunities in guideline recommended care pathway leading to HF and/or CKD diagnosis in DM vs non-DM patients and the patients characteristics along the care pathway.
- To use combined unsupervised and supervised statistical learning to identify sub-phenotypes of HF, DM and CKD.
- To use machine learning (supervised, unsupervised or combined) to identify clusters in the overlap syndromes between HF, DM and CKD
Heart failure
Diabetes mellitus
Chronic kidney disease
Dialysis
Renal transplant
Coronary artery disease
Cardiovascular disease
All-cause mortality
Cardiovascular mortality
GP appointments
Outpatient appointments
Emergency department attendances
Diagnostic imaging tests
Percutaneous coronary interventions
Coronary artery bypass graft surgery
Heart transplant surgery
Limb amputation
Jil Billy Mamza - Chief Investigator - AstraZeneca Ltd - UK Headquarters
Laura Pasea - Corresponding Applicant - University College London ( UCL )
Amitava Banerjee - Collaborator - University College London ( UCL )
Arturo Gonzalez-Izquierdo - Collaborator - University College London ( UCL )
Folkert Asselbergs - Collaborator - University College London ( UCL )
Gehan Lyu - Collaborator - University College London ( UCL )
George Godfrey - Collaborator - AstraZeneca Ltd - UK Headquarters
Harry Hemingway - Collaborator - University College London ( UCL )
Mehrdad Alizadeh Mizani - Collaborator - University College London ( UCL )
Mohamed Mohamed - Collaborator - University College London ( UCL )
Muhammad (Ashkan) Dashtban - Collaborator - University College London ( UCL )
Spiros Denaxas - Collaborator - University College London ( UCL )
Suliang Chen - Collaborator - University College London ( UCL )
Tamsin Morris - Collaborator - AstraZeneca Ltd - UK Headquarters
Tom Lumbers - Collaborator - University College London ( UCL )
WEI ZHOU - Collaborator - University College London ( UCL )
Xiaoyu He - Collaborator - University College London ( UCL )
Xiaoyu He - Collaborator - University College London ( UCL )
WEI ZHOU - Collaborator - University College London ( UCL )
Gehan Lyu - Collaborator - University College London ( UCL )
HES Accident and Emergency;HES Admitted Patient Care;HES Diagnostic Imaging Dataset;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation