Heart failure (HF) is widespread and affects up to 15 million people in Europe. It is a complex clinical syndrome presenting with reduced heart function. Individuals diagnosed with HF have poor quality of life. HF is an expensive disease to manage and requires engagement of entire teams of healthcare professionals. There is a high clinical demand for novel techniques to be developed to assist in early diagnosis and prediction of HF. Artificial intelligence (AI) has enormous potential in the medical field for early and accurate prediction of HF and its progression.
Such tools are essential to allow early initiation of prevention and treatment strategies which will improve patient’s quality of life and reduce overall HF burden on patients and healthcare. STRATIFYHF (EU Horizon-, UKRI-funded project) aims to develop, validate and implement novel AI-methods to advance early prediction of HF and to detect those who may already have disease but do not know they have it. The AI methods will integrate patient-specific demographic and clinical data using existing and novel technologies and develop AI-based tools (computer software) for predicting HF. STRATIFYHF project aims to change the way in which HF is diagnosed today and thereby improve the quality and length of patients’ lives and lead to efficient and sustainable healthcare systems.
This study aims to address the global challenge of identifying individuals at risk of HF and predicting prognosis of those living with HF. A retrospective cohort study will be conducted from 1st January 2004 to 31st January 2024. We aim to retrospectively identify two populations cohorts (patients diagnosed with HF, patients at-risk of HF (low and high)) using SMOMED-CT clinical codes and will construct a predictive model (Decision Support System, DSS, STRATIFYHF) for risk prediction of HF and prognosis of those living with HF using sex, age, ethnicity and linked to HES Admitted Patient Care for hospitalisations, HES Diagnostic Imaging Dataset for frequency of imaging, HES Outpatient for frequency of appointments and ONS Death Registration. The statistical methods will include advanced models such as Random Forest, XGBoost and Neural Networks. Both population cohorts will be identified and permanently registered on CPRD for >5 years. The primary outcome for the patients diagnosed with HF is hospitalisation rate during the five years of diagnosis. Secondary outcomes for patients diagnosed with HF will include: incidence rates for comorbidities, frequency of HF symptoms, frequency of appointments, prescriptions and mortality. The primary outcome for patients at risk (low and high) of HF is diagnosis of HF.
This study and the development of the DSS (STRATIFYHF) may improve risk prediction and prognosis of HF thus enhancing patient’s quality of life and leading to a more cost-effective health care system.
The primary outcome for the patients diagnosed with HF is hospitalisation rate during the five years of diagnosis.
The secondary outcomes are:
2. Heart failure symptoms (including the presence of breathlessness, fatigue, or ankle swelling, as recorded in primary care records using SNOMED CT symptom codes)3. Frequency of primary care clinical appointments per year (including 4+ minute GP appointment, 4+ minute practice nurse appointment, home visit appointment, out of hours appointment)
7. Mortality using the ONS Death Registration Data
The primary outcome for the patients at high and low risk of HF is diagnosis of HF.
There are no secondary outcomes for the patients are high and low risk of HF.
Sarah Charman - Chief Investigator - Newcastle University
Sarah Charman - Corresponding Applicant - Newcastle University
Amy Fuller - Collaborator - Coventry University
David Sinclair - Collaborator - Newcastle University
Djordje Jakovljevic - Collaborator - Coventry University
Duncan Edwards - Collaborator - University of Cambridge
Jonathan Mant - Collaborator - University of Cambridge
Matej Pičulin - Collaborator - University of Ljubljana
Nduka Okwose - Collaborator - Coventry University
Petar Vračar - Collaborator - University of Ljubljana
Zoran Bosnic - Collaborator - University of Ljubljana
Bogdan Milicevic - Collaborator - BIOIRC - Bioengineering Research and Development Center
Nenad Filipovic - Collaborator - BIOIRC - Bioengineering Research and Development Center
Tijana Geroski - Collaborator - BIOIRC - Bioengineering Research and Development Center
HES Admitted Patient Care;HES Diagnostic Imaging Dataset;HES Outpatient;ONS Death Registration Data;Practice Level Index of Multiple Deprivation;CPRD Aurum Ethnicity Record