Why understanding your cardiovascular risk matters
Cardiovascular risk is influenced by a combination of factors rather than a single measurement. Blood pressure, cholesterol, blood glucose, physical activity, tobacco exposure, body composition, family history and other factors can all contribute to an individual's overall risk.
Many cardiovascular risk factors can change over time, while others cannot. Understanding the factors that contribute to risk can therefore provide a useful foundation for prevention, monitoring and informed conversations with healthcare professionals.
What powers HeartAware?
HeartAware brings together structured health information, predictive modelling and clear presentation to make cardiovascular risk easier to understand.
Structured health data
HeartAware brings relevant cardiovascular, health and lifestyle factors together into a structured assessment so relationships between multiple variables can be considered together.
Predictive machine learning
Machine-learning models can identify patterns across combinations of health variables and use learned relationships to estimate cardiovascular risk for new assessments.
Understandable risk insights
Complex predictive outputs are translated into understandable cardiovascular information designed to help users interpret their individual risk profile.
From assessment to insight
A structured pipeline connects the information provided by the user with predictive analysis and understandable results.
Assessment
Relevant cardiovascular, health and lifestyle information is collected through a structured assessment.
Data processing
The collected information is prepared into the appropriate format for the predictive modelling pipeline.
Prediction
A trained predictive model evaluates the assessment and generates a cardiovascular risk-related prediction.
Insight
The prediction is translated into understandable information that helps users become more aware of their cardiovascular risk profile.
Why machine learning matters
Cardiovascular risk is inherently multivariable. Modern risk prediction therefore goes beyond looking at a single measurement and considers combinations of established cardiovascular and health-related factors.
Machine learning provides a framework for learning statistical relationships from historical datasets and applying those learned relationships to new observations. Depending on the dataset and modelling objective, this can involve classification, regression and other supervised learning techniques.
The important part is not simply using a sophisticated algorithm. A useful cardiovascular prediction system also requires appropriate feature selection, data quality, validation, calibration and evaluation of whether the model provides meaningful predictive and clinical value.
Multivariable prediction
Cardiovascular risk models can integrate multiple variables to produce an overall estimate rather than treating every measurement independently.
Validation matters
Predictive performance should be evaluated using appropriate validation strategies, calibration, discrimination and other measures relevant to the intended application.
Explore the science behind cardiovascular risk
HeartAware sits within a broader field of cardiovascular prevention, risk prediction and data-driven healthcare. Explore these resources to learn more about the science behind cardiovascular risk assessment.
Cardiovascular diseases
An overview of cardiovascular disease, major risk factors, prevention and the importance of understanding cardiovascular health.
Read articleEvaluating cardiovascular risk prediction models
Scientific guidance covering predictive performance, clinical utility and the evaluation of cardiovascular risk prediction models.
Read articlePREVENT cardiovascular risk equations
A contemporary example of multivariable cardiovascular risk prediction incorporating cardiovascular, kidney and metabolic health factors.
Read articleUnderstand your risk for heart disease
An evidence-based overview of cardiovascular risk factors and why knowing your risk can support prevention and conversations with healthcare professionals.
Read articleBuilt around responsible use
Predictive models are designed to estimate patterns and support cardiovascular risk awareness. Their outputs should be interpreted within the appropriate healthcare context and should not be treated as definitive medical diagnoses.
Developed by Nightingale Heart Project
Nightingale Heart Project develops artificial intelligence and machine-learning technologies focused on cardiovascular wellness. Its work explores how data-driven methods can be applied to cardiovascular health, risk assessment and digital health innovation.
Visit Nightingale Heart Project