Computational medicine.

Heart Rate Simulation Dataset

This project generates synthetic patient datasets focused on heart rate (HR) and clinical parameters that may influence heart rate, including age, sex, ejection fraction (EF), NYHA functional class, heart failure phenotype, physical activity, and other variables.

The relationships between variables are currently modeled empirically to create realistic educational and research-oriented datasets. The simulator allows users to explore how different clinical factors may affect heart rate distributions and correlations.

The model includes adjustable effect sizes, enabling users to modify the strength of associations between variables and investigate different physiological and pathological scenarios.

Potential applications include:

• Learning statistics and data science
• Teaching clinical epidemiology
• Practicing data visualization and machine learning
• Testing statistical methods
• Generating synthetic datasets for hypothesis development
• Exploring potential relationships between heart rate and clinical variables

To interactive app.

Scientific background

Numerous studies have demonstrated statistically significant associations between heart rate and individual clinical variables, such as age, left ventricular ejection fraction, physical activity, heart failure severity, and pharmacological treatment. While correlation coefficients or regression estimates for individual predictors have been reported, a comprehensive multivariable mathematical model integrating the major determinants of heart rate has not yet been established.

The long-term objective of this project is to develop and validate such a model using clinical data. Conceptually, heart rate may be represented as a multivariable regression model:

HR = β₀ + β₁X₁ + β₂X₂ + ⋯ + β_nX_n + ε,

HR is the predicted heart rate,

β₀ is the intercept,

β₁n are regression coefficients,

X₁ … X_n represent relevant physiological and clinical variables (e.g., age, ejection fraction, stroke volume, hemoglobin concentration, physical activity, medication, and heart failure phenotype),

ε denotes the unexplained residual error.

Such a validated model could have important clinical applications, including the assessment of chronotropic competence, optimization of pacemaker programming, and quantitative evaluation of heart rate responses under different physiological and pathological conditions.