Prediction models for early risk detection of cardiovascular event

Area of Science:

• Medical Informatics
• Computational Biology
• Cardiology

Background:

• Cardiovascular disease (CVD) is a leading global cause of mortality.
• CVD disproportionately affects low and middle-income countries, with nearly equal prevalence in males and females.
• Early detection of cardiovascular events is crucial for effective intervention and improved patient outcomes.

Purpose of the Study:

• To develop and evaluate computational models for the early prediction of cardiovascular events.
• To compare the performance of Bayesian Networks, Multilayer Perceptron, Radial Basis Function, and Logistic Regression models in CVD risk detection.
• To assess model accuracy using both combined and sex-specific heart attack datasets.

Main Methods:

• Utilized a dataset of 929 heart attack cases (626 male, 303 female).
• Constructed predictive models using Bayesian Networks, Multilayer Perceptron, Radial Basis Function, and Logistic Regression algorithms.
• Validated models on combined and separate male and female patient data.

Main Results:

• The Multilayer Perceptron model achieved the highest accuracy in predicting cardiovascular events.
• Performance variations were observed when models were tested on combined versus sex-specific datasets.
• All evaluated computational models provided a means for early cardiovascular risk assessment.

Conclusions:

• Computational models, particularly the Multilayer Perceptron, show significant promise for early cardiovascular disease risk detection.
• Sex-specific data may be important for optimizing the accuracy of CVD prediction models.
• Further research into advanced computational approaches can enhance cardiovascular event prediction and patient management.