Hybrid Machine Learning Method and Standard Data Analysis Approaches for Predicting Treatment Outcomes of Cardiovascular Diseases: A Randomized Controlled Trial.

Background: This study aimed to assess the effectiveness of data analysis using a hybrid method in predicting and diagnosing cardiovascular diseases compared to standard methods. Methods. The study involved 200 patients diagnosed with cardiovascular diseases (arterial hypertension, ischemic heart disease, heart failure) in Moscow, Russia. Patients were randomly assigned to two equal groups. Group A underwent analysis of clinical data using random forest, support vector machines, and linear regression methods. Group B was subjected to hybrid method analysis. For Group B, patient survival was higher by 5%, and complication frequency was lower by 3%. The hybrid method demonstrated superior forecasting and treatment efficacy (p < 0.001) compared to similar indicators in Group A. Results: PCA analysis revealed that principal components explained over 70% of the variability among clinical parameters. Kaplan-Meier survival curves showed a statistically significant influence of cholesterol levels on survival and complication frequency (p < 0.05). Correlation analysis identified an inverse relationship between cholesterol levels and survival (p < 0.05). A hybrid data analysis method proves more effective than standard methods in predicting cardiovascular treatment outcomes and improving patient survival. The use of a hybrid method demonstrates the success of new data processing techniques in clinical practice, enabling the optimization of therapies and improving the quality of care for patients with cardiovascular disease. Conclusion: The use of a hybrid method demonstrates the success of new data processing techniques in clinical practice, enabling the optimization of therapies and improving the quality of care for patients with cardiovascular disease.