An Artificial Intelligence Approach to Guiding the Management of Heart Failure Patients Using Predictive Models: A...

Area of Science:

• Cardiology
• Biomedical Informatics
• Data Science

Background:

• Heart failure (HF) is a major global cause of mortality and hospitalizations.
• Existing predictive models for HF have shown limited accuracy.
• Accurate prediction of mortality and readmission is vital for patient management.

Purpose of the Study:

• To evaluate machine learning (ML) predictive models in heart failure (HF) patients.
• To assess the external validation of these ML models.
• To compare the performance of ML models against traditional risk scores.

Main Methods:

• Systematic review of studies using HF patient data for predictive model development.
• Literature search across four databases and reference lists.
• Analysis of models predicting mortality, rehospitalization, treatment response, and medication adherence.
• Comparison of Area Under the Receiver Operating Characteristic Curve (AUC) values.

Main Results:

• Nine studies were included in the final analysis.
• ML models achieved AUCs ranging from 0.6494 to 0.913 in independent datasets.
• Traditional statistical scores had AUCs ranging from 0.622 to 0.806.
• External validation of ML models in HF populations remains infrequent.

Conclusions:

• Machine learning models demonstrate superior performance over conventional risk scores for HF.
• ML approaches hold significant potential for improving heart failure management.
• Further research should focus on the external validation of ML models in diverse HF cohorts.