Artificial Intelligence in Predictive Healthcare: A Systematic Review

Area of Science:

• Healthcare AI
• Machine Learning in Medicine
• Predictive Analytics

Background:

• AI and ML significantly enhance healthcare predictive analytics.
• Data-driven approaches and AI integration are gaining traction in clinical settings.
• Growing publications highlight innovative AI applications in healthcare.

Purpose of the Study:

• Synthesize recent evidence on AI/ML applications in disease prediction.
• Identify common AI/ML models, evaluation metrics, and challenges.
• Provide insights into the current state and future directions of AI in healthcare.

Main Methods:

• Systematic literature review from 2021-2025.
• Searches conducted on Web of Science and Google Scholar.
• Included studies focused on AI/ML techniques for disease prediction.

Main Results:

• Twenty-two studies met the inclusion criteria.
• Tree-based ensembles (Random Forest, XGBoost) and deep learning (CNN, LSTM) were prevalent.
• AUROC, F1-score, accuracy, and sensitivity were common evaluation metrics.
• Key challenges include data privacy, workflow integration, interpretability, and data quality.

Conclusions:

• Future research should focus on interpretable AI models for clinical trust.
• Robust privacy-preserving techniques are essential for safeguarding patient data.
• Standardized evaluation frameworks are needed to assess model performance effectively.