Machine learning and artificial intelligence: applications in healthcare epidemiology

Area of Science:

• Computational epidemiology
• Health informatics
• Machine learning applications in medicine

Background:

• Artificial intelligence (AI) enables machines to perform tasks typically requiring human intelligence.
• Machine learning (ML), a subset of AI, allows computers to learn from data without explicit programming.
• ML is gaining traction in healthcare epidemiology for its potential to enhance disease prediction and patient care.

Purpose of the Study:

• To provide an overview of machine learning in healthcare epidemiology.
• To present practical examples of ML tools supporting healthcare decision-making.
• To illustrate ML applications across four hospital care stages: triage, diagnosis, treatment, and discharge.

Main Methods:

• Review of machine learning applications in healthcare epidemiology.
• Identification and description of ML tools used in hospital-based care.
• Analysis of ML model-building efforts for specific clinical scenarios.

Main Results:

• ML tools assist in emergency department triage and predicting septic shock onset.
• ML aids in detecting community-acquired pneumonia and classifying COVID-19 risk.
• Examples demonstrate ML's utility in clinical decision support.

Conclusions:

• Increasing electronic health record (EHR) data and computing power facilitate ML adoption.
• ML offers opportunities to improve patient safety and clinical management efficiency.
• ML has the potential to reduce overall healthcare costs.