Using artificial intelligence to analyse and teach communication in healthcare

Area of Science:

• Medical Informatics
• Health Communication
• Artificial Intelligence in Healthcare

Background:

• Effective healthcare communication is crucial for patient and doctor outcomes but difficult to analyze and teach.
• Traditional human-based coding systems are time-consuming and expensive.
• There is a growing need for automated solutions in health communication analysis.

Purpose of the Study:

• To introduce health communication and its significance.
• To discuss the application of artificial intelligence (AI) in automating health communication coding.
• To review the evidence for AI coding reliability, validity, and applications in training.

Main Methods:

• Review of historical interaction coding systems.
• Discussion of machine learning algorithms (supervised and unsupervised) for automated coding.
• Analysis of studies on AI's reliability, validity, and application in healthcare communication.

Main Results:

• Machine learning enables accurate transcription and analysis of communication elements like prosody, emotion, and style.
• AI coding shows moderate to good reliability, often matching or exceeding human coders.
• AI has identified associations between communication variables and patient satisfaction.

Conclusions:

• AI, particularly machine learning, provides a reliable and valid method for analyzing health communication.
• AI applications in communication skills training, including feedback and avatar use, show potential.
• AI-based training may serve as a valuable supplement to traditional human-led training methods.