Technical Aspects of Developing Chatbots for Medical Applications: Scoping Review.

Area of Science:

• Computer Science
• Medical Informatics
• Artificial Intelligence

Background:

• Chatbots facilitate natural language conversations for various applications, including critical patient information access in healthcare.
• The evolution of chatbots in medicine spans decades, with continuous development in diverse methodologies.

Purpose of the Study:

• To explore technical aspects and development methodologies of medical chatbots.
• To identify optimal development methods and guide future research in medical chatbot development.

Main Methods:

• A comprehensive literature search was conducted across 8 databases, supplemented by reference checking.
• Study selection involved single reviewer with dual review for 50% of studies; a narrative approach synthesized results.
• Chatbots were classified by technical aspects, identifying key components and implementation techniques.

Main Results:

• 45 studies were included, with English being the predominant communication language.
• Four core modules were identified: text understanding, dialog management, database layer, and text generation.
• Pattern matching was the most common technique for text understanding and dialog management, while fixed output was prevalent for text generation.

Conclusions:

• Medical chatbot development is rapidly increasing, with a notable shift towards machine learning approaches.
• Further research is needed to correlate clinical outcomes with specific chatbot development techniques and technical features.