Improving and Evaluating eMMA's Communication Skills: A Chatbot for Managing Medication

Area of Science:

• Health Informatics
• Human-Computer Interaction
• Artificial Intelligence

Background:

• A mobile application for medication self-management (eMMA) previously featured a basic conversational user interface (CUI).
• Enhancing patient engagement and adherence in medication management is a critical healthcare challenge.

Purpose of the Study:

• To extend the conversational user interface (CUI) of the eMMA application by integrating the RiveScript chatbot framework and an instruction interface.
• To evaluate the impact of these technical extensions on task success, dialog quality, and efficiency in medication self-management.

Main Methods:

• A theoretical evaluation was conducted.
• A quantitative evaluation was performed.
• A usability test was carried out to assess user experience and system performance.

Main Results:

• The technical extensions, including RiveScript integration, demonstrably improved the chatbot's dialog quality.
• Evaluations indicated that the enhancements were beneficial for the conversational aspect of the application.
• Despite improvements, the system's knowledge base requires significant expansion for real-world application.

Conclusions:

• The integration of RiveScript and an instruction interface offers a promising pathway to enhance conversational agents for medication self-management.
• While technical improvements are evident, the system's practical utility is currently limited by the scope of its knowledge base.
• Future work should focus on substantial knowledge base augmentation to enable widespread adoption of the eMMA system.