The first AI-based Chatbot to promote HIV self-management: A mixed methods usability study.

Area of Science:

• Digital Health
• Artificial Intelligence in Healthcare
• HIV Self-Management

Background:

• MARVIN is an AI-powered chatbot designed to offer continuous, expert-validated information for individuals managing HIV.
• The study aimed to evaluate the feasibility, usability, and acceptability of MARVIN.

Purpose of the Study:

• To assess the feasibility of integrating MARVIN into HIV self-management.
• To evaluate MARVIN's usability, including perceived ease of use, usefulness, attitude, and behavioral intention.
• To determine the acceptability of MARVIN among users.

Main Methods:

• A mixed-methods approach was employed at the McGill University Health Centre.
• Participants engaged in 20 conversations with MARVIN over three weeks and completed usability questionnaires.
• Qualitative data was gathered through semi-structured interviews to identify barriers and facilitators.

Main Results:

• Study retention was 70% with 28 participants surveyed and nine interviewed.
• MARVIN demonstrated high usability (mean 69.9/68) and acceptability (mean 23.8/24).
• Facilitators included real-time information, accessibility, and perceived emotional safety; barriers involved limited comprehension and platform constraints.

Conclusions:

• MARVIN exhibits significant global usability for HIV self-management.
• Findings support MARVIN's potential as a digital health tool for people with HIV.
• Recommendations for future development include improving comprehension and feature sets.