Computer-Controlled Virtual Humans in Patient-Facing Systems: Systematic Review and Meta-Analysis

Area of Science:

• Health Informatics
• Human-Computer Interaction
• Digital Health

Background:

• Virtual humans (VH) are advanced computer-generated characters simulating human-like conversations with verbal and nonverbal cues.
• VH integrate conversational agent capabilities with interactive avatar features, distinguishing them from simpler chatbots or smart speakers.
• While VH show promise in patient-facing systems, their effectiveness in healthcare applications requires systematic evaluation.

Purpose of the Study:

• To systematically review the effectiveness of virtual humans in patient-facing health systems.
• To examine the design and implementation characteristics of virtual human systems used in healthcare.

Main Methods:

• A systematic review and meta-analysis of peer-reviewed articles identified through electronic bibliographic databases.
• Studies were included if they designed or evaluated VH in patient-facing systems; randomized controlled trials were used for meta-analysis.
• The PICOTS framework was utilized for summarizing studies, with random-effects models employed to calculate effect sizes and assess risk of bias.

Main Results:

• The review included 53 articles describing 33 unique VH systems, with 16 studies (26 trials) contributing to the meta-analysis.
• Meta-analysis of 44 primary outcomes revealed a significant positive effect of VH interventions (SMD = .166, P=.012), despite some heterogeneity (I²=49.3%).
• Most studies utilized personal computers for VH delivery, with a mix of cross-sectional and longitudinal designs.

Conclusions:

• Evidence supports the efficacy of virtual humans in patient-facing health applications.
• Future research should focus on more specific analyses considering diverse populations and outcomes.
• Identifying key features that contribute to the effectiveness of virtual human interventions is crucial for future development.