Real-Time Voice AI Virtual Patient Trial for Clinical Communication Training

A randomized controlled study of 134 medical residents compared communication training outcomes using real-time voice generative AI-based virtual patients (RT-VP), standardized patients (SP), and peer role-play (PR). The RT-VP was developed on the Doubao platform and demonstrated SPIKES communication performance scores (24.9 ± 2.9) comparable to SP (26.6 ± 3.0), significantly outperforming PR (20.4 ± 4.3). This research addresses the global challenge of scalable, authentic communication training in clinical education.

Use Cases

• Comparing the efficacy of AI-based versus traditional simulation modalities for communication skills training based on SPIKES scores
• Analyzing perceived realism across linguistic, contextual, emotional, and engagement dimensions for different training methods
• Evaluating self-efficacy outcomes in medical residents after practicing with different simulation modalities
• Assessing the scalability of real-time voice generative AI platforms for clinical dialogue simulation

Strengths

• Dataset originates from a randomized controlled study with 134 participants, providing a structured experimental basis
• Includes detailed outcome measures: communication performance (SPIKES scores), self-efficacy, and multi-dimensional perceived realism
• Cross-platform presence on figshare suggests the dataset has undergone some level of curation and is publicly shared

Limitations

• Column names and specific data structure are not provided, limiting understanding of the dataset's internal organization
• The dataset size (7,974,001 bytes) is given, but the number of rows or observations is unknown
• The description is rich in study context but sparse on the actual variables and measurements contained in the data files

Provenance

Source

Wei Hu

Collection Method

Data collected from a randomized controlled trial where 134 medical residents were assigned to RT-VP, SP, or PR training groups.

Time Range

null

Freshness

Last updated on 2026-03-19

Geography

null