An AI-Driven Virtual Patient Platform (CBT Trainer) for Training Cognitive Behavioral Therapy Practitioners Against Competencies: Mixed Methods Pilot Study

TL;DR

A new AI-based virtual patient platform called CBT Trainer was tested as a tool to train cognitive behavioral therapy practitioners, showing promise in improving skills through real-time feedback and accessible practice.

Contribution

The CBT Trainer is the first virtual patient platform to provide real-time feedback aligned with established CBT competence frameworks.

Findings

Participants engaged with the platform for an average of 95 minutes and rated platform usability as excellent (mean SUS score of 82.20).

Self-reported improvements were highest in assessment skills (96.7%) and information gathering (66.7%).

Qualitative feedback highlighted strengths in competency-aligned feedback and a psychologically safe practice space.

Abstract

Cognitive behavioral therapy (CBT) training faces significant challenges, including supervised practice with diverse cases, inconsistent feedback, resource-intensive supervision, and difficulties standardizing competence assessment.

This study evaluated the acceptability and feasibility of CBT Trainer (TTZ), the first virtual patient platform to provide real-time feedback aligned with established competence frameworks. The mobile app trains psychological practitioners using standardized artificial intelligence patient interactions and the evaluation of therapist responses against competence frameworks to enable structured skill development in a controlled, repeatable environment that complements traditional training methods.

This mixed methods pilot study used a 2-stage approach. Stage 1 involved usability testing with 4 participants. Stage 2 included 59 participants from psychological practitioner training programs (a Low Intensity CBT Interventions Program and a Doctorate in Clinical Psychology) who engaged with the CBT Trainer voluntarily for over 1 month. Measures of impact included the System Usability Scale (SUS), platform naturalistic engagement, poststudy questionnaire on perceived competency development, comparative evaluation against traditional role-play, and qualitative feedback.

Participants engaged voluntarily with the platform for an average of 95.24 (SD 134.58; median 45.34, IQR 11.57–105.15) minutes of active role-play. Platform usability was rated as excellent (mean SUS 82.20, SD 12.93). Self-reported competence improvement improved most in assessment skills (96.7%) and information gathering (66.7%). When compared to traditional peer role-play exercises, participants rated CBT Trainer moderately favorably (mean 5.90/10, SD 1.94). Qualitative feedback highlighted strengths in competency-aligned feedback, on-demand accessibility, and a psychologically safe practice space.

This pilot study provides evidence that an artificial intelligence–based patient simulation shows promise as a supplementary training tool for psychological therapists who use CBT in their practice, particularly regarding accessibility and immediate feedback. Future research should use randomized controlled designs with objective competence assessments.

Figures