Validation of the Usability and Effectiveness of a Thought Record Chatbot Based on a Large Language Model

Usability and Effectiveness Evaluation of a Large Language Model-Based Chatbot for Identifying Automatic Thoughts and Cognitive Distortions in CBT

Status

Active, not recruiting

Phases

Unknown

Study type

Interventional

Source

CRIS

Registry ID

KCT0012233

Enrollment

105

Registered

2026-07-06

Start date

2026-07-22

Completion date

Unknown

Last updated

2026-07-13

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

None listed

Interventions

Others(Digital intervention) : In this study, participants will be randomly assigned to one of three groups: the intervention group, mock-intervention group, or waitlist control group. The study will

Eligibility

Sex/Gender

All

Inclusion criteria

Inclusion criteria: 1.Individuals who voluntarily agree to participate in the study and provide written or online informed consent 2.Individuals who are able to use the chatbot program via a PC or smartphone 3.Adults aged 18 to 64 years 4.Individuals with mild or greater depressive and/or anxiety symptoms (PHQ-9 score of 5 or higher or GAD-7 score of 5 or higher) 5. Individuals who agree to the study procedures, questionnaire assessments, chatbot use, and data collection

Exclusion criteria

Exclusion criteria: 1.Individuals who have difficulty using a PC or smartphone 2.Individuals who have insufficient Korean language ability to understand the study procedures, complete questionnaires, or use the chatbot 3.Individuals with suicidal ideation 4.Individuals currently receiving mental health services 5.Individuals who do not agree to the chatbot use requirements or study participation conditions

Design outcomes

Primary

Measure	Time frame
Change in depressive symptom scores (Patient Health Questionnaire-9; PHQ-9);Change in anxiety symptom scores (Generalized Anxiety Disorder-7; GAD-7)	—

Secondary

Measure	Time frame
Change in self-reflection and insight scores (Self-Reflection and Insight Scale-Korean version; K-SRIS);Change in negative automatic thought scores (Short Form of the Korean Version of the Automatic Thoughts Questionnaire-Negative; KATQN-SF);Change in positive automatic thought scores (Short Form of the Korean Version of the Positive Subscale of the Automatic Thoughts Questionnaire-Revised; KATQ-RP-SF);Working alliance score with the chatbot (Korean version of Working Alliance Inventory-Short Revised; WAI-SR-K);CBT chatbot user experience score (Self-developed CBT chatbot user experience questionnaire);Chatbot usage patterns and interaction indicators (login records, content viewing records, conversation logs, and thought record entries);Chatbot performance in thought record completion and cognitive distortion classification (expert ratings, accuracy, sensitivity, specificity, F1 score, and Cohen’s kappa)	—

Measure

Time frame

Change in self-reflection and insight scores (Self-Reflection and Insight Scale-Korean version; K-SRIS);Change in negative automatic thought scores (Short Form of the Korean Version of the Automatic Thoughts Questionnaire-Negative; KATQN-SF);Change in positive automatic thought scores (Short Form of the Korean Version of the Positive Subscale of the Automatic Thoughts Questionnaire-Revised; KATQ-RP-SF);Working alliance score with the chatbot (Korean version of Working Alliance Inventory-Short Revised; WAI-SR-K);CBT chatbot user experience score (Self-developed CBT chatbot user experience questionnaire);Chatbot usage patterns and interaction indicators (login records, content viewing records, conversation logs, and thought record entries);Chatbot performance in thought record completion and cognitive distortion classification (expert ratings, accuracy, sensitivity, specificity, F1 score, and Cohen’s kappa)

—

Countries

Korea, Republic of

Contacts

Public ContactSeou Jung

Yonsei University

seouj0511@google.com+82-2-2123-7536

Outcome results