Modeling of Therapist-Client Dynamics in Psychotherapy Using Large Language Model–Based Assessments: Retrospective Observational Study of 1573 Psychotherapy Sessions
JMIR Mental Health · 05.11.2026 · 13 · e95389
DOI: 10.2196/95389 ↗Kurz gefasst
Beobachtungsstudie in der Zeitschrift JMIR Mental Health, online erschienen am 05.11.2026. Verfasst von Angela H Chen und 6 weiteren Autor:innen. Beteiligte Einrichtung: Carnegie Mellon University, Pittsburgh, PA, United States. Hauptthemen laut MEDLINE: Professional-Patient Relations, Psychotherapy. Die Arbeit nennt 61 Literaturangaben.
Abstract (Originalsprache)
Background
Psychotherapy is a primary treatment for many mental health conditions, yet the dynamic interplay among therapist behaviors, client responses, and the therapeutic relationship remains difficult to disentangle at scale. Traditional process research relies on labor-intensive human coding, limiting the ability to model moment-to-moment therapeutic dynamics across large datasets.
Objective
This study aimed to develop and validate a computational framework for modeling therapist-client interaction processes and to examine how therapist behaviors and relational qualities relate to client self-disclosure and emotional expression in psychotherapy sessions.
Methods
We developed automated assessment pipelines using large language models (LLMs) to measure therapist behaviors (empathy or exploration), relational quality (rapport), and client outcomes (self-disclosure, self-directed negative emotion, or outward-directed negative emotion). Model-generated scores were validated against human annotations. We then applied these measures to approximately 2000 hours of psychotherapy transcripts from the Alexander Street corpus. This yielded an estimation sample of 119,233 client utterances from 1573 sessions and 37 clients. Because utterances are nested within sessions and sessions within clients, we estimated a saturated multivariate structural equation model with SEs clustered at the session level. Results were confirmed in a multilevel regression model with random intercepts for session and client.
Results
Following comparative validation of 6 candidate language models, automated measures showed good to excellent agreement with human annotations (intraclass correlation coefficient [ICC]=0.60-0.81) for rapport (ICC=0.81), self-disclosure (ICC=0.78), empathy components (ICC=0.60-0.75), and emotion measurements (ICC=0.66-0.80). Structural equation modeling analyses revealed that therapist empathy and exploration were each associated with increased client disclosure (empathy: β=.086, 95% CI 0.078-0.095; exploration: β=.086, 95% CI 0.078-0.094; both P
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Carnegie Mellon University, Pittsburgh, PA, United States.
ORCID-Profile: Angela H Chen, Siwei Jin, Canwen Wang, Holly A Swartz, Tongshuang Wu, Robert E Kraut, Haiyi Zhu
Zitieren
APA
Chen, A. H., Jin, S., Wang, C., Swartz, H. A., Wu, T., Kraut, R. E., & Zhu, H. (2026). Modeling of Therapist-Client Dynamics in Psychotherapy Using Large Language Model–Based Assessments: Retrospective Observational Study of 1573 Psychotherapy Sessions. JMIR Mental Health, 13, e95389. https://doi.org/10.2196/95389
Vancouver
Chen AH, Jin S, Wang C, Swartz HA, Wu T, Kraut RE, et al. Modeling of Therapist-Client Dynamics in Psychotherapy Using Large Language Model–Based Assessments: Retrospective Observational Study of 1573 Psychotherapy Sessions. JMIR Mental Health. 2026;13:e95389. doi:10.2196/95389
BibTeX
@article{Chen2026,
author = {Chen, Angela H and Jin, Siwei and Wang, Canwen and Swartz, Holly A and Wu, Tongshuang and Kraut, Robert E and Zhu, Haiyi},
title = {Modeling of Therapist-Client Dynamics in Psychotherapy Using Large Language Model–Based Assessments: Retrospective Observational Study of 1573 Psychotherapy Sessions},
journal = {JMIR Mental Health},
year = {2026},
volume = {13},
pages = {e95389},
doi = {10.2196/95389}
}
EUVIMED