EUVIMEDEUVIMED entdecken
Crossref

Detecting Therapist Burnout and Flourishing From Brief Speech: A Machine‐Learning Approach

Matteo Bugatti, Seungbin Oh, Emma Hills, Colby Schramel, Patty Kuo, Christina S. Soma, Jordan Pruett, Jesse Owen

Counselling and Psychotherapy Research · 2026-12 · 26 (4)

DOI: 10.1002/capr.70246 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Counselling and Psychotherapy Research, online erschienen am 11.11.2026. Verfasst von Matteo Bugatti und 7 weiteren Autor:innen. Die Arbeit nennt 83 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Burnout is pervasive among therapists and is associated with impaired well‐being and reduced clinical effectiveness. Although flourishing and character strengths may function as protective factors, intervention implementation is limited by a lack of feasible, scalable methods for monitoring therapist well‐being. Machine‐learning (ML) approaches may be suited to address these limitations.

Aims

This proof‐of‐concept study examined the use of a transformer‐based natural language processing (NLP) model to detect therapist burnout, flourishing, mental health continuum, employee engagement and character strengths (humility, gratitude and forgiveness) from automatically transcribed therapists' verbal responses. Materials and Methods Participants ( n = 365) included licensed and training therapists, who completed self‐report measures of burnout, flourishing, mental health continuum, employee engagement and character strengths (humility, gratitude and forgiveness). A transformer‐based RoBERTa model was trained, tuned and evaluated using separate training, validation and test sets, with performance assessed using Spearman rank correlations with permutation tests.

Results

Moderate correlations were observed for burnout ( ρ = 0.456), mental health continuum ( ρ = 0.46) and flourishing ( ρ = 0.387).

Discussion

These findings suggest that transformer‐based NLP models might be able to detect therapist burnout, flourishing and some positive functioning indicators relevant to organisational well‐being models. Limitations include the relatively small and potentially unrepresentative cross‐sectional sample, possible transcription and prompt‐related measurement error, and the absence of external validation, warranting cautious interpretation of these findings as preliminary evidence of feasibility rather than as definitive estimates of detection performance. Research should also examine ethical concerns related to privacy, bias, misclassification and potential misuse.

Conclusion

This study provides preliminary evidence supporting the use of transformer‐based NLP models for the detection of therapist burnout, flourishing and other positive functioning indicators.

Zitieren

APA

Bugatti, M., Oh, S., Hills, E., Schramel, C., Kuo, P., Soma, C. S., Pruett, J., & Owen, J. (2026). Detecting Therapist Burnout and Flourishing From Brief Speech: A Machine‐Learning Approach. Counselling and Psychotherapy Research, 26(4). https://doi.org/10.1002/capr.70246

Vancouver

Bugatti M, Oh S, Hills E, Schramel C, Kuo P, Soma CS, et al. Detecting Therapist Burnout and Flourishing From Brief Speech: A Machine‐Learning Approach. Counselling and Psychotherapy Research. 2026;26(4). doi:10.1002/capr.70246

BibTeX
@article{Bugatti2026,
  author = {Bugatti, Matteo and Oh, Seungbin and Hills, Emma and Schramel, Colby and Kuo, Patty and Soma, Christina S. and Pruett, Jordan and Owen, Jesse},
  title = {Detecting Therapist Burnout and Flourishing From Brief Speech: A Machine‐Learning Approach},
  journal = {Counselling and Psychotherapy Research},
  year = {2026},
  volume = {26},
  number = {4},
  doi = {10.1002/capr.70246}
}
Kostenlos starten →