EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Can machine learning predict PTSD symptoms from trauma narratives of children and adolescents?

Alessandra Giuliani iD, Tamsin Sharp, Yeukai Chideya, Richard Meiser-Stedman iD, Mark Tomlinson iD, Sarah L Halligan

European Journal of Psychotraumatology · 2025

DOI: 10.1080/20008066.2025.2589709 ↗

Kurz gefasst

Fachartikel in der Zeitschrift European Journal of Psychotraumatology, 2025. Verfasst von Alessandra Giuliani und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Psychology, University of Bath, Bath, UK. Hauptthemen laut MEDLINE: Stress Disorders, Post-Traumatic, Machine Learning, Narration, Memory, Episodic. Die Arbeit nennt 58 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

Machine learning approaches are being increasingly tested as a potential means of identifying mental health conditions. Narrative features of trauma memories are proposed to play a significant role in the development of post-traumatic stress disorder (PTSD), meaning that trauma narratives provide an excellent context in which to test machine learning capabilities. The potential for children's trauma narratives to predict post-traumatic stress remains particularly poorly studied. Here, we tested whether the application of machine learning to trauma narrative characteristics can predict PTSD symptoms in young individuals exposed to trauma.Study methodology: Two pre-trained large language models and two benchmark models were fine-tuned and trained to predict PTSD symptom severity from children's autobiographical narratives of a traumatic event. Data comprised narratives collected one month post-trauma from 400 individuals aged 7-17 years old who experienced a psychological trauma that led to attendance at emergency departments in the United Kingdom (N = 178) and South Africa (N = 222), as well as self-reported PTSD symptoms and trauma memory features.

Findings

Both pre-trained and benchmark models demonstrated poor predictive performance across trauma narratives in the United Kingdom, South Africa, and the combined datasets (e.g. RoBERTa R² = -.05; LASSO R² ≈ 0). However, adding self-reported trauma memory features, disorganisation, and sensory vividness improved the benchmark models' performances, especially in the UK dataset (e.g. LASSO R² = .57; XGBoost R² = .45).

Conclusions

These findings indicate that while trauma narratives alone offer limited predictive value, incorporating self-reported trauma memory characteristics substantially enhances model performance, highlighting the importance of focusing on subjective reports to develop scalable automated tools for PTSD risk prediction in youth. Trauma narratives alone offer limited predictive value for PTSD in youth. Advanced machine learning models, including large language models, failed to predict post-traumatic stress symptoms based on children’s and adolescents’ trauma accounts.The study used trauma narratives from children in both the UK and South Africa. Despite cultural and linguistic differences, predictive performance was similarly limited across both samples, suggesting a broader developmental challenge in detecting PTSD through narrative language.Subjective memory features substantially enhance predictive performance. Incorporating self-reported ratings of memory disorganisation and sensory vividness led to significant improvements in model accuracy.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Stress Disorders, Post-TraumaticMachine LearningNarrationMemory, EpisodicPsychological TraumaHumansAdolescentChildFemaleMaleUnited KingdomSouth Africa

Stichwörter der Autor:innen

Aprendizaje automáticoLLMsMachine learningPTSDTEPTchildren and adolescentsmodelos de lenguaje grandes (LLMs)modelos predictivosnarrativas del traumaniños y adolescentespredictive modelstrauma narratives

Autor:innen und Institutionen

  • Department of Psychology, University of Bath, Bath, UK.
  • Department of Global Health, Institute for Life Course Health Research, Stellenbosch University, Stellenbosch, South Africa.
  • Department of Clinical Psychology & Psychological Therapies, Norwich Medical School, University of East Anglia, Norwich, UK.
  • School of Nursing and Midwifery, Queen's University Belfast, Belfast, UK.

ORCID-Profile: Alessandra Giuliani, Richard Meiser-Stedman, Mark Tomlinson

Zitieren

APA

Giuliani, A., Sharp, T., Chideya, Y., Meiser-Stedman, R., Tomlinson, M., & Halligan, S. L. (2025). Can machine learning predict PTSD symptoms from trauma narratives of children and adolescents?. European Journal of Psychotraumatology. https://doi.org/10.1080/20008066.2025.2589709

Vancouver

Giuliani A, Sharp T, Chideya Y, Meiser-Stedman R, Tomlinson M, Halligan SL. Can machine learning predict PTSD symptoms from trauma narratives of children and adolescents?. European Journal of Psychotraumatology. 2025. doi:10.1080/20008066.2025.2589709

BibTeX
@article{Giuliani2025,
  author = {Giuliani, Alessandra and Sharp, Tamsin and Chideya, Yeukai and Meiser-Stedman, Richard and Tomlinson, Mark and Halligan, Sarah L},
  title = {Can machine learning predict PTSD symptoms from trauma narratives of children and adolescents?},
  journal = {European Journal of Psychotraumatology},
  year = {2025},
  doi = {10.1080/20008066.2025.2589709}
}
Kostenlos starten →