Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Replicability and Validity of a New Artificial-Intelligence Assessment of Posttraumatic Stress Disorder From Patient Language : A Sequential Evaluation With Model Preregistration

Kjell, Oscar LU orcid ; Ganesan, Adithya V. ; Boyd, Ryan L. ; Oltmanns, Joshua ; Rivero, Alfredo ; Feltman, Scott ; Carr, Melissa A. ; Alves, Jorge ; Luft, Benjamin and Kotov, Roman , et al. (2026) In Clinical Psychological Science
Abstract

Artificial intelligence (AI) shows promise in identifying psychopathology through language, but replicability in AI models remains challenging. We develop an AI-based language assessment of posttraumatic-stress-disorder (PTSD) severity and introduce the sequential evaluation with model preregistration to rigorously evaluate its validity and replicability. This design includes two phases: development with preregistration and evaluation. Data included development (N = 1,437) and prospective (N = 346) samples, in which participants described their lives during automated interviews. In the prospective sample, preregistered models correlated with PTSD CheckList scores (r = .38, p < .001) and converged with PTSD diagnosis (area under the... (More)

Artificial intelligence (AI) shows promise in identifying psychopathology through language, but replicability in AI models remains challenging. We develop an AI-based language assessment of posttraumatic-stress-disorder (PTSD) severity and introduce the sequential evaluation with model preregistration to rigorously evaluate its validity and replicability. This design includes two phases: development with preregistration and evaluation. Data included development (N = 1,437) and prospective (N = 346) samples, in which participants described their lives during automated interviews. In the prospective sample, preregistered models correlated with PTSD CheckList scores (r = .38, p < .001) and converged with PTSD diagnosis (area under the curve [AUC] = .76; outperforming demographics and trauma exposures: AUC = .61, p < .01). We found that for each standard-deviation increase, mental-health-care expenditure rose by $696.50 (p < .001). Our preregistered PTSD model assessments are replicable in prospectively collected clinical data and showed external validity against expense criteria. With further development, such models can be used to screen for PTSD or monitor treatment response, especially in telehealth or automated interviews, in which deployment can be seamless.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; ; ; and , et al. (More)
; ; ; ; ; ; ; ; ; and (Less)
organization
publishing date
type
Contribution to journal
publication status
in press
subject
keywords
depression, disaster responders, language-based assessments, open materials, oral interviews, posttraumatic stress disorder, preregistration, World Trade Center
in
Clinical Psychological Science
publisher
SAGE Publications
external identifiers
  • scopus:105037757276
ISSN
2167-7026
DOI
10.1177/21677026261439026
language
English
LU publication?
yes
id
bd5e42c2-ceef-4e36-b878-85b00da778b1
date added to LUP
2026-08-10 13:09:47
date last changed
2026-08-10 13:09:47
@article{bd5e42c2-ceef-4e36-b878-85b00da778b1,
  abstract     = {{<p>Artificial intelligence (AI) shows promise in identifying psychopathology through language, but replicability in AI models remains challenging. We develop an AI-based language assessment of posttraumatic-stress-disorder (PTSD) severity and introduce the sequential evaluation with model preregistration to rigorously evaluate its validity and replicability. This design includes two phases: development with preregistration and evaluation. Data included development (N = 1,437) and prospective (N = 346) samples, in which participants described their lives during automated interviews. In the prospective sample, preregistered models correlated with PTSD CheckList scores (r = .38, p &lt; .001) and converged with PTSD diagnosis (area under the curve [AUC] = .76; outperforming demographics and trauma exposures: AUC = .61, p &lt; .01). We found that for each standard-deviation increase, mental-health-care expenditure rose by $696.50 (p &lt; .001). Our preregistered PTSD model assessments are replicable in prospectively collected clinical data and showed external validity against expense criteria. With further development, such models can be used to screen for PTSD or monitor treatment response, especially in telehealth or automated interviews, in which deployment can be seamless.</p>}},
  author       = {{Kjell, Oscar and Ganesan, Adithya V. and Boyd, Ryan L. and Oltmanns, Joshua and Rivero, Alfredo and Feltman, Scott and Carr, Melissa A. and Alves, Jorge and Luft, Benjamin and Kotov, Roman and Schwartz, H. Andrew}},
  issn         = {{2167-7026}},
  keywords     = {{depression; disaster responders; language-based assessments; open materials; oral interviews; posttraumatic stress disorder; preregistration; World Trade Center}},
  language     = {{eng}},
  publisher    = {{SAGE Publications}},
  series       = {{Clinical Psychological Science}},
  title        = {{Replicability and Validity of a New Artificial-Intelligence Assessment of Posttraumatic Stress Disorder From Patient Language : A Sequential Evaluation With Model Preregistration}},
  url          = {{http://dx.doi.org/10.1177/21677026261439026}},
  doi          = {{10.1177/21677026261439026}},
  year         = {{2026}},
}