Large Language Models and Text Embeddings for Detecting Depression and Suicide in Patient Narratives
This cross-sectional study of SCT narratives from psychiatric patients suggests that LLMs and text-embedding models may effectively detect depression and suicide risk, particularly using self-concept narratives. However, while these models demonstrated potential for detecting mental health risks, further improvements in performance and safety are essential before clinical application.
Authors
- Hyeonjin Kim (ORCID: https://orcid.org/0000-0002-3505-0054)
- Soomin Jang (ORCID: https://orcid.org/0000-0003-3515-7211)
- Da Young Oh
- Su Mi Park
- So Young Yoo (ORCID: https://orcid.org/0000-0001-9964-126X)
- Silvia Kyungjin Lho (ORCID: https://orcid.org/0000-0001-7164-9358)
- Hyon Lee (ORCID: https://orcid.org/0000-0002-0273-8960)
- Hee Yeon Jung (ORCID: https://orcid.org/0000-0002-0578-2211)
- Junyoung Lee (ORCID: https://orcid.org/0009-0005-8080-2082)
- Sang Cheol Park
Institutions
- Hannam University (KR)
- Seoul National University (KR)
- Korea University (KR)
- Seoul Metropolitan Government (KR)
- SNUH SMG-SNU Boramae Medical Center (KR)
Publication Details
- Journal
- JAMA Network Open
- Published
- 2025-05-23
- DOI
- https://doi.org/10.1001/jamanetworkopen.2025.11922
- Citations
- 9
- Primary Topic
- Mental Health via Writing
- Type
- article
- Field-Weighted Citation Impact
- 21.95
Funders
- National Research Foundation
- National Research Foundation of Korea