Estimating depression severity in narrative clinical notes using large language models
A foundational LLM performed poorly but consistently across subgroups in imputing PHQ-9 scores from notes when actual PHQ-9 reporting was ablated. This result suggests the extent to which inclusion of PROMs may impoverish documentation of psychiatric symptoms.
Authors
- Thomas H. McCoy (ORCID: https://orcid.org/0000-0002-5624-0439)
- Víctor M. Castro (ORCID: https://orcid.org/0000-0001-7390-6354)
- Roy H Perlis
Institutions
- Massachusetts General Hospital (US)
Publication Details
- Journal
- Journal of Affective Disorders
- Published
- 2025-04-05
- DOI
- https://doi.org/10.1016/j.jad.2025.04.014
- Citations
- 5
- Primary Topic
- Treatment of Major Depression
- Type
- article
- Field-Weighted Citation Impact
- 6.97
Funders
- National Institutes of Health