Identifying Deprescribing Opportunities With Large Language Models in Older Adults: Retrospective Cohort Study
This study highlights the potential of LLMs to support deprescribing decisions in the ED by effectively filtering relevant criteria. However, challenges remain in applying these criteria to complex clinical scenarios, as the LLM demonstrated poor performance on more intricate decision-making tasks, with its reported confidence often failing to align with its actual success in these cases. The findings underscore the need for clearer deprescribing guidelines, improved LLM calibration for real-world use, and better integration of human-artificial intelligence workflows to balance artificial intelligence recommendations with clinician judgment.
Authors
- Naga Sasidhar Kanaparthy (ORCID: https://orcid.org/0000-0002-2342-2412)
- Andrew J. Loza (ORCID: https://orcid.org/0000-0002-1754-5133)
- David Chartash (ORCID: https://orcid.org/0000-0002-0265-330X)
- Brian W. Patterson (ORCID: https://orcid.org/0000-0002-4584-3808)
- Christine Dien (ORCID: https://orcid.org/0009-0008-8850-3046)
- Mark Iscoe (ORCID: https://orcid.org/0000-0003-4446-2488)
- Catherine Xie Wright (ORCID: https://orcid.org/0000-0001-7638-4053)
- D. Wright (ORCID: https://orcid.org/0000-0002-2564-7754)
- Ling Chi (ORCID: https://orcid.org/0000-0002-8270-9245)
- J. D. M. Albano (ORCID: https://orcid.org/0000-0002-4451-0438)
- Soraya Fereydooni (ORCID: https://orcid.org/0000-0002-1264-6461)
- Thomas Huang (ORCID: https://orcid.org/0000-0001-9056-7016)
- Vimig Socrates (ORCID: https://orcid.org/0000-0001-7955-9875)
- Richard A. Taylor (ORCID: https://orcid.org/0000-0002-9082-6644)
Institutions
- University College Dublin (IE)
- United States Department of Veterans Affairs (US)
- Yale New Haven Hospital (US)
- University of Wisconsin–Madison (US)
- Yale University (US)
- VA Connecticut Healthcare System (US)
Publication Details
- Journal
- JMIR Aging
- Published
- 2025-02-21
- DOI
- https://doi.org/10.2196/69504
- Citations
- 3
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 1.96