Preprint · 2026 · Open access
Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review
Research Square · Published 28 May 2026 · DOI 10.21203/rs.3.rs-9576482/v1
In brief
“Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review” is a 2026 preprint by Ehsan Safaee and 4 co-authors, published in Research Square. OpenAlex counts 0 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at Research Square.
- Citations
- 0
- References
- 17
- Authors
- 5
Abstract
The rapid integration of Large Language Models (LLMs) into clinical workflows has outpaced the establishment of rigorous frameworks to ensure patient safety and model reliability. We conducted a systematic scoping review of n = 1,054 full-text articles published between 2017 and 2025 to characterize the current landscape of LLM evaluation in healthcare. Using an automated extraction pipeline validated for high-fidelity data retrieval, we mapped study characteristics across the MedHELM taxonomy and 16 distinct evaluation constructs. Our analysis reveals a profound "safety-utility gap": while technical validity is nearly universal (95.1%), safety-critical dimensions including bias, toxicity, and uncertainty are evaluated in fewer than 10% of clinical decision support tasks. Furthermore, we identify extreme heterogeneity in evaluation metrics, with leading measures such as accuracy accounting for less than 20% of reported evaluations in most constructs. This fragmentation, combined with a lack of "ecological validity" in testing against real-world clinical documentation, prevents meaningful cross-study comparisons and complicates regulatory oversight. These findings demonstrate that current benchmarking is insufficient for clinical deployment. We conclude that a transition toward standardized, safety-first evaluation frameworks is essential to bridge the gap between technical performance and bedside utility.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Safaee, E., Mehrabi, A., Shariatnia, M. M., Rouzrokh, P., & Rouzrokh, P. (2026). Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review. Research Square. https://doi.org/10.21203/rs.3.rs-9576482/v1
MLA 9
Safaee, Ehsan, et al. “Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review.” Research Square, 2026. https://doi.org/10.21203/rs.3.rs-9576482/v1.
Chicago
Safaee, Ehsan, Ali Mehrabi, M. Moein Shariatnia, Parsa Rouzrokh, and Pouria Rouzrokh. 2026. “Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review.” Research Square. https://doi.org/10.21203/rs.3.rs-9576482/v1.
Harvard
Safaee, E. et al. (2026) ‘Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review’, Research Square. doi:10.21203/rs.3.rs-9576482/v1.
IEEE
E. Safaee, A. Mehrabi, M. M. Shariatnia, P. Rouzrokh, and P. Rouzrokh, “Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review,” Research Square, 2026, doi: 10.21203/rs.3.rs-9576482/v1.
BibTeX
@misc{safaee2026evaluation,
title = {{Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review}},
author = {Safaee, Ehsan and Mehrabi, Ali and Shariatnia, M. Moein and Rouzrokh, Parsa and Rouzrokh, Pouria},
howpublished = {Research Square},
year = {2026},
doi = {10.21203/rs.3.rs-9576482/v1},
url = {https://doi.org/10.21203/rs.3.rs-9576482/v1}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- large language models
- LLM evaluation
- health care
- clinical decision support
- patient safety
- bias
- toxicity
- uncertainty
Work with this paper
Most-cited papers in Medicine
- Yaşlı Kadınlarda Üreme SağlığıDemet Kokanalı, Yaprak Engin Üstün · Jinekoloji-Obstetrik ve Neonatoloji Tıp Dergisi · 2019 · 169,630 citations
- Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 CountriesHyuna Sung, Adalberto M. Filho, Mathieu Laversanne and 5 more · CA A Cancer Journal for Clinicians · 2021 · 119,885 citations
- “Mini-mental state”Marshal F. Folstein, Susan E. Folstein, Paul R. McHugh · Journal of Psychiatric Research · 1975 · 90,424 citations
- Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countriesFreddie Bray, Jacques Ferlay, Isabelle Soerjomataram and 3 more · CA A Cancer Journal for Clinicians · 2018 · 87,396 citations
- Hallmarks of Cancer: The Next GenerationDouglas Hanahan, Robert A. Weinberg · Cell · 2011 · 69,758 citations
Questions about this paper
How do I cite this paper in APA style?
Safaee, E., Mehrabi, A., Shariatnia, M. M., Rouzrokh, P., & Rouzrokh, P. (2026). Evaluation Practices for Large Language Models in Healthcare: A Systematic Scoping Review. Research Square. https://doi.org/10.21203/rs.3.rs-9576482/v1
How many times has this paper been cited?
0 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at Research Square: sign in and open the PDF from this page.
Who wrote this paper?
Ehsan Safaee, Ali Mehrabi, M. Moein Shariatnia, Parsa Rouzrokh, Pouria Rouzrokh.
What is the DOI of this paper?
10.21203/rs.3.rs-9576482/v1, permanently at https://doi.org/10.21203/rs.3.rs-9576482/v1. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.