AI Science Tracker
EN

University of Delaware.

1 discovery from 2024

Discoveries

Bias Patterns in the Application of LLMs for Clinical Decision Support: A Comprehensive Study.

Source check: Source checked. Not independent scientific confirmation.

Finding: The journal version reports bias patterns across eight LLMs and clinical-vignette tasks; an earlier arXiv version reported the same study, and the journal version adds an author and expanded analyses. The findings are task- and prompt-specific and do not establish bias across all clinical use. AI role: eight general-purpose and clinically trained LLMs evaluated across three standardized question-answering datasets using demographic red-teaming and prompt comparisons. The later peer-reviewed journal version appeared in Delaware Journal of Public Health in 2026.

Technology: Eight large language models tested on clinical vignettes

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
Eight large language models tested on clinical vignettes
Test and result
Primary journal and earlier arXiv version study bias patterns in eight LLMs across clinical vignette tasks. Journal version adds a coauthor and expanded analyses; both are one study.
Limitation / still unknown
The results depend on the selected eight models, three vignette datasets, demographic attributes, prompts, and scoring; clinical outcomes and model behavior in deployed care were not evaluated. The arXiv and journal versions are one study.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://arxiv.org/abs/2404.15149 · https://arxiv.org/html/2404.15149 · https://doi.org/10.32481/djph.2026.03.10 · https://pmc.ncbi.nlm.nih.gov/articles/PMC13048765/ · https://pubmed.ncbi.nlm.nih.gov/41943726/

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: linked separately.

Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open the interactive profile for the timeline, or browse all organizations and model families.