Enhancing clinicians’ trust in large language models via transparent source attribution: A randomized controlled evaluation in uro-oncology
Our findings show that combining LLMs with RAG with in-line references and source text previews markedly enhances perceived source attribution and verifiability compared to state-of-the-art conventional LLMs. Importantly, this approach is readily transferable across medical subspecialties, enabling reliable and up-to-date clinical decision support.
Authors
- Jana Theres Winterstein (ORCID: https://orcid.org/0000-0003-3346-9537)
- Thomas Stefan Worst (ORCID: https://orcid.org/0000-0001-5429-126X)
- Maurice Stephan Michel (ORCID: https://orcid.org/0000-0001-7424-9085)
- Martin J. Hetz
- Lasse Maywald (ORCID: https://orcid.org/0000-0001-5163-3383)
- Frederik Wessels (ORCID: https://orcid.org/0000-0003-4213-8692)
- Christoph Wies (ORCID: https://orcid.org/0000-0001-7136-298X)
- Niklas Westhoff (ORCID: https://orcid.org/0000-0003-2945-2302)
- Sarah Haggenmüller (ORCID: https://orcid.org/0000-0002-7961-1781)
- Titus Josef Brinker (ORCID: https://orcid.org/0000-0002-3620-5919)
- Nicolas Carl (ORCID: https://orcid.org/0009-0002-0080-3455)
- Maurin Helen Mangold
Institutions
- German Cancer Research Center (DE)
- Heidelberg University (DE)
- University Hospital Heidelberg (DE)
- University Medical Centre Mannheim (DE)
Publication Details
- Journal
- European Journal of Cancer
- Published
- 2025-12-11
- DOI
- https://doi.org/10.1016/j.ejca.2025.116168
- Citations
- 1
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 2.48