Can Large Language Models be a Viable Tool for Consensus Working Groups? Experience of the Ventral Rectopexy Expert Consensus Group
OpenEvidence outperformed Gemini 1.5 Pro and ChatGPT-4o on content appropriateness and peer-reviewed citation quantity and quality. Chatbot-generated text was indistinguishable from expert-authored consensus to subject matter experts. Large language models as an early-stage research tool may be viable for future consensus working groups, provided transparent disclosure and rigorous oversight. See Video Abstract.
Authors
- André D’Hoore
- Frank G. Lee (ORCID: https://orcid.org/0000-0003-0140-3963)
- William Perry (ORCID: https://orcid.org/0000-0001-8718-8306)
- P. Christensen
- Andrew R. L. Stevenson (ORCID: https://orcid.org/0000-0001-5639-3526)
- Jane Vermunt (ORCID: https://orcid.org/0000-0002-0906-7258)
- Ellen L. Larson (ORCID: https://orcid.org/0000-0002-9420-5941)
- Brooke Gurland (ORCID: https://orcid.org/0000-0002-3868-7747)
- Anders Mellgren
- Liliana Bordeianou (ORCID: https://orcid.org/0000-0003-3418-6774)
- Rowan Collinson (ORCID: https://orcid.org/0000-0002-5678-5586)
- Frédéric Ris (ORCID: https://orcid.org/0000-0001-7421-6101)
- Julie Cornish (ORCID: https://orcid.org/0000-0003-4360-4472)
- Carlo Ratto
- Cornelius A. Thiels
- On behalf of the Ventral Rectopexy Consensus AI Group and the Ventral Rectopexy Working Group
Institutions
- Cardiff and Vale University Health Board (GB)
- Mayo Clinic (US)
- The University of Queensland (AU)
- Universitair Ziekenhuis Leuven (BE)
- King Faisal Specialist Hospital & Research Centre (SA)
- Auckland City Hospital (NZ)
- Aarhus University Hospital (DK)
- WinnMed (US)
- Agostino Gemelli University Polyclinic (IT)
- Massachusetts General Hospital (US)
- Hôpital Beau-Séjour (CH)
- Stanford Medicine (US)
Publication Details
- Journal
- Diseases of the Colon & Rectum
- Published
- 2026-01-06
- DOI
- https://doi.org/10.1097/dcr.0000000000004058
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00