Large Language Model–Assisted Risk-of-Bias Assessment in Randomized Controlled Trials Using the Revised Risk-of-Bias Tool: Evaluation Study
LLMs achieved commendable accuracy when guided by structured prompts, particularly through processing methodological details through structured reasoning. While not replacing human assessment, LLMs demonstrate strong potential for assisting RoB2 evaluations. Larger studies with improved prompting could enhance performance.
Authors
- Bei Pan (ORCID: https://orcid.org/0000-0002-0370-9571)
- Honghao Lai (ORCID: https://orcid.org/0000-0001-7913-6207)
- Jianing Liu (ORCID: https://orcid.org/0000-0001-5048-4566)
- Jinhui Tian (ORCID: https://orcid.org/0000-0002-0054-2454)
- Weilong Zhao (ORCID: https://orcid.org/0000-0001-8909-5322)
- Long Ge (ORCID: https://orcid.org/0000-0002-3555-1107)
- J. Huang (ORCID: https://orcid.org/0000-0001-7925-5409)
- Danni Xia (ORCID: https://orcid.org/0009-0004-2226-6923)
- Mingyao Sun (ORCID: https://orcid.org/0009-0005-4457-6401)
- Jiayi Liu (ORCID: https://orcid.org/0009-0001-5127-2116)
- Chunyang Bai (ORCID: https://orcid.org/0009-0002-4183-686X)
Institutions
- Peking University (CN)
- Gansu University of Traditional Chinese Medicine (CN)
- Southern Medical University (CN)
- Lanzhou University (CN)
Publication Details
- Journal
- Journal of Medical Internet Research
- Published
- 2025-05-05
- DOI
- https://doi.org/10.2196/70450
- Citations
- 7
- Primary Topic
- Meta-analysis and systematic reviews
- Type
- article
- Field-Weighted Citation Impact
- 9.79